OPUS 4 - Der Technischen Fakultät der Friedrich …...cells. We demonstrate that fairly efficient...
182
Engineering Stable Interfaces for Printed Solar Cells by Rationalizing Material Induced Loss Mechanisms Rationale Entwicklung und Optimierung von stabilen Ladungsträgerextraktionsschichten für die gedruckte Photovoltaik Der Technischen Fakultät der Friedrich-Alexander-Universität Erlangen-Nürnberg zur Erlangung des Doktorgrades Dr.-Ing. vorgelegt von Hong Zhang aus Yunnan, China
Transcript of OPUS 4 - Der Technischen Fakultät der Friedrich …...cells. We demonstrate that fairly efficient...
Engineering Stable Interfaces for Printed Solar Cells
by Rationalizing Material Induced Loss Mechanisms
Rationale Entwicklung und Optimierung von stabilen
Ladungsträgerextraktionsschichten für die gedruckte Photovoltaik
Der Technischen Fakultät
der Friedrich-Alexander-Universität
Erlangen-Nürnberg
zur
Erlangung des Doktorgrades Dr.-Ing.
vorgelegt von
Hong Zhang
aus Yunnan, China
Als Dissertation genehmigt
von der Technischen Fakultät
der Friedrich-Alexander-Universität Erlangen-Nürnberg
Tag der mündlichen Prüfung: 13.10.2016
Vorsitzender des Promotionsorgans: Prof. Dr.-Ing. Reinhard Lerch
Gutachter: Prof. Dr. Christoph J. Brabec
Gutachter: Prof. Dr. Dieter Neher
Dedicated to my beloved parents
i
Acknowledgments
First of all, my deepest appreciation goes to my advisor Professor Dr. Christoph J. Brabec for
his guidance through my entire research at Erlangen. I am very fortunate to have an
opportunity to conduct my doctoral research in such a wonderful group (i-MEET) with a
perfect combination of freedom and guidance. Your enthusiasm on science, the highest
standards of research, great scientific taste and sense of humor inspire me to achieve the level
that I could never dream before and will definitely have an intense impact in my future
career. Thank you, sir, for everything.
I would like to express my sincere gratitude to all my colleagues and friends at i-MEET and
ZAE for their great cooperation and atmosphere during my doctoral study. In particularly, I
would like to thank Dr. Tayebeh Ameri for the academic support and guidance. I would like
to thank Dr. Tobias Stubhan who served as my mentors at my early stage in the group and
taught me a lot about organic photovoltaics both in theory and experiment. I would like to
thank all the people who have contributed to this thesis, especially Dr. Ning Li, Dr. Hamed
Azimi, Dr. Gebhard J Matt, Dr. Fei Guo, Dr. Jie Min, Yi Hou, Lili Ke for the fruitful
collaboration. I thank Thomas Przybilla, Dr. Stefanie Fladischer and Prof. Erdmann Spiecker
from the Center for Nanoanalysis and Electron Microscopy for the great help with SEM and
TEM characterizations, Prof. Marcus Halik from the OMD group at the Institute of Polymer
Materials for the access to AFM measurements.
I wish to thank Dr. Zakutayev and all the members of the Materials Discovery group at the
National Renewable Energy Laboratory (NREL) for having and helping me during my
research stay.
Meanwhile, I wish to thank Prof. Neal R. Armstrong and Dr. R. Clayton Shallcross at
University of Arizona for the help with UPS and XPS measurements. I am indebted to Prof.
Ullrich Scherf and Mario Kraft from IfP, Bergische Universität Wuppertal and Prof. Xuhui
Zhu’s group from South China University of Technology (SCUT) for offering amazing
organic materials, particularly the functionalized small molecules. I learned a lot of the
chemistry and material properties about organic molecules in a close collaboration with them.
ii
I also wish to thank Prof. L.J.A. (Jan Anton) Koster at University of Groningen for the help
with simulation the effects of bimolecular recombination at the interface of active
layer/contact with 1D drift-diffusion modeling.
Last but not the least I am indebted to my family. I would like to thank my parents and my
sister for their consistent encouragement and support throughout my life.
Hong Zhang
Erlangen
March, 2016
iii
Abstract
Solar energy is almost infinitely available and a clean energy source of the future. Organic
solar cells (OSCs) are continuously drawing attention from both the academic and industrial
communities and considered as a promising candidate for renewable energy sources of the
next generation due to their non-toxicity, low-costs, high sustainability and especially their
light weight and compatibility with flexible substrates. This dissertation targets on the
development and understanding of high efficiency OSCs with interface modification layer
and the simplified technological process to fabricate efficient devices using roll-to-roll
compatible processing techniques.
The first part of this thesis focuses on the development and understanding of interface contact
between the PCBM and metal oxides for OSCs. Modification of metal oxides with a solution
processed low-cost alkali hydroxide layers increases the efficiency of the inverted
architecture device by dominantly tuning the barrier between the conduction band of metal
oxides and the LUMO of PCBM in the active layer. Since the presence of a large interface
dipole and a new interface state between the Fermi energy and the fullerene HOMO for alkali
hydroxide-modified metal oxides contacts, alkali hydroxide-modified metal oxides contacts
enhances electron extraction, reduces contact resistivity, and suppresses bimolecular
recombination, leading to a remarkable PCE enhancement for pDPP5T-2:PCBM based OSCs
from 3.3% to approximately 6.0%. These novel interfacial gap states are hypothesized to be
electronically hybridized with the contact. Combined with a numerical device model based
on a 1D drift–diffusion approach, the effect of reduced interfacial majority carrier at the
electrodes is simulated. The majority carrier density at the respective interface is reduced
when inserting an interface modification layer. The bulk bimolecular recombination with the
photo generated minority carriers gets reduced. The alkali hydroxide-modified metal oxides
contacts induces suppression of the bulk recombination in organic bulk heterojunction
photovoltaics.
iv
In the second part of this thesis, design of interface layers by engineering is demonstrated to
print efficient devices using roll-to-roll compatible processing techniques under ambient
conditions. A simple approach to printing efficient, inverted OSCs with a self-organized
cathode interface layer Phen-NaDPO blended into a pDPP5T-2/PCBM system is introduced.
We observe a spontaneous, surface energy driven migration of Phen-NaDPO towards the
ZnO interface and a subsequent formation of electron selective and barrier free extraction
contacts. In the presence of 0.5 wt% Phen-NaDPO, a PCE of 5.4% is achieved for inverted
device based on an ITO/ZnO cathode. In addition, we successfully develop an s-MoOX/PEG
ink as HTL to fabricate all solution processed inverted OSCs via doctor-blading in air, which
shows a performance comparable to those with evaporated MoOX. Excellent wetting of s-
MoOX:PEG solution on the active layer leads to highly uniform layers with complete surface
coverage and superior hole selectivity. As an alternative to PEDOT:PSS and evaporated
MoOX, s-MoOX:PEG is established as a highly promising hole transport material for efficient
and stable inverted organic solar cells.
Finally, we focus on the development of organic–metal interfaces based on perovskite solar
cells. We demonstrate that fairly efficient perovskite solar cell can be fabricated by inserting
an ultrathin polyelectrolyte layer based on the interface layers all solution-processed at low
temperature, either PEIE or P3TMAHT. The PCE increases from 8.53% to 12.01% (PEIE)
and 11.28% (P3TMAHT) for the solution-processed polyelectrolyte-modified interfaces.
Moreover, we introduce a water-free dispersion of PEDOT to replace spiro-MeOTAD in
perovskite solar cells, and the device shows a maximum PCE of 11.75% and more than 800 h
stability under ambient environmental conditions without packaging. We find that the density
of sub bandgap states at the interface of the perovskite/HTL seems to depend on the nature of
PEDOT. This water-free PEDOT can become a very promising candidate for upscaling
production of perovskite solar cells.
v
Zusammenfassung
Solarenergie ist eine beinahe unerschöpfliche und saubere Energiequelle für die Zukunft.
Organische Solarzellen (OSCs) ziehen kontinuierlich mehr Aufmerksamkeit von der
akademischen und industriellen Gemeinschaft auf sich und werden als ein vielversprechender
Kandidat für erneuerbare Energien der nächsten Generation angesehen, da sie nicht toxisch,
günstig, flexibel und leicht sind. Diese Dissertation zielt auf die Entwicklung und das
Verständnis von Hocheffizienzsolarzellen mit Zwischenschichtmodifikationen und
vereinfachten technologischen Prozessen zum Drucken von Bauteilen, die kompatibel zur
Rolle-zu-Rolle-Prozessierung sind.
Der erste Teil der Arbeit fokussiert sich auf die Entwicklung und das Verständnis des
Kontaktverhaltens zwischen PCBM und Metalloxiden in OSCs. Modifikation von
Metalloxiden mit lösungsprozessierten, günstigen Alkalihydroxidschichten verbessert die
Effizienz von invertierten Bauteilen vorwiegend durch das Einstellen der Barriere zwischen
Leitungsband der Metalloxide und dem LUMO des PCBM in der aktiven Schicht.
Alkalihydroxid-modifierte Metalloxide verbessert die Elektronenextraktion, reduziert den
Kontaktwiderstand und unterdrückt die Bimolekulare Rekombination, was zu einer
erstaunlichen Effizienzerhöhung für pDPP5T-2:PCBM basierte OSCS von 3.3 % auf ca.
6.0 %. Diese neuartigen Greznschichtzustände werden vermutlich mit dem Kontakt
elektronisch hybridisiert. Kombiniert mit einem numerischen Bauelementmodell basierend
auf dem 1D „drift-diffusion“ Ansatz wird der Effekt der reduzierten
Grenzflächenmajoritätsladungsträger an den Elektroden simuliert. Die
Majoritätsladungsträgerdichte an den entsprechenden Elektroden wird durch die
Modifikationsschichten reduziert. Die bimolekulare Rekombination in der aktiven Schicht
mit den photogenerierten Minoritätsladungsträgern wird reduziert. Die Alkalihydroxide
induzieren reduzierte Rekombination in der aktiven Schicht von organischen
Photovoltaikzellen.
Im zweiten Teil wird das Design der Zwischenschichten durch das Einstellen der
vi
Druckprozesse an Luft eingeführt. Es konnte erfolgreich dargestellt werden, dass durch einen
einfachen Ansatz invertierte, hocheffiziente OSCs mit selbstorganisierenden Phen-NaDPO
Kathodenschichten gedruckt werden können, wenn Phen-NaDPO in das pDPP5T-2/PCBM
System hineingemischt wird. Es wird spontane oberflächenenergiegetriebene Migration des
Phen-NaDPO zum Zno beobachtet, was einen elektronenselektiven, barrierefreien Kontakt
erzeugt. Durch 0,5wt% Phen-NaDPO wird eine Effizienz von 5,4% erreicht. Zudem wurde
eine MoOx/PEG Tinte entwickelt, durch die an Luft gedruckte OSCs vergleichbare Effizienz
zu denen mit aufgedampften MoOx HTLs erreichten. Sehr gutes Benetzen der MoOx:PEG
Tinte führt zu gleichmäßigen und sehr lochselektiven Schicht. Damit ist dies eine gute
Alternative zu PEDOT:PSS und aufgedampftem MoOx für stabile invertierte OSCs, die
kompatibel zur Rolle-zu-Rolle Prozessierung sind.
Am Ende fokussiert sich die Arbeit auf die Entwicklung für Zwischenschichten für
Perowskitsolarzellen. Hier werden effiziente Perowskitsolarzellen durch den Einsatz von
ultradünnen Polyelektrolytschichten bei niedrigen Temperaturen mit PEIE oder P3TMAHT
erzeugt. Die PCE steigt von 8,53% auf 12,01% (PEIE) und 11,28% (P3TMAHT). Zudem
wird eine wasserfreie PEDOT-Dispersion eingeführt, mit der spiro-MeOTAD in
Perowskitsolarzellen ersetzt werden kann. Diese Zellen zeigten eine maximale Effizienz von
11,75% und mehr als 800 Stunden Stabilität an Luft ohne Verkapselung. Hier wurde auch
herausgefunden, dass die Dichte der Zustände unterhalb der Bandlücke an der Grenzfläche
Perowskit/HTL von der Natur des PEDOT abhängt. Dieses wasserfreie PEDOT kann auch
ein sehr vielversprechender Kandidat für die industrielle Prozessierung von
Perowskitsolarzellen werden.
vii
Abbreviations
AFM Atomic Force Microscopy Ag Silver
Al Aluminum AOH Alkali hydroxide
Au Gold
AZO Aluminum doped Zinc Oxide
Ba Barium
Ba(OH)2 Barium Hydroxide
BCP Bathocuproine
BDEW Bundesverband der Energie- und Wasserwirtschaft
BHJ Bulk-heterojucntion
Ca Calcium
CB Conduction band
CNL Charge neutrality level
CPEs Conjugated polyelectrolytes
C60 Fullerene
C60-SAM Modified fullerene derivatives
C-V Capacitance versus voltage
DFT Discrete Fourier Transform
DIP Diindenoperylene
DPP Diketopyrrolopyrrole
D-A Donor - Acceptor
EA Electron affinity
EEL Electron extraction layer
EDXS Energy-dispersive X-ray spectroscopy
Eg Bandgap
EHOMO Energy level of Highest Occupied Molecular Orbital
ELUMO Energy level of Lowest Unoccupied Molecular Orbital
viii
EQE External quantum efficiency
ETL Electron transporting layer
FIB Focused ion beam
HOMO Highest occupied molecular orbital
HTL Hole transporting layer
HTM Hole transporting materials
ICBA Indene-C60 bisadduct
ICT Intramolecular charge transfer
IE Ionization energy
ITO Indium tin oxide
j-V Current density-voltage
KPM Kelvin probe microscopy
LCOE levelized cost of electricity
LUMO Lowest unoccupied molecular orbital
MOX Metal oxides
MoOX Molybdenum trioxide
OLED Organic light-emitting diodes
OPV Organic photovoltaic
OSC Organic solar cell
PCBDAN [6,6]-phenyl-C61-butyric acid 2-((2-(dimethylamino)-ethyl)(methyl)amino)ethyl ester
PCBM [6,6]-Phenyl-C61-Butyric-acid-Methyl ester
PCE Power conversion efficiency
PE Polyelectrolyte layers
PEDOT:PSS Poly(ethylenedioxythiophene):poly(styrene sulfonic acid)
PEG Polyethylene glycol
PEIE Polyethylenimine, 80% ethoxylated
PFN Poly[(9,9-bis(3’-(N,N-dimethylamino)propyl)-2,7- fluorene)-alt-2,7-(9,9-dioctylfluorene)]
Phen-NaDPO (2-(1,10-phenanthrolin-3-yl)naphth-6- yl) diphenylphosphine oxide
PHJ Planar heterojunction
PL Photoluminescence
PV Photovoltaic
ix
P3HT Poly(3-hexylthiophene-2,5-diyl)
P3TMAHT Poly(3-(6-trimethylammoniumhexyl)thiophene)
RMS Root mean square
R2R Roll-to-roll
SAM Self-assembled monolayer
SEM Scanning electron microscope
SCLC Space-charge-limited current
spiro-MeOTAD 2,2',7,7'-Tetrakis-(N,N-di-4-methoxyphenylamino)-9,9'-spirobifluorene
SRH Shockley-Read-Hall
TEM Transmission electron microscopy
TiOX Titanium oxide
TPC Transient photocurrent
TPV Transient photovoltage
UPS Ultraviolet photoelectron spectroscopy
VB Valence band
WF Work function
XPS X-ray photoelectron spectroscopy
ZnO Zinc oxide
ZMO Mg doped ZnO
1-D One-dimentional
x
Symbols
A device active area cm2
C capacitance μF/cm2 D diffusion coefficient cm2/s
EF Fermi level eV
ECNL the charge neutrality level of the interface states eV
FF fill factor %
I light intensity mW/cm2
Jo reverse saturation current mA/cm2
JSC short-circuit current mA/cm2
k Boltzmann constant J/K
L diffusion length nm
n0 diode ideality factor n/a
Pin input power mW/cm2
PCE power conversion efficiency %
q elementary charge C
R resistance Ω cm2
t0 lifetime s
T Temperature ℃
Vbi built-in potential V
VOC open-circuit voltage V
Φ work function eV
ε0 dielectric constant of vacuum n/a
θ contact angle o
Δ activation energy kJ/mol
xi
Contents
Acknowledgments....................................................................................................................... i
Abstract .................................................................................................................................... iii
Zusammenfassung...................................................................................................................... v
Abbreviations ........................................................................................................................... vii
Symbols...................................................................................................................................... x
Contents .................................................................................................................................... xi
Chapter 1 Introduction and Motivation ...................................................................................... 1
1.1 Introduction ...................................................................................................................... 2
1.2 Organic semiconductors ................................................................................................... 3
1.3 Evolution of organic solar cells ........................................................................................ 5
1.4 Aim of this thesis .............................................................................................................. 8
1.5 Outline of this thesis ......................................................................................................... 9
Chapter 2 Theory ..................................................................................................................... 11
2.1 Working principle of organic solar cells ......................................................................... 12
2.1.1 Photon absorption and exciton generation ............................................................... 12
2.1.2 Exciton diffusion & dissociation ............................................................................. 13
2.1.3 Electron-hole pair separation ................................................................................... 14
2.1.4 Charge transport ....................................................................................................... 14
2.1.5 Charge collection ..................................................................................................... 15
2.2 Characterization of organic solar cells ........................................................................... 15
2.2.1 Light current density-voltage (j-V) characteristics .................................................. 16
2.2.2 Dark current density-voltage (j-V) characteristics ................................................... 17
2.3 Recombination processes ............................................................................................... 19
2.3.1 Geminate recombination .......................................................................................... 20
2.3.2 Non-geminate recombination................................................................................... 20
2.3.3 Surface recombination (Recombination at contacts) ............................................... 23
2.3.4 A distinction of typical recombination .................................................................... 24
2.4 The electrode–organic semiconductor interface ............................................................. 26
xii
2.4.1 Standard charge neutrality level (CNL) models ...................................................... 26
2.4.2 Integer charge transfer (ICT) models ....................................................................... 28
Chapter 3 State of the Art: Role and Function of Interfaces in Organic Solar Cells ............... 33
3.1 Function of interfacial layers .......................................................................................... 34
3.2 Surface recombinations from non-ideal contacts ........................................................... 35
3.2.1 an S-kink in the j-V curves ....................................................................................... 35
3.2.2 The VOC of OSCs is reduced by surface recombination .......................................... 37
3.3 Surface doping of semiconductors ................................................................................. 37
3.3.1 Doping of inorganic semiconductors ....................................................................... 37
3.3.2 Surface contact doping of organic semiconductors ................................................. 38
3.4 Tuning of charge injection / extraction barriers ............................................................. 40
3.4.1 Cathode modification for efficient electron injection .............................................. 40
3.4.2 Anode modification for efficient hole injection ....................................................... 51
Chapter 4 Materials and Methods ............................................................................................ 55
4.1 Materials ......................................................................................................................... 56
4.1.1 Active layer materials .............................................................................................. 56
4.1.2 Interface materials .................................................................................................... 56
4.2. Device preparation ........................................................................................................ 59
4.2.1. Deposition methods ................................................................................................ 59
4.2.2. Device Architectures and Sample Layout ............................................................... 61
4.2.3. Fabrication of organic devices ................................................................................ 62
4.3. Characterization ............................................................................................................. 63
Chapter 5 Overcoming Electrode-Induced Losses by Tailoring a Quasi-Ohmic Contact via Alkali Hydroxide Layers.......................................................................................................... 69
5.1 Solution-processed alkali hydroxide interlayers in organic solar cells .......................... 70
5.1.1 A solution-processed Ba(OH)2 modified AZO for inverted OSCs .......................... 70
5.1.2 Green solution-processing of the AOH layer for inverted OSCs ............................. 73
5.1.3 The effect of an AOH layer on the device contact barrier ....................................... 76
5.1.4 The impact of the AOH layer on light soaking ........................................................ 79
5.1.5 The environmental stability of devices with AOH layers ........................................ 81
5.2 The mechanism for the interfacial energetics of AOH-modified AZO cathode contacts .............................................................................................................................................. 83
5.2.1 The work function of MOx cathode was measured using the Kelvin Probe Method........................................................................................................................................... 83
xiii
5.2.2 The interfacial energetics of KOH-modified AZO cathode contacts with fullerene 84
5.2.3 Core level X-ray photoelectron spectra for the AZO and AZO/KOH substrate signals upon deposition of C60 .......................................................................................... 87
5.2.4 A semi-quantitative energy level alignment between the contacts and the C60 layer........................................................................................................................................... 89
5.3 Interface-induced Suppression of the Bulk Recombination in Organic Bulk Heterojunction Solar Cells ................................................................................................... 91
5.3.1 The electrical properties of the AOH-modified diode devices ................................ 91
5.3.2 The recombination kinetics by measuring the j-V characteristics of the AOH modified devices ............................................................................................................... 94
5.3.3 1D drift-diffusion modeling ..................................................................................... 96
5.4 Conclusions .................................................................................................................. 100
Chapter 6 Roll to Roll Compatible Fabrication of Inverted Organic Solar Cells .................. 103
6.1 A simple approach to printing inverted OSCs with a self-organized charge selective cathode interface layer ........................................................................................................ 104
6.1.1 The performance of solar cells with Phen-NaDPO as the additive........................ 104
6.1.2 The performance of solar cells with Phen-NaDPO as a thin interlayer ................. 108
6.1.3 The distribution of Phen-NaDPO in the “ternary” active layer ............................. 110
6.2 Deposition of annealing-free MoOx/PEG Hybrids in inverted OSCs by doctor-blading ............................................................................................................................................ 114
6.2.1 Comparison of s-MoOX, PEDOT:PSS and e-MoOX as HTLs in inverted OSCs .. 114
6.2.2 s-MoOX/PEG hybrids as HTL in inverted OSCs ................................................... 116
6.2.3 Surface coverage and the morphology of s-MoOX/PEG film on active layer ....... 117
6.2.4 The environmental stability of devices with s-MoOX/PEG ................................... 120
6.3 Conclusions .................................................................................................................. 122
Chapter 7 Interfacial Engineering for Perovskite Solar Cells ................................................ 123
7.1 A brief introduction of perovskite solar cells ............................................................... 124
7.2 Polyelectrolyte Interlayers for Perovskite PHJ Solar Cells .......................................... 126
7.2.1 The device structure of the perovskite PHJ solar cell ............................................ 126
7.2.2 Electron injection efficiency at PCBM/Ag interface ............................................. 128
7.2.3 The device performance of the perovskite PHJ solar cell...................................... 129
7.2.4 Surface coverage of the perovskite films ............................................................... 131
7.3 Perovskite Solar Cell with a Water-Free PEDOT ........................................................ 133
7.3.1 Optical properties of perovskite films with water-free PEDOT ............................ 133
7.3.2 Morphology of perovskite films with water-free PEDOT ..................................... 135
xiv
7.3.3 Photovoltaic properties of perovskite solar cells with water-free PEDOT ............ 136
7.3.3 Defect states at the interface to the perovskite ....................................................... 138
7.4 Conclusions .................................................................................................................. 140
Chapter 8 ................................................................................................................................ 141
Summary and Outlook ........................................................................................................... 141
8.1 Summary ...................................................................................................................... 142
8.2 Outlook ......................................................................................................................... 144
Appendix A ............................................................................................................................ 147
Bibliography .......................................................................................................................... 147
Appendix B ............................................................................................................................ 161
Publications and Presentations ............................................................................................... 161
1
Chapter 1
Introduction and Motivation
Although it is said that Organic Solar Cells (OSCs) cannot be competitive with the most widely
used inorganic solar cells, there has been a significant progress on the enhancement of the
efficiencies of organic devices in the last decade up to 12 %. In order to improve the
industrialization of OSCs, three important factors of cost, stability and performance are crucial
junctures for sustainable development. In this chapter, the motivation and frame of the thesis are
laid out.
2
1.1 Introduction
Solar energy is almost infinitely available and clean source of the future. Photovoltaic (PV)
technology, which converts solar energy from sunlight to electricity, is a renewable way to
relieve the increase in global energy demand and large amounts of carbon emissions, and can
be seen an alternative energy resource to decrease our reliance on fossil fuels, such as oil,
natural gas and coal, which can also irreparably pollute the natural environment and promote
global warming. Regarding the trend of fossil energy exhaustion and the demand of
improving environment, the need for renewable and clean energies is rising. Among these,
PV technology has a special place as it bears the largest potential for cost-effective and clean
source. According to estimates from the BDEW (Bundesverband der Energie- und
Wasserwirtschaft) in Germany, PV generated power amounted to 34.9 TW hand covered
approximately 6.8 percent of Germany’s net electricity consumption in 2014 (Figure1-1).
The total nominal power of PV installed in Germany rose to ca. 38.5 GW, distributed over
1.5 million power plants, by the end of 2014. PV is now, after biomass and wind power, the
third most important renewable energy source in terms of globally installed capacity. With
such growth of PV technologies, worldwide photovoltaic capacity is projected to be doubled
or even tripled to 430 GW in 2018. Solar Power Europe (formerly known as EPIA) also
estimates that photovoltaics will meet 10% to 15% of Europe's energy demand in 2030.1, 2
Traditional 1st generation PV devices are made from silicon wafer, and are usually flat-plate,
and generally the most efficient. However a large amount of silicon materials in processing
are disposed, which may cause secondary pollution and resource-wasting. To support such
growth, PV technologies would need to be developed with resource constraints in mind.
Second-generation PV devices are called thin-film solar cells because they are made from
amorphous silicon or non-silicon materials such as copper indium gallium (di)selenide and
cadmium telluride. Thin film solar cells use layers of semiconductor materials with only a
few micrometers thickness. However, partially due to the high production cost and related
environmental issues, conventional PV technology hasn’t successfully replaced grid-
electricity.
3
Figure 1-1. Percentage of renewable energy in net electricity consumption (final energy) in Germany from
2004-2014, Press Release Bundesverband der Energie- und Wasserwirtschaft. Reproduced from ref. 1.
Continuing advances to lower levelized cost of electricity (LCOE) of PV devices and avoid
secondary pollution in processing, while achieving promising highly efficiency and lifetime
is a task for the entire PV technology community to solve. During the past two decades, a
great deal of research has shifted from inorganic- to organic-based materials duo to
remarkable properties of organic solar cells (OSCs) such as mechanical flexibility, light-
weight, transparent, and rich in color and potentially being a low-cost, especially low-
temperature and easy to manufacture technology.3-6 Now organic light-emitting diodes
(OLEDs), which are similar to OSCs, are on the verge of widespread commercialization.
Dramatic improvements have also been achieved in the performance and stability of OSCs in
the past two decades, which are currently at levels worthy of consideration for applications in
low-cost modules using fully roll-to-roll (R2R) processes for the production of domestic
electricity.7, 8
1.2 Organic semiconductors
Organic semiconductors are a group of conjugated carbon-based compounds, and their
derivatives. Single molecules, oligomers, and organic polymers can be semi-conductive. The
optoelectronic properties of organic semiconductors arise from the π-conjugated system. In a
π-conjugated system, C-atoms can form sp2 hybridized where the sp2-orbitals form a triangle
4
within a plane and the pz-orbitals are in the plane perpendicular to it.9 A schematic illustration
of the sp2-hybridized orbitals is shown in Figure 1-2a. An orbital overlap of two sp2-orbitals
formed a σ-bond between two carbons. The resulting occupied bonding orbitals (σ) and
unoccupiedanti-bonding orbitals (σ∗) have a large energy difference, which is quite large and
well beyond the visible spectral range, leading to insulating properties.9 On the other hand,
the un-hybridized pz-orbitals overlap and form a π-bond, creating a delocalized electron
density. The filled π sate with highest energy is called the highest occupied molecular orbital
(HOMO) and the empty π* sate with lowest energy is called the lowest unoccupied molecular
orbital (LUMO).9 The energy difference between the bonding and anti-bonding π-orbitals is
much smaller HOMO and LUMO, leading to strong absorption in or near the visible spectral
range, thus contributing to semiconducting properties. 9
In organic semiconductor, two classes can be distinguished with (i) lower molecular weight
materials, e.g. small molecules and oligomers, and (ii) high molecular weight materials, e.g.
polymers. Small organic or organometallic molecules may be processed by either solution or
thermal deposition techniques. If carbon atoms form larger molecules, for example with
benzene rings as the basic unit, six pz-bonds become delocalized and form a π-system which
often has the extensions of the molecule (Figure 1-2b). Due to the close coupling of the π-
systems of the molecules in molecule crystals, they show in a purified form remarkable
transport properties, including band transport up to room temperature with mobilities of 1-10
cm2/Vs.10, 11 The polymers contain an extended π-conjugated organic backbone, giving rise to
their unique opto-electrical properties. The transport properties of semiconductor polymers
are usually determined by defects in the one-dimensional (1D)-chains or by hopping from
chains to chains. If a long chain of carbon atoms is formed, the pz-bonds become delocalized
along the chain and form a 1D electronic system (Figure 1-2c). The 1D-band has
considerable band width (on the scale of an eV), i.e., we have a 1D semiconductor with a
filled valence band originating from the HOMOs and an empty conduction band originating
from the LUMOs.9 Polymer semiconductors are usually deposited using solution processes,
like spin-coating, blading or printing.
5
For well-established inorganic semiconductors such as silicon, germanium, and GaAs,
intrinsic conductivities are in the range of about 10−8 to 10−2 Ω−1 cm−1,12 and the dielectric
constant is as large as 𝜀r = 10 so that coulomb effects between electrons and holes are
unimportant due to dielectric screening, and light absorption at room temperature creates free
electrons and holes. 12
Unlike traditional inorganic semiconductors, the molecules in organic materials are held by
van der Waals bonds. The dielectric constant is low, taking a value of about 𝜀r = 3.5.9, 13 This
implies that coulomb interactions are significant, so that any electron–hole pair created by
optical (or thermal) excitation is bound by a coulomb energy of about 0.5–1.0 eV.9, 13, 14 The
part conductivity of organic materials results from the dissociation of photo-generated
electron–hole pairs that are bound by their mutual coulomb attraction.
Figure 1-2. (a) a schematic of the sp2-hybridized orbitals with σ- andπ-bonds; (b) a orbital picture of
benzene as a simple example for small organic or organometallic molecules; (c) a orbital scheme of a
pol(acetylene) subunit. Reproduced from ref. 15.
1.3 Evolution of organic solar cells
c b
a
6
Organic semiconductors show great promise owing to their synthetic variability, low-
temperature processing, and exceptional microstructure in combination with unique
advantages such as being light-weight, flexible, transparent, and rich in color.9, 13, 16, 17
Although the main and key breakthrough in the development of organic solar cells can be
clearly found within the last three decades, their beginning can be dated back to the early 20th
century with the first discovery on the photoconductivity in solid anthracene.18, 19 In the
development of organic semiconductors, an important and remarkable milestone happened in
the late 1970s. Alan J. Heeger, Alan G. MacDiarmid and Hideki Shirakawa discovered that
the conductivity of conjugated polymers can be increased by more than 5 orders of magnitude
to the semiconductor level (~10 S cm-1) by oxidative (halogen) doping.18, 20, 21 The Nobel
Prize in Chemistry 2000 was awarded jointly to them “for the discovery and development of
conductive polymers”.
In the days which followed the first report of conductive polymers in 1977, new classes of
conjugated organic materials were synthesized for use in traditional semiconductor electronic
devices to replace inorganic semiconductor. In 1980s, first organic solar cell (OSC) with very
low (< 0.1%) power conversion efficiency (PCE) was fabricated, which consisted of a single
layer of organic conductive materials between two electrodes of different work functions.22 In
1986 a major breakthrough in OSCs was achieved by Tang via employing a donor-acceptor
planer heterojunction (bilayer) structure (Figure 1-3a), which was the first organic
heterojunction cell with an efficiency of 1%.23 The first report of photo-excited charge
transfer from a conjugated polymer to a fullerene (C60), was actually demonstrated by
Sariciftci et al. in 1992.24 Owing to C60 is fairly transparent and also has fair electron
conductance (10−4 Scm−1), this property makes fullerenes a good component in organic
heterojunction cells.24, 25 This heterojunction architecture consisting of two components
(polymer and fullerene) had different electron affinity (EA) of the electron donating
polymer and ionization energies (IE) of the accepting fullerene, therefore thermodynamic
driving forces are generated at the interface between the two components for separating the
coulombically bound electron-hole pairs, so called excitons, into free charge carriers.26 The
100-400 meV energy is required to split an electron–hole pair compared to a few meV for
7
crystalline inorganic semiconductors.27, 28 The thermal energy kT at standard conditions and
the electric field is not sufficient to dissociate these excitons. The materials are chosen to
make the differences large enough that these local electric fields are strong, which split
excitons much more efficiently than single layer photovoltaic cells.29 The layer with higher
electron affinity and ionization potential is the electron acceptor, and the other layer is the
electron donor. 26
A typical organic semiconducting material system which enables effective exciton splitting is
P3HT and PCBM. Owing to the diffusion length of excitons is typically in the range of 1–10
nm, a bilayer device of acceptor and donor is thus strongly limited to the diffusion length of
excitons.30 Therefore, the thickness of the photoactive layer is thin for a bilayer device. Yu et
al.25 developed a so called buck hetero-junction (BHJ) structure for photo-active layer to
improve their efficiencies, where donor and acceptor are either thermally-evaporated together
or mixed in the solution and processed for active layer (Figure 1-3b). The bulk-
heterojunction structure is not only a random mixture of two materials but consists of
homogeneously distributed acceptor and donor phases with appropriate dimensions to
effectively split excitons in the whole bulk. The subsequent efficiency improvements of BHJ
devices from 2.5% up to 12%31 during the last two decades strongly correlate with the
attempts to design of novel active materials32-34, morphological control33, 35, 36, interfacial
engineering37, 38, employing novel device architectures39, 40, meticulous device optimization41,
42, etc. Although the PCE of OSCs is still lower than that of commercial solar panels, owing
to their merit, as mentioned previously, with the technological development various kinds of
products based on organic photovoltaic will occupy an important position in the future PV
infrastructure.
b a
8
Figure 1-3. Two typical device configurations of OSCs where the active layers are sandwiched between two electrodes of different work functions: a) planar bilayer heterojunction with the donor and acceptor layers stacked on top of each other, b) bulk-heterojunction with a mixed blend of donor and acceptor materials.
1.4 Aim of this thesis
In an OSC, electrons and holes mentioned in previous section, resulting from exciton
dissociation at relevant donor–acceptor interfaces, are separated in the interfacial band
bending regions that also depend on the organic–electrode interfaces. Hence, the physical
process involved in charge injection, extraction, transfer and recombination organic–
electrode interfaces also play a large role in device performance and lifetime.43 In order to
improve the industrialization of OSCs, three important factors of cost, stability and
performance are crucial junctures for sustainable development. The goal of this work is to
achieve control of interfaces in order to optimize extraction and transport of charge carriers,
thus improving devices performance and lifetime along with effectively reduced production
costs for printing solar cells with large area processed interface layers.
In this thesis, i) we focus on the development and understanding of interface contact with
alkali hydroxides modification to improve devices performance and lifetime. Meanwhile, a
systematic investigation of the interfacial energetics with alkali hydroxides functionalized
contacts is carried out to further understand the electron extraction and recombination
mechanism. ii) We develop a new materials ink to print efficient OSCs using roll to roll
compatible processing techniques under ambient conditions, which effectively reduces
production costs and simplifies technological process. iii) The research on organic–metal
interfaces based on perovskite solar cells can provide new insight into the evolution of
interface design rules.
Although this thesis is focused on understanding and design of device interfaces to push
OSCs to commercial applications, the novel materials, approaches and promising mechanism
demonstrated in this thesis pave the way for roll-to-roll processing.
9
1.5 Outline of this thesis
Chapter 1 gives a background introduction about organic semiconductors, the evolution,
development, operational principles and characterization of OSCs. Specially, the relative
recombination mechanisms in a device are reviewed.
Chapter 2 is devoted to fundamentals including operational principles and characterization of
organic solar cells, the relative recombination mechanisms and interface models used in this
thesis.
Chapter 3 introduces the role and function of interfaces in state-of-the-art OSCs.
Correspondingly, some representative works on various interface treatment for improving
device performance are summarized and discussed. Meanwhile, the challenge and current
issues on interface layer are also presented.
Chapter 4 summarizes all the materials used in this thesis. The deposition techniques used in
this thesis for device fabrication and the parameters for fabrication of OSCs are discussed in
detail. At last, experimental methods used in this thesis are introduced.
In Chapter 5, the macroscopic effects of alkali hydroxides as interface modification layers are
studied and analyzed. From the microscopic mechanism, we investigate the interfacial
energetics of these electron collecting contacts with fullerenes using UPS measurements to
reveal energy level alignment with band bending. In addition, the recombination kinetics is
discussed, combined with a numerical device model based on a 1D drift–diffusion approach.
Based on these results and discussion, it is found that tailoring a quasi-ohmic contact to
fullerenes via solution-processed alkali hydroxides could overcome electrode induced losses
in OSCs.
Chapter 6 introduces the work on the design of interface layers by engineering to fabricate
efficient devices using roll-to-roll compatible processing techniques. a), a simple one-step
solution processing based on self-organization of interfacial materials in BHJ blend leading to
a de facto bilayer with a bottom cathode interfacial layer and a top BHJ photoactive layer is
advantageous for high-volume roll-to-roll printing. b), a successful s-MoOX/PEG ink is
10
developed to replace evaporated MoOX to fabricate solution processed inverted OSCs via
doctor-blading in air. The part of this thesis is aiming to exploit fully solution processed
inverted devices using simplify technological process towards large scale roll-to-roll
production of fully printed solar cells.
Chapter 7 focuses on the organic–metal interfaces based on perovskite solar cells. We
fabricate fairly efficient conventional perovskite solar cells with an ultrathin polyelectrolyte
layer modified Ag electrode and inverted perovskite solar cells with water-free PEDOT to
replace spiro-MeOTAD. It is found that the density of sub bandgap states at the interface of
the perovskite/HTL seems to depend on the nature of interface materials. Based on these
results, the interface design rules for the organic solar cell can apply to the perovskite solar
cell technology.
Chapter 8 summarizes the main achievements as presented in this thesis. Limitations and
future challenges for OSCs are supplied as an outlook.
11
Chapter 2
Theory
In the following chapter, the theoretical aspects of the thesis will be explained. The fundamental
operational principles and characterization of organic solar cells are supplied. Specially, the
relative recombination mechanisms (geminate recombination and non-geminate recombination)
in a BHJ active layer and surface recombination are reviewed in detail. In the end, the current
interface models on energy level alignment are outlined, including charge neutrality level (CNL)
models and integer charge transfer (ICT) models.
12
2.1 Working principle of organic solar cells
Like all solar cells, the OSC converts photon energy into electric energy, by converting a flux
of photons (light) into a flux of charged particles (current). This fundamental process is made
possible by the combination of several types of layers, but most importantly is the
photoactive layers in organic solar cells. The fundamental operation principle of OSCs is
shown in Figure 2-1, using a bulk heterojunction structure OSC as an example. The process
of OSC operation can be summarized in brief. Details about all processes are given in the
text.44
Figure 2-1. An energetic scheme to describe the working principle of the BHJ organic solar cell.
1. An incident photon on a device, having an energy that exceeds the donor semiconductor band gap,
excites an electron from HOMO to LUMO above band gap, generating an electron-hole pair (1).
2. The electron-hole pair diffuses in the BHJ until it either recombines or reaches a donor-acceptor
interface, where it separates into electrons (blue) and holes (red) (2) over a built-in gradient in the
electrochemical potential of the solar cell.
3. Finally, the electron and hole then move through the donor and acceptor material phase to the
corresponding electrodes (3) and recombine after being put to work in an external circuit.
2.1.1 Photon absorption and exciton generation
When a photon with energy greater than gap energy (Eg) of the organic donor material is
absorbed, an electron from the HOMO will be excited into the LUMO, leaving behind an
13
electron vacancy, which is equivalent to a positively charged carrier and is referred to as a
hole. Together with the remaining hole on the HOMO the electron on the LUMO forms
instantly an excited singlet exciton which exhibits a high binding energy and is therefore
called an exciton.
Ephoto = c∙h
λphoto ≥ Eg equation (2-1)
where λphoto is the wavelength of light, c is the speed of light and h is Planck’s constant.
Different from inorganic semiconductor materials, as mentioned previously, an absorbed
photon leads to the promotion of an electron from the valence band (VB) to the conduction
band (CB), resulting in a free electron in the conduction band and a free hole in the valence
band, due to their high dielectric constants. In contrast, for organic semiconductor, the
electron in this excited state quickly relaxes to the lowest vibronic state of S1 via non-
radiative decay, resulting in relative low dielectric constant.45 For example, a distance of 1
nm between hole and electron together with a value of about εr = 3 would estimate a rough
binding energy EB = 0.5 eV.46 To achieve efficient exciton dissociation in active layer, BHJ
donor-acceptor systems with phase separation in nanometer regimes has to be introduced.
2.1.2 Exciton diffusion & dissociation
Owing to the exciton diffusion length in organic semiconductors is typically in the range of
5-20 nm,22, 47-49 the exciton located on the donor, has a finite lifetime to reach the donor–
acceptor interface, and then the electrons transferred from the donor to the acceptor. If the
lifetimes of the excitons are very short, they recombine before they reach to the interface. An
intrinsic drawback of the disordered organic materials is that the exciton diffusion length is
typically quite short, within only typically around several tens of nanometers. For example in
P3HT, values of 2.6 to 8.5 nm have been reported.47, 50-52 Therefore, in order to obtain as
many excitons as possible reaching the interface, the active layer of OSCs is commonly kept
to be thin and the microstructure of BHJ has to be carefully controlled. As mentioned before,
due to the difference in chemical potential, exciton dissociation becomes energetically
14
favorable at the donor/acceptor interface, where electrons transfer from the donor to the
acceptor material occurs. The resulting state of the electron and hole is termed as a polaron
pair, which is bound by less strong Coulombic force.
2.1.3 Electron-hole pair separation
In the next step, in order to generate free electrons and holes which can contribute to the
photocurrent, the polaron pairs have to be further separated. Beside many models, the
separation process of the polaron pairs can be well described by the Braun-Onsager theory53-
55. The model is based on three transition rates: the decay rate of polaron pairs to the ground
state kf, the dissociation rate of polaron pairs to quasi free polarons kd and the reverse
transition of free polarons to polaron pairs kr.45 As shown in Figure 2-2, the dissociation
process is balanced by a reverse process. During the process, a fraction of the polaron pairs
can recombine geminately with a decay rate of kf. The geminate recombination is a
monomolecular process, with the recombination rate being proportional to the concentration
of the polaron pairs. The recombination in detail will be described below.
Figure 2-2. A schematic description of polaron pair dissociation according to the Braun–Onsager model is
given. The desired dissociation takes place with the dissociation rate kd. The free charge carriers can
recombine again with a recombination rate kr into the bound polaron pair state, which can once more decay
with the rate kf. Figure is modified from ref.45.
2.1.4 Charge transport
Upon separation of the exciton into an electron and a hole at a donor/acceptor interface, the
electron and hole will move in the acceptor and donor phases respectively, where electrons
are localized in the acceptor phase whereas the holes remain in the donor phase.
Subsequently, the free charge carriers have to be transported via percolated acceptor and
15
donor pathways towards the corresponding electrodes.42 The transport is driven by the built-
in electric field in OSCs, which is induced by the difference in the intrinsic work functions of
the two electrodes before contact. Due to low mobilities of major organic materials, this step
can be a limiting step where a large number of carriers can be trapped. In this process,
nongeminate recombination of the free charge carriers would occur, which is contrast to the
geminate recombination during polaron pair dissociation.42, 56 The relative analysis of both
geminate and non-geminate recombination loss mechanisms will discussed in detail in
section 2.3.
2.1.5 Charge collection
Following charge transport, electrons and holes are collected at the respective electrodes to
generate photocurrent. Instead the current flow is depended on the use of electrodes having
sufficiently different work functions. To efficiently extract charge carriers, the contacts have
to be chosen carefully. In the case of the BHJ OSCs the phase orientations are random and
percolation paths of pure donor or acceptor material can connect the two electrodes.42 As
shown in Figure 1-4, the holes will travel to the high work function anode and the electrons
to the low work function cathode. If charge carriers are extracted at the wrong electrode by a
high energy barrier, meaning electrons (holes) at the anode (cathode), the tremendous
recombination will happen at the wrong electrode. Selective electrodes can be experimentally
achieved by (electron / hole) blocking layers between the active material and the respective
electrode. If the charge carriers can only be extracted with a lower rate than they are
generated, they will accumulate at the contact and form a space charge. This would result in
an s-shaped current–voltage (j-V) behavior and insufficient solar cell efficiencies.57
Appropriate modification of the electrodes by interface engineering having desired energy
level alignment with the LUMO and HOMO of the acceptor and donors is a critical
importance point to control interface barriers for charge harvesting.
2.2 Characterization of organic solar cells
16
One of typical characterization methods of organic solar cells is to measure their current
density-voltage (j-V) characteristics with and without illumination. A functional organic solar
cell exhibits a typical diode behavior in dark, i.e. the electric current is supposed to pass
through the device when a certain forward bias, higher than the threshold voltage, is applied.
When the applied forward bias is lower than the threshold voltage or a reverse bias is applied,
the current that passes through the device should be as low as possible. Under illumination,
charge carriers are generated and then extracted to the electrodes due to the built-in electric
field, creating a photocurrent in the external circuit. Photocurrent is generated within the solar
cell. In the ideal case the j–V characteristic of a solar cell is thus the additive of the dark
characteristic and photocurrent. The typical j–V characteristics of an organic solar cell with
and without illumination are represented in Figure 2-3, and the series resistance (Rs) and
shunt resistance (Rsh) from the equivalent device model of OSCs are shown in Figure 2-3c.
2.2.1 Light current density-voltage (j-V) characteristics
Several parameters can be extracted from the characteristic, namely, power conversion
efficiency (PCE), short-circuit current density (JSC), open-circuit voltage (VOC) and fill factor
(FF). MPP stands for the point at which maximal power can be extracted from a solar cell.
The FF is defined by
𝐹𝐹 = 𝐽𝑀𝑃𝑃×𝑉𝑀𝑃𝑃
𝐽𝑆𝐶×𝑉𝑂𝐶 equation (2-2)
The cell output PCE can be simply calculated using the j-V curve under illumination driving
JSC, VOC and FF values.
𝑃𝐶𝐸 =𝐽𝑆𝐶×𝑉𝑂𝐶×𝐹𝐹
𝑃𝑖𝑛 equation (2-3)
Where JSC (mA/cm2) is the current density under short-circuit conditions (zero applied
voltage); VOC (V) is the applied voltage under open-circuit conditions (zero current); fill
factor (FF) is defined as the area ratio of current to voltage product at maximum power point
(PMPP) (mW/cm2) divided by the product of JSC and VOC; Pin(mW/cm2) is the input power
during the j-V measurement.
17
2.2.2 Dark current density-voltage (j-V) characteristics
Studying the asymmetric dark j-V current response of an organic solar cell allows us to
evaluate device electrical parameters, including diode ideality factor, reverse saturation
current, and shunt and series resistances, which reflect material/contact properties. A typical
dark j-V curve on a log-linear scale is shown in Figure 2-3b. To determine these electrical
parameters, a modified ideal diode equation is often used58, 59, as shown in Equation
𝐽 = 𝐽0 {𝑒𝑥𝑝 [𝑞(𝑉−𝐽𝐴𝑅𝑠)
𝑛𝑘𝑇] − 1} + (𝑉 − 𝐽𝐴𝑅𝑠)/𝐴𝑅𝑠ℎ − 𝐽𝑝ℎ(𝑉) equation (2-4)
where J0,q, n, Rs,Rsh, k, T, and A is reverse saturation current density, elementary charge,
ideality factor, series resistance, shunt resistance, Boltzmann constant, absolute temperature,
and the active area of the device, respectively.
b
a
c
18
Figure 2-3. a) Typical j-V curve of an OSC under illumination, with important j-V parameters labeled. b)
Typical j-V curve of an OSC in dark on a log-linear scale, with the regions highlighted that are affected by
each electrical parameter in a modified ideal diode equation. c) Equivalent circuit diagram of a solar cell,
which consists of a current source JSC and a diode in parallel connection. Rs and Rsh account the realistic
behavior of a real solar cell.
𝐽0 is strongly dependent on the dark injection current from the acceptor HOMO to the donor
LUMO through thermionic emission, neglecting current leakage pathway in the band gap
and surface state at contacts. This energy barrier is typically larger than 2 eV, which is two to
three orders of magnitude higher than the thermal energy at room temperature (kT, ~25meV).
Therefore, the reverse saturation current is thermodynamically unfavorable and expected to
be very small (on the order of 10-5 to 10-6 mA/cm2 or less for a standard P3HT:PCBM
system). Mathematically, the magnitude of 𝐽0 can be estimated from the current intercept
from a linear fitting in the exponential range of the dark log j-V curve.
n is a qualitative value that describes the deviation of the diode behavior from its ideal
condition. The value of n determines the shape of the current inflection in the middle range of
positive bias, at around 0.3 to 0.6 V, where the electron injection from the cathode starts to
dominate the current response and FF can be affected. n is associated with the type of
majority carrier and charge recombination within the diode, as described by geminate (n = 1)
and non-geminate (n = 2) recombination.58, 60, 61 As mentioned previously, geminate
recombination is defined as the occurrence of electron and hole recombination before the
charges fully dissociate from one another, whereas for non-geminate recombination the
electron and hole recombine after the charges become free carriers. Geminate and non-
geminate recombination have been characterized with different lifetimes and spectral
signatures, as revealed by the light intensity dependent VOC and FF, and transient
photovoltage (TPV) and transient photocurrent (TPC) measurements. n values typically vary
between 1 and 2, and reflect competition between geminate and non-geminate recombination.
Higher orders (n >2) have been reported although their physical meanings are under
discussion. Mathematically, the magnitude of n can be estimated by the slope (linear fitting)
in the exponential range of dark log j-V curve. 59
19
In a diode equivalent circuit of a solar cell, as shown in the Figure 2-3c, Rsh is in parallel
with the diode. Significant power losses caused by the presence of Rsh, are typically due to
manufacturing defects, rather than poor solar cell design. Rsh is desired to be large compared
to the resistive component of the diode, so that most of the current will pass through the diode.
Low Rsh causes PCE losses in devices by providing an alternate current path for the light-
generated current. Such a diversion reduces the amount of current flowing through the solar
cell junction and reduces the voltage from the device. The effect of Rsh is particularly severe
at low light levels, since there will be less light-generated current. In addition, at lower
voltages where the effective resistance of the solar cell is high, the impact of a resistance in
parallel is large. The determination of Rsh in OSCs is one way to evaluate the quality of active
layer and interlayer films as well as the compatibility of interface between the two contacts,
since the presence of current leakage pathways (pinholes) allows for more non-geminate
recombination and shorting the devices. Ideally, the Rsh should be infinite and a low Rsh can
lead to a poor rectifying behavior of the diode.
In the equivalent circuit, Rs is in series with the diode. Rs is desired to be small compared to
the resistive component of a diode, so that the current magnitude in positive bias becomes
large. Series resistance in a solar cell has three causes: firstly, the electrode sheet resistance;
secondly, the contact resistance between the interlayer and active layer; and finally the bulk
resistance of interlayer and active-layer components. The main impact of series resistance is
to reduce the fill factor, although excessively high values may also reduce the short-circuit
current.
2.3 Recombination processes
The recombination of electrons and holes is a major loss mechanism in OSCs that has a
severe impact on the device performance of an organic solar cell. On the one hand, the
geminate recombination process of excitons or still Coulomb bound electron–hole pairs in the
charge transfer state is considered. On the other hand, the non-geminate charge carrier
recombination is defined by electron–hole annihilation of quasi free polarons toward the
electrodes. Thus, non-geminate recombination occurs after a polaron pair dissociation and
20
can therefore be distinguished from geminate processes by different decay dynamics and
timescale. For geminate recombination, electrons and holes of the polaron pairs are from the
same origin (a monomolecular process), while for non-geminate recombination, the free
electrons and holes are independent and recombine by a second order process (a bimolecular
process).30, 42, 56
2.3.1 Geminate recombination
As mentioned previously, light absorption in BHJ OSCs leads to the formation of a tightly
bound singlet exciton that subsequently dissociate at the donor-acceptor interface via charge
transfer. After dissociation, the bound interfacial charge transfer states are formed, due to a
geminate pair of a hole on the donor and an electron on the acceptor. Such bound charge pairs
then decay to the ground state via geminate recombination.62-65 Polaron pairs that do not split
up will recombine geminately; geminate recombination is the reunion of charge particles
originating from the same molecule. Typically, many reports described that this geminate
recombination at a D/A interface occurs after formation of a charge transfer state where the
electron resides in the acceptor and the hole in the donor.63, 66
Geminate recombination is considered a single body, monomolecular process. The number of
geminate pairs that are able to recombine geminately scales linearly with the number of
absorbed photons.63 However, it is important to note that at light intensity comparable to one
sun the probability of any given geminate pair recombining is independent of the total density
of geminate pairs. The consequence of this is that the fraction of geminate pairs lost to
geminate recombination is the same at low light intensities as it is at higher light intensities.
Thus, the photocurrent of a system limited only by geminate recombination would scale
linearly with light intensity. At light intensities exceeding several suns, other recombination
processes such as exciton–exciton or exciton–charge annihilation may become significant. 67
2.3.2 Non-geminate recombination
Once a photo-generated charge carrier successfully separates from its geminate counter
charge, the internal electric field in the device drives it toward the electrodes. To generate a
photocurrent, unbound charges, formed after exciton dissociation and the resulting free
21
charge carriers, have to diffuse through the photoactive layer to anode and cathode.62-65
Regardless of the mechanism, exciton dissociation leads to the formation of mostly free
charges and greatly reduced geminate recombination. Thus, in such devices, geminate
recombination is not a significant factor, and non-geminate recombination (bimolecular
recombination) is the main recombination channel.
The term non-geminate recombination encompasses all the recombination of any free charge
carriers. With different reaction orders non-geminate recombination can be of different types,
for example; trap-assisted recombination (monomolecular, first order, Figure 2-4b),
bimolecular recombination (second order, Figure 2-4a), and auger recombination
(trimolecular, third order, Figure 2-4c) mono-molecular recombination (first order), bi-
molecular recombination (second order) or tri-molecular recombination (third order).63 The
rate equation for photo-generated charges with non-geminate recombination of order α is
given by:
𝑑𝑛
𝑑𝑡= G − R = G − r𝑛𝑎 equation (2-5)
where n, t, G, R and r are density of charge carriers, time, charge generation, charge
recombination and the recombination coefficient, respectively.
Figure 2-4. Non-geminate recombination mechanisms: a) bimolecular recombination, b) trap-assisted
(Shockley–Read–Hall) recombination, c) Auger recombination. Figure is modified from ref.63.
2.3.2.1 Bimolecular recombination
In BHJ OSCs the most commonly observed non-geminate recombination mechanism is
usually direct (band-to-band) bimolecular recombination, which is a second order process.
a b c
22
Bimolecular recombination is typically observed to be of Langevin type due to the fact that
the mean free path of the charge carriers is smaller than the Coulomb capture radius.
Langevin recombination is described by the Langevin expression68 following the relation,
𝑅𝐿 =𝑞
𝜀(𝜇𝑛 + 𝜇𝑝)(𝑛𝑝 − 𝑛𝑖
2) equation (2-6)
where q, ɛ, 𝜇𝑛 , 𝜇𝑝 are the elementary charge, the dielectric constant, the mobility of the
electrons through the LUMO of the acceptor and the mobility of the holes through the
HOMO of the donor, respectively. n and p represent the electron and hole charge density
respectively and 𝑛𝑖 is the intrinsic carrier concentration.
As mentioned above, geminate recombination losses are not very significant in the current
generation of devices. This means that bimolecular recombination effectively controls j-V
characteristics of OSCs.69 From short-circuit to open-circuit conditions of j-V characteristics,
the charge density within the active layer increases, leading to an increase in bimolecular
recombination and a decrease in the fill factor of the device. Durrant et al.70 demonstrated
that bimolecular recombination limits the open-circuit voltage (VOC) of the device by limiting
the charge density within the photoactive layer. From their transient experiment and analysis,
they demonstrated that VOC follows the following empirical expression62:
𝑉𝑂𝐶 =1
𝑒(𝐼𝐸𝑑 − 𝐸𝐴𝑎) −
𝜂𝑘𝑇
𝑒𝑙𝑛 (
𝐽𝐵𝑅
𝐽𝑆𝐶) equation (2-7)
where IEd, EAa, η, JBR, and JSC are the ionization potential of the donor, electron affinity of
the acceptor, diode ideality factor, bimolecular recombination current, and short-circuit
current, respectively.
Koster et al.71 described the dependence of the short circuit current density (JSC) on intensity
(I)
𝐽𝑆𝐶 = 𝛽(𝐼)𝛼 equation (2-8)
where β is a constant, α is the exponential factor and the α ranges typically from 0.85 to 1.71-
73 Thus far, the deviation from α = 1 has been conjectured to arise from a small loss of
carriers via bimolecular recombination, and the fraction of charge carriers that recombined (η)
and established a relationship, η = α−1 – 1.
23
3.3.2.2 Trap-assisted recombination
Trap-assisted recombination is a first order process which mainly occurs between trapped and
free charges through a localized energetic trap. A model for trap-assisted recombination in
inorganic semiconductors, known as Shockley–Read–Hall (SRH) recombination, has recently
been applied to organic semiconductor systems. In most BHJ solar cells, trap-assisted
recombination is governed by the diffusion of the free carrier (hole/electron) toward the
trapped carrier (electron/hole), similar to the Langevin recombination for free carriers.
Therefore, traps in the organic material or general impurities are the origin of most trap-
assisted recombination processes observed in organic solar cells. It should be considered the
mobility of the trapped charge is zero hence the recombination coefficient is dependent only
on the mobility of free charge. Here, 𝑅 = 𝐶𝑝𝑛𝑡𝑝 where 𝐶𝑝and 𝑛𝑡 are capture coefficient and
the density of electron traps, respectively, considering free holes recombining with trapped
electrons.74, 75
Although the known presence of traps in most donor materials used for OSCs, the majority of
efficient devices do not appear to be limited by trap-assisted recombination. However, the
intrinsic phase separation in high performing devices may help explain the apparently limited
role of trap-assisted recombination
3.3.2.3 Auger recombination
Auger recombination is the third possible non-geminate recombination process, which is a
third order process because it is a three-particle process (Figure 2-4c). In Auger recombination,
an electron in the LUMO and a hole in the HOMO recombine after which the energy is given
to a third carrier, an electron in the conduction band, which is then excited to a higher
energetic state. Auger recombination is most important at a high charge density caused by
heavy doping or high level injection under concentrated sunlight.76 However, unlike
inorganic semiconductor, auger recombination is generally not believed to be present in
organic solar cells.
2.3.3 Surface recombination (Recombination at contacts)
24
In the above discussions, charge-extracting contacts were found to be the most critical factor
determining device characteristics. Within bimolecular recombination models, there is a
competition between the extraction of charges at the electrode and recombination within the
active layer. To improve charge extraction and to suppress bimolecular recombination, have
mainly emerged as a critical success factor via the use of solvent additives and annealing, or
by increasing the mobility of the active layer components in order to minimize transit times
to the electrodes.77-79 These means were based on the understanding that bimolecular
recombination is mainly controlled by the properties of the active layer, such as morphology,
charge mobility and thickness of active layer. However, Recently Kumar et al. 80
demonstrated that the time charge carriers taken within an OPV device is controlled by the
interface between the active layer and contact, not entirely controlled by active layer
properties. When this interface of electrode is modified, via treating the surface of electrode
with conjugated polyelectrolytes (CPEs) or polar solvents, the rate of charger extraction at
short circuit increases by a factor of several and as a result device efficiency increases.
Without changing the active layer properties, modification of the electrode properties can
have significant effects on the extraction rate. This dramatic improvement in extraction
shows that the role of the active layer/contact interface is critical in enabling or suppressing
charge extraction.
Surface recombination is governed by the charge injection/extraction behavior of the contacts.
The presence of a minority carrier at a contact will govern the recombination. Typically the
surface recombination velocity of this process is assumed infinite, all minority carriers
recombine.81 Naturally the presence of minority carriers at a contact will reduce the device
performance because those charges will not be collected. Since surface recombination is a
bimolecular process the fastest carrier governs its rate.82 Consequently, increasing the
influence of surface recombination reduce the device performance.
2.3.4 A distinction of typical recombination
25
Literature typically distinguishes between two types of recombination in organic
semiconductors plus surface recombination. Direct recombination of carriers follows a
recombination rate12, 83 given by
𝑅𝑑𝑖𝑟 = 𝑘𝑟𝑒𝑐(𝑛 ∗ 𝑝 − 𝑛𝑖2) with 𝑘𝑟𝑒𝑐 = 𝑞(𝑚𝑛 + 𝑚𝑝)/𝜀𝑟𝜀0 equation (2-9)
where krec is a constant for a given semiconductor and ni2 is the familiar “intrinsic carrier
concentration”. Please note that equation (2-9) assumes a carrier density independent
mobility. Carrier recombination according to equation (2-9) results in an ideality factor nid,d
= q/kT d(Vexternal)/dln(Jdark) of unity. In the absence of other recombination processes,
equation 2-9 further predicts an illuminated ideality factor nid,l = q/kT dVOC/dln(f) of unity
and, consequently, Voc scaling linearly with an exponent = 1 as a function of the logarithm
of the light intensity. On the other hand, defect assisted or Shockley-Read-Hall (SRH)
recombination leads to an ideality factor of 2, as a single deep defect will capture minority
carriers with high yield.
𝑅𝑆𝑅𝐻 = (𝑛𝑝 − 𝑛𝑖2)/((𝑛 + 𝑝)𝑡𝑆𝑅𝐻) equation (2-10)
With tSRH being the minority carrier lifetime for holes and electrons in the presence of a single
level deep trap. Please note the interpretation of tSRH as minority carrier lifetime slightly
changes in the case of a shallower, long lived distribution of traps as commonly assumed for
organic semiconductors, while the equation still remains relevant.
Surface recombination was not too intensely discussed in literature as it is difficult to
uniquely identify. Kirchartz et al.84 suggested that the observation of nid,l < 1 is a unique sign
for surface recombination and thus provided excellent experimental access to study this
phenomena. The electron and hole contact currents Jsurface,(n,p) summarizing all contributions
from extraction, injection and recombination85 are given by:
𝐽𝑠𝑢𝑟𝑓𝑎𝑐𝑒,(𝑛,𝑝) = 𝑞𝑆(𝑛, 𝑝 − 𝑛0, 𝑝0) equation (2-11)
with surface recombination velocity S. Assuming the cathode at a distance d (device
thickness) away from the anode, the equilibrium concentration for electrons and holes in the
Boltzmann approximation is governed by the contact barriers Φn and Φp rather than by the
statistics of the bulk equilibrium quasi Fermi levels.
𝑛0(𝑑) = 𝑁𝑐 𝑒𝑥𝑝(−𝛷𝑛/𝑘𝐵𝑇) and 𝑝0(𝑑) = 𝑁𝑣 𝑒𝑥𝑝((−𝐸𝑔 + 𝛷𝑛)/𝑘𝐵𝑇) equation (2-12)
26
where Nc and Nv are the effective densities of states (DOS) for LUMO and HOMO,
respectively. The band gap of the organic layer in contact with the electrode is named as Eg.
It is important to rephrase the mechanisms influencing the carrier density at the interfaces. In
the presence of a significantly large contact barriers of several 𝑘𝐵𝑇 we expect the total carrier
density n = 𝑛0 + 𝐷𝑛 and p = 𝑝0 + 𝐷𝑝 at the contacts to be determined by the excess carrier
densities 𝐷𝑛 and 𝐷𝑝. Combined with the identity 𝑛𝑖2=𝑛0*𝑝0 this will simplify equation 1 to
𝑅𝑑𝑖𝑟 ≈ 𝑘𝑟𝑒𝑐(𝐷𝑛 ∗ 𝐷𝑝) equation (2-13)
It is important to highlight that equation (2-13) suggests that an interface barrier for carrier
extraction can induce enhanced bimolecular recombination as long as a significant minority
carrier concentration is photo-generated at the majority extracting contact. Varying the bias
from Vbi to 0 or to negative values may reduce the recombination rate as photo-generated
minority carriers are extracted to their majority contact.
2.4 The electrode–organic semiconductor interface
2.4.1 Standard charge neutrality level (CNL) models
Standard electrode–organic semiconductor interface leads to the following equation for the
charge carrier injection barrier12, 86:
ΦBn = S (ΦM – EA) + (1– S) ECNL equation (2-14)
where the work function of the electrode (ΦM), and the ionization energy (IE) and electron,
and affinity (EA) of the organic semiconductor are defined as the energy difference between
vacuum and Fermi levels, and vacuum level and HOMO or LUMO, respectively, ECNL is the
charge neutrality level of the interface states. If the interface EF is above (below) ECNL, the
net charge in the interface states is negative (positive).
To discuss the different functions of interfacial layers, an understanding of contact formation
between the photoactive layer and the contact of electrode materials is required. Interfaces
between organic semiconductors and metal or conducting electrodes are the gateway to
injection of charge carriers into (or extraction of charge carriers from) devices. As shown in
27
the energy level diagram in Figure 2-5a, the position of the molecular levels with respect to
the electrode Fermi level (EF) follows from vacuum level alignment (EF equal to the ECNL),
the interface electronic structure of a typical metal- (organic) semiconductor interface
depends in first approximation on the electrode contact work function (ΦM) and the organic
semiconductor electron affinity (EA) and ionization energy (IE). As a non-interactive metal–
semiconductor interface example, that is the Schottky–Mott limit12, as electrons in the metal
occupy states up to the Fermi edge, the distance between the Fermi level (EF) and the organic
semiconductor IE (EA) describes the injection barrier for holes ΦBh (electrons, ΦBn) under the
assumption of vacuum level alignment.
Electron injection barrier: ΦBn = ΦM – EA; equation (2-15)
Hole injection barrier: ΦBp = IE – ΦM. equation (2-16)
A useful interface parameter to describe the dependence of the barrier on the electrode is the
interface parameter S defined as the rate of change of the electron barrier ΦBn with the metal
work function ΦM : S = dΦBn /d ΦM. In the Schottky–Mott limit, the parameter S is equal to 1. 12, 87
Figure 2-5. a) Electronic structure of a typical electrode–(organic) semiconductor interface in the
Schottky–Mott limit with vacuum level alignment across the interface; b) energy diagram of an interface
affected by a density of interface gap states, leading to an interface dipole Δ, showing relevant energy
levels on both sides, the electrode work function (ΦM), the electron affinity (EA) and ionization energy (IE)
a b
28
of the (organic) semiconductor, the interface density of gap states and charge neutrality level ECNL, and the
electron and hole injection barriers (ΦBn and ΦBp). Interface states below the Fermi level (red bar) are filled.
Figure is adapted from ref.88.
Moving away from the common type of non-interactive interfaces, an EF equal to the ECNL
represents a net surface charge of zero. However, if there is a difference between the EF and
the ECNL, the reverse surface charge is then located on the metal and a dipole layer is formed.
At an intimate clean electrode–organic semiconductor interface, for example metal as
electrodes, the proximity to the continuum of metallic states and penetration of the metal
wave function in the first few molecular layers (1~5 Å) of the organic semiconductor render
this region of the material metallic, resulting in a formation of metal-induced gap states at the
surface of organic semiconductors.89-91 Other cases involving defects at metal-organic
interfaces can be of structural or chemical nature, for example chemical reaction that breaks
chemical bonding in the molecule and introduces a new density of states in the semiconductor
gap. The charge neutrality level (CNL) of this density of state tends to align with the metal
Fermi level in order to minimize the charge transfer between metal and induced gap states,
and thus reduces the Fermi level movement in the process, leading to S < 1.88 As shown in
Figure 2-5b the reverse charge is then located on the metal and a charge transfer induced
dipole layer, Δ, between the electrode and organic semiconductor is formed.
Another mechanism, called “push-back” or “pillow” effect, always introduces a dipole that
lowers the vacuum level from the electrode into the organic film, when an atom or a molecule
is deposited on an electrode surface. This dipole can reach up to 0.5–1 eV, e.g. Au, upon
deposition of conjugated molecules, depending on the initial electronic structure of the gold.91
It's worth noting that the reduction in work function associated with this effect should not be
viewed as resulted from the formation of an interface dipole, as is with the reverse surface
charge transfer, but simply as the modification (or suppression) of the ‘‘external’’ surface
component of the work function of the metal.
2.4.2 Integer charge transfer (ICT) models
29
In principle, the CNL theory holds for metal-organic semiconductor interfaces as well.
However, unlike conventional inorganic semiconductors, organic materials are found in a low
screening and in the presence of significant structural disorder. On one hand, a low screening
within organic materials results in that the gap close to the metal becomes smaller due to the
much higher screening capability of the metal. On the other hand, structural disorder at the
interface creates energetic disorder with a variance that is typically several tenths of eV. Thus,
structural disorder effectively generates a deep occupation of intragap/tail states, which cause
a pinning of EF to values within the energy gap away from the charge transport level (LUMO
or HOMO).92, 93 A second theory to explain this pinning of EF is the integer charge transfer
(ICT) model, as described in the reviews by Fahlman et al.94 ICT model assumes self-
localized states called polarons (single charge) or bipolarons (double charge) for organic
semiconductors at the metal-organic interface. The self-localized states are further relaxed
states due to the screening from the metal substrate. Therefore, a charge transfer to the
electrode is possible with the work function of the metal reaching the polaronic level,
consequently, an interface dipole is created.
There are two polaronic levels in this model94: (i) the energy of a negative integer charge-
transfer state, EICT– which is defined as the energy gained when one electron is added to the
organic semiconductor to reach the fully relaxed state, that is, both electronic and geometrical
relaxation are included as well as screening from the substrate; (ii) the energy of a positive
integer charge-transfer state, EICT+ which is defined as the energy required to take away one
electron from the organic semiconductor to reach the fully relaxed state, that is, both
electronic and geometrical relaxation are included as well as screening from the substrate. In
the ICT model, three separate regimes are identified to control the interface energy level
alignment behavior (Figure 2-6). The Schottky-Mott regime, characterized by vacuum level
alignment has been described above and is depicted as case 2 in Figure 2-6b. From this, an
abrupt transition is made into a region characterized by creation of a positive (or negative)
interface dipole that scales with a decrease (or increase) in substrate work function, case 1
and 3. The transitions occurs when the substrate work function crosses the threshold values of
the positive or negative integer charge transfer energies EICT+ and EICT– as demonstrated in
30
Figure 2-6a where the charge transfer process and interface dipole formation upon contact
between the electrode and the organic materials is displayed.
1. Φsub < EICT– Transfer of electron to organic semiconductor
2. EICT– < Φsub < EICT+ Vacuum level alignment
3. Φsub > EICT+ Transfer of electron from organic semiconductor
a
b
31
Figure 2-6. a) Dependence of the work function of organic molecule-coated substrates, Φorg/sub, on the
work function of bare substrates, Φsub. The solid line of the slope = 1 dependence expected for vacuum
level alignment, and the slope = 0 dependence expected for a Fermi-level pinned interface. b) A schematic
illustration of the different energy level alignment regimes possible at weakly interacting organic-electrode
interfaces when a π-conjugated organic molecule or polymer is absorbed on a substrate surface. In this
viewgraph three different substrates with a low, intermediate and a high work function are chosen for the
interaction with the molecule. Upon adsorption electron transfer is facilitated by non-resonant transport
processes. The illustration is based on a model previously published by Braun et.al in ref. 94.
The ICT model is valid for interfaces where direct coupling between the electrode and the
organic material is prevented because of contamination of hydrocarbons and/or oxide layers.
This is typical for solution processed organic devices with contacts between conductive
interfaces and organic semiconductors. At these interfaces, the injection barrier should follow
the Schottky-Mott rule, i.e. the injection barrier can be estimated by the difference between
electrode work function and the HOMO and LUMO level of the organic semiconductor. This
behavior has indeed been observed for many organic/metal interfaces prepared under ambient
conditions, for which the metal surfaces are “passivated” by hydrocarbon contamination
and/or oxides and the interface is governed by weak interaction. Recent studies showed that
the situation for these interfaces is different and that significant interface dipoles may be
formed.95, 96
The origin of the interface dipole is explained by spontaneous integer charge transfer across
the interface from/to the Fermi level of the substrate to/from the organic materials, creating a
fully relaxed charge transfer state. The process results in what is referred to as Fermi level
pinning due to the alignment that occurs between the EICT value and the Fermi level of the
substrate. The energy of the charge transfer state, EICT corresponds to a charged state that is
fully relaxed regarding electronic and geometric interaction with surroundings.
32
33
Chapter 3
State of the Art:
Role and Function of Interfaces
in Organic Solar Cells
This chapter presents the current research stage of role and function of interfaces in organic
solar cells. Representative works on interface treatment to improve the device performance and
to reduce the relative recombination are summarized and discussed. Meanwhile, this chapter
concludes with a brief discussion on the relative challenge and current issues on interface layer.
34
Organic solar cells are based on thin film architectures that are generally composed of
multiple interfaces between organic semiconductors (small molecules or polymers), inorganic
semiconductors (metal oxides, MOX), organic or inorganic dielectrics, and metals or other
types of conducting electrodes.88 After photon absorption and charge transfer, the photo-
generated excitons dissociate into free charge carriers, which must be transported cross
several interfaces along their paths and collected by the corresponding electrodes. The
process of charge injection/extraction dominates the performance of organic solar cells.
Therefore, the performance of devices is affected, and often dominated, by the electronic
structure and electrical behavior of the interfaces. Owing to the rapid development of organic
electronics over the past decade, interfacial engineering attracted much research interest from
the science and engineering research community.
3.1 Function of interfacial layers
As described in Section 2.2, organic semiconductors have a wide band gap and narrow
bandwidth compared to their inorganic counterparts.26 Therefore, the density of thermally
excited charge carriers in organic films is not sufficient to sustain high current density. In
these BHJ OSCs, light absorption leads to formation of a tightly bound singlet exciton that
subsequently dissociates at the donor-acceptor interface via charge transfer. After
dissociation, the bound interfacial charge transfer state is formed, due to a geminate pair of a
hole on the donor and an electron on the acceptor. To generate photocurrent, unbound
charges, formed after exciton dissociation and the resulting free charge carriers, have to
diffuse through the photoactive layer to the corresponding electrode, i.e. electrons by the
acceptor to the electron selective contact and holes by the donor to the hole selective contact.
In this thesis the hole selective contact is named as anode and the electron selective contact as
cathode. The physical process involved in charge injection, extraction, transfer and
recombination at an electrode–organic semiconductor interface also play a large role, and
thus control of the interfacial energy level alignment is the key technology required for high
performance and stable organic solar cells.
35
Depending on the material system used, the main functions and impacts of interface materials
are:
to tune the energetic barrier between the photoactive layer and the electrode
to form selective contacts for charge carriers
to determine the polarity of the inverted or normal architecture
to enhance the stability and lifetime of device
to influence the light distribution and propagation as an optical spacer
3.2 Surface recombinations from non-ideal contacts
3.2.1 An S-kink in the j-V curves
To ensure efficient charge extraction, as known extract the “right” type of charge carrier and
to block the “wrong” one, the ideal selectivity of the contacts is given when the electron
contact exchanges only electrons with the photoactive layer and the hole contact only holes.
If however the selectivity is not sufficient, a part of the photogenerated charge carriers will be
lost by surface recombination and the power output of the device will be reduced.45, 73 The
most prominent feature of imperfect contacts in OSCs is the appearance of an S-kink in the j-
V curves.57, 97 This surface recombination mechanism becomes more important when the
electrode is nonselective. So many factors have been considered as possible sources, such as
doping caused by oxygen exposure98, 99, diffusion and accumulation of organic impurities to
an electrode interface100, non-ideal vertical phase segregation101, energy barriers to
extraction102, or increased injection barriers103, 104.
Usually charged defects in the organic materials can give rise to undesigned doping of the
active layer. Some charged species, such as from oxygen, organic impurities, or metal atoms
from the electrodes, in the active layer can result in the S-shape degradation feature. Water
and molecular oxygen98, for example, have been found to diffuse into the active layer through
grains and pinholes in the electrodes, respectively. The presence of water and oxygen near the
electrodes might, therefore, give rise to a doping profile close to the contacts, then resulting
36
in free charge carriers recombination at the contacts. Thus generally speaking, the aged
devices under illumination give rise to a strongly reduced FF and JSC. Because any
nongeminate recombination is effectively eliminating charge carriers that could otherwise
contribute to the photocurrent. Wagenpfahl et al. calculated dependence of Voc, Jsc, and FF
on charge carrier mobility assuming reduced Langevin recombination (ζ = 0.1). The key to
limiting the effects of nongeminate recombination on Jsc and FF seems to be first and
foremost to establish efficient (see Figure 3-1) and balanced charge carrier mobilities.
Figure 3-1. Calculated dependence of Voc, Jsc, and FF on chare carrier mobility assuming reduced
Langevin recombination with different surface recombination velocity. Reproduced with permission from
ref.97, Copyright © 2010. IEEE Photonics Soc.
The S-shape degradation feature, symptomatic of low device lifetimes, appears to be linked
to the presence of low molecular weight contaminants, which may be trapped within samples
of high-molecular weight polyme. McGehee et al.100 showed that replacing the cathode
suppressed the S-shape feature in degraded. The inflection points attribute the S-shape
features to be a consequence of diffusion of organic impurities accumulating at and possibly
reacting with the cathode.
37
3.2.2 The VOC of OSCs is reduced by surface recombination
The VOC is often related to losses due to non-geminate recombination of separated charge
carriers at the D/A interface. This can be understood by considering that at the open circuit
condition the photo-generated current is entirely cancelled out by the nongeminate
recombination current. However, the VOC can also be reduced by recombination at the
“wrong” electrodes.84, 105, 106 From both the dark and light ideality factors differentially as a
function of voltage, Kirchartz et.al84 find that both the dark and light ideality factors are
sensitive to bulk recombination mechanisms at the internal donor/acceptor interface, and the
VOC of BHJ solar cells is limited by surface recombination, which leads to light ideality
factors decreasing below one at high voltage. In order to identify the impact of surface
recombination at the electrodes, a strong correlation between VOC, the intensity of
electroluminescence signal and the current in forward direction was set up by Reinhardt and
Würfel.107 Device with a lower VOC showed strongly reduced electroluminescence intensity
and reduced current in forward direction. This is ascribed to an enhanced rate of surface
recombination. Through interface modification, such as using a conjugated polyelectrolyte
(PFN) as a cathode interlayer, enhanced photovoltaic performance was achieved by adjusting
the PFN thickness, due to a reduction in recombination of holes at the cathode as the main
cause for improving VOC.105
3.3 Surface doping of semiconductors
3.3.1 Doping of inorganic semiconductors
The possibility of controlled doping was an extremely important factor for the application of
semiconductor technology because it allows taking control over conduction properties and
Fermi energy. Doping of semiconductors is usually achieved by intentionally introducing
impurities into an extremely pure intrinsic semiconductor. The impurities are dependent upon
the type of semiconductor and the properties that it needs to have for its intended purpose.
The dopants release an excess electron as a free negative charge carrier to the n-type doping
38
semiconductor. In contrast, the dopants consume one more electron for chemical bonding as a
free positive charge carrier to the p-type doping semiconductor.12, 108, 109
In all classical inorganic semiconductor devices, the dopants are impurity atoms introduced
into the bulk of semiconductors. However, doping at semiconductor surface can also be
achieved via an electron exchange between a semiconductor and dopants, which can control
the electronic structure and electrical behavior of these interfaces for charge carrier injection /
extraction.12, 108, 109 One difficulty in extending our knowledge from crystalline to amorphous
molecular semiconductors arises because the latter are not conventional semiconductors:
Charge transport is no longer by free propagation in extended states, but rather by hopping in
a manifold of localized states.42, 89, 110 It should be noted that the microscopic understanding
of the doping process of organic semiconductors is still very rudimentary.
3.3.2 Surface contact doping of organic semiconductors
In OSCs, the dopants are impurity atoms introduced into the bulk of organic semiconductors,
which often become recombination centers for free charge carriers. Although most of organic
semiconductors are usually undoped, the prospective creation of efficient organic-based
devices will require controlled and stable doping to adjust the Fermi levels to the transport
states.9, 42 The Ohmic losses can be lowered while charge carrier injection is supported. This
kind of p-type surface transfer doping has recently been demonstrated for fullerene and
fluorofullerene molecules serving as surface acceptors on hydrogen-terminated diamond.111-
113 The hydrogen termination leads to an exceptionally low ionization energy for the diamond;
the fullerenes were chosen for their high electron affinities. For C60F48111, the activation
energy Δ is even negative, and each molecule brought onto the diamond surface creates a
hole.
The electronic states at the surface are in some ways associated with molecular adsorbates,
which involved a complex electrochemical (redox) system. Molecular adsorbates acted as
surface donors / acceptors. These ions were usually created unintentionally when hydrogen-
terminated diamond was exposed to the atmosphere. In this electrochemical variant of surface
39
transfer doping, the redox potential of the hydrated ions effectively determines the effective
acceptor level of the electronic system.
Figure 3-2. Band-level diagrams of a model system in which the surface charge is a redox species located
exterior to the dipole. Starting conditions (a) and (d) depict no dipole, and the redox species equilibrated to
the Fermi level of the semiconductor resulting in a small amount of band bending. Introduction of a dipole,
but prior to any electron transfer will cause the redox species to be shifted up relative to the Fermi level for
positive dipoles (b) and down for negative dipoles (e). Redox species above the Fermi level will donate
electrons to the semiconductor resulting in more downward band bending (c), while those below the Fermi
level will accept electrons causing the bands to flatten (f). This model could be expanded by adding
counterions to achieve positive band bending. Reproduced with permission from ref.114, Copyright ©
2013, American Chemical Society.
Surface contact doping thus appears to be the mechanism behind a variety of surface
electronic phenomena. When controlled, it may become a valuable tool for engineering
micrometer- and nanometer-scale electronic devices. The balance between electrons localized
in acceptor states and free holes in the valence band is expressed by the constant Fermi
energy EF. The closer EF is to Ev, the higher the local density of holes. Under similarly
favorable conditions, surface transfer doping was very recently observed for silicon. As
described in Ref 114, a surface dipole will move all levels of a species located outside the
dipole layer relative to the Fermi level of the surface (see Figure 3-2). Thus a positive δ moves
outside levels toward the vacuum level, while a negative δ will move the outside levels
40
toward lower energy. As shown in Figure 3-2, for example, the Siδ+−Brδ− dipole decreasing
the distance from the core level to Fermi level, will also shift any energy levels at the Fermi
level to a position above the Fermi level (Figure 3-2b). Filled states above the Fermi level
will then donate their electrons to the semiconductor. This process will continue until all of
the filled donors lose their electrons or the energy level of the redox couple equilibrates with
the Fermi level (Figure 3-2c). In reverse, for a negative dipole, such as the Cδ−−Hδ+ bond on
a Si−CH3 surface, all of the energy levels of the adsorbed molecules should be shifted
downward. This shift will move any energy levels at the Fermi level to a position below the
Fermi level. Any unfilled acceptor would be filled, decreasing the net surface charge and
reducing the band bending (Figure 3-2f).
3.4 Tuning of charge injection / extraction barriers
Interface engineering, in the context of organic electronics, involves tailoring the energy-
level alignment between the electrode and the organic material, often via some sort of surface
treatment to the electrode material,115-119 or by incorporating a thin modification layer
between the electrode and the organic materials120-123. These interface engineering methods
could effectively reduce non-geminate recombination in the OSCs and interface losses at the
non-ideal contacts.
3.4.1 Cathode modification for efficient electron injection
3.4.1.1 Doping of n-type metal oxides as ETLs
Depending on the work function offsets and the type of doping, typical examples for the
electronic structure at interfaces, so-called ohmic or Schottky contacts, in OSCs
schematically show in Figure 3-3. Take a n-type semiconductor for example, a low metal
work function leads to an ohmic contact (a), whereas a high metal work function results in a
Schottky contact (b). Such a contact is characterized by the already defined electron injection
barrier and a built-in potential Vbi due to the work function difference. The common approach
of making a Schottky contact ohmic (c), i.e. not rectifying, consists of using a very high
doping concentration at the contacts, which decreases the width of the space charge region
41
and allows for a high tunneling probability for electrons in both directions. Therefore doping
of semiconductors is another effective way to improve the charge injection/extraction in
semiconductor devices.
Figure 3-3. Cathode-organic semiconductor interfaces: (a) an ohmic contact and (b) a Schottky contact
with electron injection barrier ϕn. (c) With a high doping concentration, cathode-organic interfaces allows
for a tunneling current through the barrier. Figure is adapted from ref.124.
From the energetics point of view, wide band gap ZnO and TiO2 are typically used as ETLs
since their conduction band (CB) energies (ca. 4 eV)12, 106, 108, 125 are close to the LUMOs of
common acceptor materials, such as PCBM and ICBA, while their VB positions are away
from the HOMOs of donors, forming an energetic barrier for hole transfer. Solution
processed ZnO and TiO2 films have been extensively employed as ETLs in OSCs during the
past decade. However, such a thin ETL makes the devices suffer from low mechanical
robustness and reduced protective properties against a chemical or physical reaction between
the active layer and the electrode.126 Doping is an efficient way to enhance the conductivity
of ZnO so as to overcome these disadvantages. For instance, for doped ZnO ETLs, the band
gap structure is a key parameter that could affect the device performance. The band gap and
energy level dependence on the doping content enables researchers to adjust the optical and
electrical properties of the doped ZnO ETLs for achieving better device performance.
The doping of ZnO with group-II elements (alkaline earth), such as magnesium (Mg),
strontium (Sr) and barium (Ba) or with group-III elements such as boron (B), gallium (Ga), or
42
indium (In), doped ZnO films have also been studied for inverted OSCs.127-130 The
improvements in device performance were attributed to the enhanced electron conductivity
and smooth surface morphology of the ZnO film and to the reduced electron trapping on the
surface associated with oxygen absorption. Yin et al.131 used Mg doped ZnO (ZMO) for
inverted OSCs achieving high performance. With an increase of the Mg content (x = 0.1–0.6),
the CB edge value of ZMO CBLs was finely adjusted in the range of ∼ −4.3 to −3.9 eV to
approach the LUMO of PC71BM. The corresponding band gap of ZMO films increased from
3.3 to 3.7 eV, which was confirmed by the gradual blue shift of the fundamental absorption
edge as the Mg content increases (Figure 3-4). It was also found that the increase in Mg
content led to a decrease in the work function using a scanning Kelvin probe technique.
Figure 3-4. a) Optical absorption spectra of Mg doped ZnO (ZMO) films. The inset shows an increase in
the bandgap of ZMO films with the increase of Mg content (x). b) Energy levels of the components in the
inverted PSCs with tunable wide-bandgaps ZMO from 3.2 ~ 3.7 eV. Reproduced with permission from ref.
131, Copyright © 2014 Wiley-VCH Verlag GmbH & Co. KGaA.
The performance of OSCs is critically dependent on the properties of each layer and the
interfacial contacts, especially the interface between electrode and organic materials. The
energy-level alignment trend has implications not only for contact resistance but also for
maximizing VOC. The maximum VOC that an OSC can achieve depends on the offset between
HOMO of the donor and LUMO of the acceptor. However, the maximum VOC can only be
achieved by appropriately tuning the energy-level alignment at the electrode/organic
interfaces. Surface treatment is an effective way to tailor the energy-level alignment between
the electrode and the organic material. Generally speaking, without interface engineering, the
a b
43
surface defects of ZnO HTL may lead to a poor interfacial contact with the organic
semiconductors, thus resulting in poor electron extraction, which is one of the main reasons
leading to a relatively high charge recombination and series resistance. Especially, in the case
of solution-processed ZnO prepared with widely used sol–gel processing or colloidal
nanoparticles, the low-temperature and solution processed ZnO with hydrocarbon ligands
frequently suffer from residual carbon impurities which lead to high densities of surface
defects, such as dangling bonds, surface groups and charged oxygen molecules, thus hamper
electron extraction and transfer.132, 133 Such surface defects or dangling bonds on the surface
of MOx nanocrystals promote the formation of defect states within the MOx bandgap, which
may act as charge trap sites or even recombination centers, decreasing the photocurrent and
power conversion efficiency as well as the device stability.133 A phenomenon commonly
observed from the j–V curve of inverted OSCs based on ZnO is the appearance of an
inflection point, i.e. the so-called “S-shape” kink feature. Schematics of various types of
energy level mismatch between the n-type MOX and the acceptor of the active layer (PCBM)
and their implication on the j-V characteristics are shown in Figure 3-5. (a) An injection
barrier is formed when the work function of the electrode is below the LUMO level of PCBM.
Such an injection barrier prevents electron injection at the MOX/PCBM interface. An
injection barrier primarily implies an increased series resistance, and, if contact formation is
following the ICT model, a potentially lower built-in potential Vbi (as compared to the
HOMOdonor−LUMOacceptor difference). A delayed injection, respectively a higher series
resistance or even an s-shaped j-V curve in the 1st quadrant in combination with a potentially
lower VOC is the consequence. (b) An extraction barrier implies that charges cannot be
extracted from the active layer to the metal oxide because the conduction band of the MeOX is
high above the LUMO of PCBM. Such an extraction barrier becomes visible as a strong bias
dependence of the photocurrent under reverse bias, i.e. in the 3rd / 4th quadrant. (c) The
combination of an extraction and an injection barrier results in an s-shaped j-V curve (1st
quadrant) in combination with bias dependent photocurrent (3rd / 4th quadrant).
Due to the surface defects or dangling bonds on the surface of the MOx, appropriate surface
engineering on MoOx is required through improving the interfacial electrical properties,
44
better aligning the energy-level and controlling the surface energy, which are highly desirable
and essential to achieve a high power conversion efficiency and good stability of OSCs.
Common efficient surface engineering methods include plasma reactions118, 134, UV-ozone117,
135 and controlled air exposure136-138, and incorporation of the surface modification interlayer,
such as the self-assembled monolayers (SAMs)120, 139, 140, conjugated polyelectrolytes
(CPEs)141-143, polyelectrolyte layers (PEs)144, 145, inorganic salts146-148, etc.
Figure 3-5.Schematic illustration of energy level mismatch between the n-type metal oxides (MOX) and
the acceptor of active layer (PCBM) and corresponding j-V curves in the organic solar cells. The red solid
line indicates dark j-V curves and the black solid line indicates illumination j-V curves. Type barrier: (a)
injection barrier; (b) extraction barrier; (c) extraction and injection barrier. Reproduced with permission
from ref.37, Copyright © 2014 The Royal Society of Chemistry.
3.4.1.2 UV illumination treatment of metal oxides
The current generation of inverted OSCs with n-type MOX semiconductors as an ETL
frequently displays so-called light soaking effects, which is best described by the appearance
of an S-shaped current−voltage curve near the open circuit point.149, 150 UV light (λ ≤ 400
nm) illumination treatment is a common way to improve the electrical characteristics and to
optimize optimum photovoltaic performance of inverted devices with MOX ETLs. The
45
origins of light-soaking effects in OSCs are multiple and were correlated to trap filling in the
metal oxide layer; chemisorbed oxygen on the metal oxide surface that creates carrier
depletion; space charge limited current effects in the metal oxide layer or to an energy barrier
at the metal oxide/organic active layer interface.149 The light induced charge generation in the
MOx is coming along with a filling of the defective trap sites and is enhancing conductivity
(photoconductivity) in parallel to reducing injection barriers. Light soaking therefore requires
exposure to light with UV light (λ ≤ 400 nm) and gradually enhances solar cell efficiency
over illumination time towards the value of maximum performance. Normally, the
conductivity of highly crystalline ZnO degrades under environmental conditions with time
due to the presence of traces of oxygen and water, while UV treatment can release this
oxygen and restore the original electronic properties. One prominent example is the
chemisorption of ambient oxygen, schematically illustrated in Figure 3-6a,b.151 Generally,
electrons in the conduction band of the ZnO layer are extracted by the chemisorption of
oxygen species (O2(g) + e– → O2–(ad)) (step (1)), and therefore an electron depletion layer
forms in the nearby regions of the surface, resulting in upward band bending (step (2)). The
band bending can result in the formation of a Schottky barrier for the electron injection and
extraction at the ZnO/organic semiconductor interface. However, upon UV light (λ ≤ 400
nm) illumination, electron–hole pairs are created in ZnO, and photo-generated holes are
likely to migrate to the surface and recombine with trapped electrons and restore the
chemisorption of oxygen species.151 Accordingly steps (3)–(5), the chemisorbed O2– species
become discharged and detach from the surface (O2–(ad) + h+ → O2(g)), resulting in a
reduction of the electron traps, and thus increases the concentration of mobile electrons. As
shown in Figure 3-6c, the inflection point problem (S-shaped) on the j–V curve of the
inverted devices with MOX could be solved by exposing the cells to UV illumination.
However, the devices underwent significant changes in their j-V curves, with the
disappearance of the kink shape after increasing UV exposure time. Lee et al.152 found that
upon UV irradiation, the electron density increased, and the trap sites were filled in the TiO2
layer of the device, resulting in “photoinduced rearrangement of the Fermi levels” at the
ITO/TiO2 interface and thereby narrowing the EB width. As EB width was significantly
reduced by UV irradiation, the accumulated electrons could be freely collected onto the ITO
46
electrode by tunneling through the barrier, leading to a remarkable enhancement in device
performances.
Figure 3-6. Schematic illustration of the oxygen adsorption–desorption process in n-type ZnO
nanostructures. In the dark (a), chemisorption of O2 molecules (1) via capture of free electrons from the
conduction band (2) leads to the formation of a depletion layer next to the surface. Under UV illumination
(b), electron–hole pairs are generated (3), and free holes likely recombine with the trapped electrons (4).
Consequently, the O2– species are neutralized and become detached from the surface (5). Reproduced with
permission from ref.151, Copyright © 2014 American Chemical Society. (c) The effect of light irradiation
at different wavelengths (i.e., red, green, blue, and UV light). (d) The effect of UV irradiation time.
Reproduced with permission from ref.152, Copyright © 2012 AIP Publishing LLC.
3.4.1.3 Fullerene based interlayer modification of MOX
In OSCs, PCBM is typically chosen as the electron acceptor, which needs to be matched by a
low work function cathode to suppress interfacial losses upon extraction of electrons. Due to
similar energy levels and structures as PCBMs, the development of water/alcohol soluble
fullerene derivatives as ETLs is thus in the focus of extensive research. The representative
a b
c d
47
examples of newly developed fullerene-based ETLs are presented in Figure 2-8a, where they
can be categorized into two parts: modified fullerene derivatives (C60-SAM) (left) and
fulleropyrrolidines (right).
Figure 3-7. (a) Molecular structures of some representative newly developed fullerene for ETLs.
Reproduced with permission from ref.153, Copyright © 2015 The Royal Society of Chemistry. (b)
Schematic illustrations of device structures and chemical structures of fullerene self-assembled monolayer
(C60-based SAM) modifiers containing different anchoring groups. Reproduced with permission from
ref.123, Copyright © 2010 American Chemical Society.
For the modified fullerene derivatives (C60-SAM) as efficient ETLs, Li et al. have been
developed a series of functional fullerenes including a PEG end-caped C60 (PEGN-C60)154 and
a –NH2 group containing C60 (DMAPA-C60)155. Meanwhile, Cao et al.156 described a
a
b
48
PC71BM-N EEL by functionalizing PC71BM with a tertiary amino group. With the work
function tuning, PC71BM-N EEL induces contact doping at the organic
semiconductor/electrode interface by forming intermediate amine: C70 complexes. Recently,
fulleropyrrolidines-based EELs (Figure 3-7a, right) have also been extensively studied.153 As
discussed in Section 2.2, considering the influence of interlayers on Fermi-level pinning of
electrodes to the EICT+/EICT− of organic semiconductor, the device polarity and electrode
selectivity can be tuned by surface treatment inserting proper interfacial materials. Jen et
al. introduced two excellent cathode-independent FPI-based ETLs, Mono-C60 and Bis-C60 for
achieving high-efficiency and stable OSCs.153 They found that the success of such ETLs lies
in self-doping and the WF-tuning capability induced by the ionic features. These C60-SAMs
can be easily processed to form a modification layer on the surface of ZnO through either a
solution immersion technique or a solution-based spin-coating method. Jen et al. developed a
series of C60-based SAM modifiers containing different anchoring groups (catechol,
carboxylic acid, and phosphonic acid), linkage location, and functionalization (Figure 3-
7b).123 They found that these C60-SAM functional layers could help in reducing the charge
recombination at the MOX/organic semiconductor interface and enhance the charge
selectivity, leading to high FF and JSC. Owning to their polar side-chains, these functional
anchoring groups attached to electrode surfaces and comprising dipolar moieties were useful
for adjusting charge injection barriers towards organic semiconductors. C60-SAM can form
interfacial dipoles at the organic semiconductor/electrode interface. These functional
anchoring groups assist in alignment of the energy levels at the organic
semiconductor/electrode interface and result in subtle dependence of VOCs on employed
interface contacts. In these C60-SAM, the proximity of molecular levels to EF assist in a
more efficient electron transfer through the SAM.
3.4.1.4 Non-fullerene based interlayer modification of MOX
Figure 3-8 shows the non-fullerene based EELs developed lately. We only discuss some
representative and the latest examples of the non-fullerene based EELs for OSCs applications
in this subsection. The structures of organic molecules can be easily modified towards
suitable energy levels and optical/electronic properties. For example, Kippelen et al. have
49
demonstrated that the WF modification of various electrodes can be achieved by simply
adhering an ultra-thin (<10 nm) layer of PEIE or PEI.157 They found that the WF modification
originates from two contributions: (i) the intrinsic molecular dipole (μMD) associated with the
neutral ethylamine group aligned along the vertical direction to the substrate, and (ii) the
surface interfacial dipole (μID) formed at the organic semiconductor/electrode interface.
Afterwards Woo et al. reported an appreciable PCE enhancement in inverted OSCs by
inserting a PEI layer modified ZnO.158 Owing to the intermolecular dipole moment, an ultra-
thin layer of insulating PEIE or PEI can induce an interface dipole pointing from the cathode
to the active layer in the device geometry, thereby effectively reducing the WF of the
cathodes and increasing the built-in potential of OSCs. Besides formation of interfacial
dipoles to result in different degrees of vacuum-level shifting, they also found that the PEI
layer could increase the surface roughness of ZnO, which noticeably decreases the series
resistance of the device.
Conjugated polyelectrolytes (CPEs), which are conjugated polymers with ionic
functionalities, have been reported to serve as a surface modification layer between the
electrode and the active layer for inverted OSCs, owing to their unique solution processed
properties and moreover their orthogonal solubility in the commonly used organic solvents
for photoactive materials.95, 159, 160 CPEs have also been employed as efficient interlayer
materials to modify MOX.160 In this type of materials, the π-conjugated main chains render
the delocalized electronic structures, while the polar pendant groups on their side chains can
improve the solubility in water and polar organic solvents. These materials can efficiently
adjust the WF of the cathodes by forming an interfacial dipole between the cathode and the
active layer, facilitating the charge transport. For instance, polyfluorene-based CPE
interlayers were incorporated into a PTB7-based PSC and a high PCE of 9.2% was
achieved.95 Bazan et al. demonstrated a homopolymer P3TMAHT in a polymer-based
conventional solar cell with a polythiophene-type electron extracting CPE interlayer
P3TMAHT which could improve the device performance from 5.0 to 6.1%.143
50
Figure 3-8. Molecular structures of some representative polymers and small-molecules for ETLs.
3.4.1.4 Alkali Metal Salts based interlayer modification of MOX
In the pioneering years, it has soon been found that interfacial losses and inefficient electron
collection can be overcome by inserting ultrathin buffer layers of low work function,
including calcium (Ca)161, barium (Ba)72 or lithium fluoride (LiF)162, between the extraction
electrode and the OSC active layer. Compared to low WF metal interlayers, solution-
processable metal salts greatly broaden the available choices of interface materials. As a
substitution for commonly used metal salt of LiF, versatile metal salts such as alkali
carbonates148, 163, cesium acetate164, cesium or sodium halides165, 166, cesium stearate167, and
disodium edentate168 have been almost equally effective as the cathode interfacial layer for
OSCs applications. It is also demonstrated that low-temperature annealing condition can also
51
be used for device fabrication as long as proper solution concentration is used. Thin alkali
carbonates modification layers can often modify the work function as depicted in Figure 3-9.
The work function variation follows one clear trend: the work function of ITO substrates
reduces with alkali carbonates layers from Li-containing to Cs-containing salts.148 These
materials favor the formation of interfacial dipoles between the active layer and the electrode,
and increase the photogenerated charge collection, resulting in better interfacial contacts and
reduced contact resistances. Nonetheless, for the most lively alkali metal (Li), the metal salts
are easy to react with oxygen and further to diffuse into the organic layer or electrodes, which
could induce degradation of OSCs.
Figure 3-9. (a) Evolution of secondary electron edge with different alkali carbonates layers on ITO, (b)
scheme for the formation of dipole layer on ITO and its effect on reducing the work function of ITO.
Reproduced with permission from ref.148, Copyright © 2008 Wiley-VCH Verlag GmbH & Co. KGaA.
3.4.2 Anode modification for efficient hole injection
A large work function difference between the anode and the cathode of an OSC is required to
create a strong built-in potential for assisting carrier transport and for generating a high open-
circuit voltage. While the ETLs have low WFs to facilitate electron collection, the hole-
transporting interface materials as HTLs should have high WFs to match the HOMO levels of
52
the donor materials in the BHJ active layer to facilitate hole-extraction.169 PEDOT:PSS is one
of the most commonly used charge selective contact layers in OSCs. However, the surface of
the PEDOT:PSS has a ~3 nm thick PSS-rich surface layer, because elongated ellipsoids of
PEDOT are surrounded by thin shells of PSS.170, 171 This morphology gives rise to its
anisotropic conductivity and refractive index. Some surface treatment modifications on
PEDOT:PSS have also been reported.172 For example, diindenoperylene (DIP) was recently
introduced as an exciton-blocking HTL to treat the surface of PEDOT:PSS for SubNc-based
devices. A spin-coated PEDOT:PSS with the low surface energy provided a smooth surface
for the deposition of a closed DIP layer, which contributed to the improved JSC and VOC. The
large improvement on the photocurrents was attributed to the reduced exciton quenching at
the anode.
Figure 3-10. Secondary photoemission onsets of (a) PEDOT:PSS, (b) ITO with (red) and without (blue)
methanol treatment. Reproduced with permission from ref.169, Copyright © 2014 Wiley-VCH Verlag
GmbH & Co. KGaA.
For one PEDOT:PSS formulation, Huang et al. discovered that during thermal annealing in
the conventional-architecture devices, surface layer of PSS could react with the donor P3HT
to form a doped interlayer between the PEDOT:PSS HTL and the BHJ active layer.173 The
acidic sulfonate moiety of PSS oxidizes P3HT to form a doped species, P3HT+. It was shown
that this doped interlayer increases Voc due to improvements in the charge selectivity of
PEDOT:PSS/BHJ interface.173 Kim et al. reported that an improvement in PCE for
P3HT:PCBM-based BHJ OSCs using photo-crosslinked P3HT (c-P3HT) as an electron
blocking/hole extraction layer.174 The incorporation of the c-P3HT layer lengthened the
a b
53
carrier lifetime and increased the hole mobility, suggesting that the c-P3HT layer not only
prevented the non-geminate recombination but also improved carrier transport. Surface
treatment of PEDOT:PSS with c-P3HT offers an effective means for preventing the non-
geminate recombination at the interfaces between the photoactive layer and the electrode in
OSCs. It is important to note that methanol treated anode can enhance the hole injection
properties for polymer solar cells. Friend et al.169 demonstrated that such methanol-induced
enhancement is caused by an in-situ modification of the anode surface, buried under a layer
of organic semiconductor (Figure 3-10). The performance enhancement originates from an
increase in the anode work function, thereby giving hole-selective ohmic contacts towards the
organic semiconductors. Moreover, the addition of Nafion to treat the surface of anode can
also increase the work function, thus improving electron blocking and reducing interfacial
barriers at the active layer/contact layer interface.175, 176
54
55
Chapter 4
Materials and Methods
This chapter presents the processing parameters and characterization methods of organic devices
in detail. All the materials used in this thesis are also summarized in this chapter.
56
4.1 Materials
The materials in this thesis were either commercially available, or were designed and
synthesized by our collaborators. All materials were used as received.
4.1.1 Active layer materials
The donor and acceptor materials applied in this thesis are listed in Table 4-1. The materials
were used as received without further purification. P3HT used for the study was provided by
Merck Chemicals Inc. (Lisicon SP001, region-regularity = 94.2%, M w = 65.5 kg/mol).
pDPP5T-2 (batch: GKS1-001) was provided by BASF. PC60BM was purchased from Solenne
B.V. The chemical structures of the donor and acceptor materials are illustrated in Figure 4-1.
Figure 4-1. Chemical structures of active layer materials (donors and acceptors) used in this thesis.
Table 4-1. Active layer materials (donors and acceptors) used in this thesis.
Material
abbreviation Provider
Product
number
Mw
[kg/mol]
Purity
[%l]
Donor P3HT Merck SP001 65.5 -
pDPP5T-2 BASF GKS1-001 47 -
Acceptor PC60BM Solenne 99.5
4.1.2 Interface materials
The n-type and p-type interface materials used in this thesis are listed in Table 4-2. The most
important materials for this thesis are metal oxides (MOX) (such as ZnO, AZO and MoOX).
All these materials are inorganic semiconductors with wide band gaps of around 3 eV (ZnO
57
≈ 3.3 eV, AZO ≈ 3.3 eV MoOX ≈ 3 eV). Through the large band gaps, these materials are
transparent for “visible” light and therefore suitable as window layers in solar cells.
N-type interface materials
The ZnO and AZO were used as received. ZnO nanoparticle suspensions solution (Lot#5039)
was provided by Nanograde Ltd. The crystalline ZnO nanoparticles with particle size of 10-
15 nm were dissolved in 2-propanol at a concentration at 2.7 wt. %.
Synthesis of Al doped zinc oxide (AZO): Zn(OAc)2•2H2O (20 g), Al(OH)(OAc)2(0.3 g) and
Zonyl FSO-100 (0.6 g) were mixed in demineralized water (200 mL). The mixture was
stirred for 2 h and filtered through a 0.45 micrometer filter to remove the insoluble material.
The functionality of AZO in OSCs was fully investigated previously.
P-type interface materials
Heraeus Clevios™ P VP AI 4083 used in this thesis is one of the most commonly used
PEDOT:PSS formulations. The chemical structure of PEDOT:PSS is illustrated in Figure 4-2.
For devices with a conventional architecture, PEDOT:PSS diluted in isopropyl at a volume
ratio of 1:3 was doctor bladed on top of the ITO-substrates. For devices with an inverted
architecture, PEDOT:PSS was diluted in isopropyl with a ratio of 1:5. The water-free PEDOT
- Clevios HTL Solar 3 (2.6% solids in toluene) and Clevios SEJ 272 (3.9% solids in anisole,
pH neutral) - used for perovskite solar cells were provided by Heraeus. When the dispersion
is spin-coated on glass and dried for 2 min at 200 °C the resulting film can show a
conductivity of 0.2 S cm–1 (Clevios HTL Solar 3) or 0.008 S cm–1 (Clevios SEJ 272).
Evaporated molybdenum oxide (e-MoOx) was purchased from Sigma-Aldrich, while the
solution-MoOX (s- MoOx) were used as received.
Synthesis of molybdenum oxide: molybdenum metal powder (0.1 g) was dispersed in 10 ml
ethanol with stirring for 30 minutes. Then 0.35 ml H2O2 (30%) solution was added into the
suspension solution. After 24 hours reaction, the molybdenum oxide solution turned from
grey to yellow and finally turned to blue, forming hydrogen molybdenum bronzes. Finally,
58
the suspension solution was dried in a vacuum chamber, and then was dispersed uniformly
into ethanol at a concentration of 1 mg ml-1.
Table 4-2. n-type and p-type interface materials used in this thesis.
Material
abbreviation Provider
Product
number Solvent
n-type
material
ZnO Nanograde Lot#5039 2-propanol
AZO i-MEET - Ethanol
TiO2 OneSun Inc. LT-TiO2 ink 1:1 water
and ethanol
p-type
material
PEDOT:PSS Heraeus AI4083 Water
water-free
PEDOT1 Heraeus
Clevios HTL
Solar 3 toluene
e-MoOX Sigma-Aldrich - -
s-MoOX i-MEET - 2-propanol
1 PSS free PEDOT material.
Interfacial modification layer
In this thesis, we focus on the understanding and development of n-type or p-type interfacial
modification materials (IMMs) for BHJ OSCs. Interfacial materials used in this thesis are
listed in Table 4-3, and the chemical structures of organic IMMs are illustrated in Figure 4-2.
P3TMAHT was provided by Prof. Ullrich Scherf from IfP, Bergische Universität Wuppertal.
P3TMAHT was diluted in methanol at a concentration of ca. 1 mg mL-1 (molecular weight of
P3TMAHT: 8000 g mol-1, estimated by GPC with PS calibration). Phen-NaDPO was
provided by Prof. Xuhui Zhu from South China University of Technology. PEIE, Ba(OH)2,
LiOH, NaOH and KOH were purchased from Sigma-Aldrich. All materials were used as
received.
PEG (Mw = 2000 g mol-1) as an interfacial modification materials to treat MoOX was
purchased from Sigma-Aldrich. The chemical structure of PEG is illustrated in Figure 4-2 as
well.
59
Table 4-3. interface modification materials used in this thesis.
Material
abbreviation Provider Solvent
Interface
modification
materials
Ba(OH)2 Sigma-Aldrich 2-methoxyethanol
LiOH Sigma-Aldrich 2-methoxyethanol
NaOH Sigma-Aldrich 2-methoxyethanol
KOH Sigma-Aldrich 2-methoxyethanol
P3TMAHT Uni-Wuppertal methanol
PEIE Sigma-Aldrich methanol
Phen-NaDPO SCUT methanol
PEG Sigma-Aldrich 2-propanol
Figure 4-2. Chemical structures of interfacial (modification) materials used in this thesis.
4.2. Device preparation
4.2.1. Deposition methods
Doctor Blading
Developing a stable process for large area coatings with highly reproducible efficiencies and
reliability remains a major challenge for the photovoltaic technology. We chose to focus on
doctor blade coating due to its simplicity and efficient material usage. Meanwhile doctor
blade coating can be employed in both sheet-to-sheet and roll-to-roll processing, which is
also relevant in main stream technologies like common in the textile, paper, photographic
60
film, printing, and ceramic industries where high uniformity over large areas is required.177
All solution processed layers of organic solar cells in this thesis have been deposited through
doctor blading in air. As shown in Figure 4-3a, the solution containing the material is applied
in the slit between the substrate and the blade. Then the blade moves with a defined speed
over the substrate to apply the wet film. After drying, the ~nm or m thickness film of the
applied material finds itself on the substrate. Normally, the thickness of the film is mainly
dependent on the speed of the blade and the solid content in the solution.
Figure 4-3. Depiction of solution deposition procedure s including a) doctor-blading and b) spin-coating.
Spin-coating
Spin-coating is another simple procedure used to deposit uniform thin films to flat substrates,
where it can be used to create uniform thin films with nanoscale thicknesses, as depicted in
Figure 4- 3b. A typical process involves depositing a small puddle of a fluid material onto
the center of a substrate and then spinning the substrate at high speed. For the spin-coating
procedure, the film thickness is mainly controlled by the concentration of the solution and the
rotation speed. In addition, for a constant spinning speed, the same dependency is obtained
for the spinning time. Spin coating procedure is the simplicity and relative ease with which a
process can be set up coupled with the thin and uniform coating that can be achieved.
Although efficiencies of OSCs have dramatically improved up to above 10% with spin-
coating, the spin-coating technology is still a batch process,178, 179 which is not compatible
with industrial-scale mass production technology, such as roll-to-roll production. Device
morphology is canonically optimized for lab scale spin coating which has little significance
a b
61
for a technical process, because it is limited to small areas and does not allow for an
independent variation of process parameters. Therefore, all organic solar cells in this thesis
were fabricated by doctor blading, which is indeed compatible to roll to roll processing under
ambient conditions. Solution processed interfacial layers of perovskite solar cells in this
thesis have been deposited through spin-coating in chapter 7.
4.2.2. Device Architectures and Sample Layout
Device Architectures
Normally, the two different architectures are defined by the position of the interface layers, as
depicted in Figure 4-4. For the inverted architecture, the holes are extracted through the hole
transport layer (HTL) on top of the active layer, while the electrons are extracted by the
electron transport layer (ETL) below the active layer. The HTL and ETL are an inverse
position for the normal architecture. Due to the stability of inverted derives, we mainly
focused on the understanding and development of inverted architecture devices.
Sample Layout
The sample layout is shown in Figure 4-4c. The solar cells were processed on glass
substrates (2.5×2.5 cm2) with pre-patterned bottom ITO electrode. Afterwards, all solution-
processed layers were deposited over the whole substrate area, and then structured before the
top metal electrode (Ag) deposition through a shadow mask via thermal evaporation. The
overlap of top and bottom electrodes defined the active area (10.4 mm 2). Each substrate had
6 devices.
a b
62
Figure 4-4. Typical layer stacks of OSCs: a) normal and b) inverted architecture and c) sample layout of
organic devices.
4.2.3. Fabrication of organic devices
Photoactive solution preparation and film deposition
pDPP5T-2:PC60BM with a ratio of 1:2 wt.% was dissolved in a mixed solvent of chloroform
and o-dichlorobenzene (9:1 vol.%). The film was deposited at 45 °C. P3HT:PC60BM with a
ratio of 1:1 wt. % was dissolved in chlorobenzene (CB), and the film was deposited at 60 °C,
annealed on a hot plate at 80 °C for 5 min.
Photovoltaic Device Fabrication
All the organic devices were fabricated using doctor-blading under ambient conditions with
the structures shown in Figure 4-4b. Pre-structured ITO coated glass substrates were
subsequently cleaned in acetone and isopropyl alcohol for 10 min each. After drying, the
substrates were coated with the 50 nm thick AZO layer or 50 nm thick ZnO layer via doctor
blading, respectively. Conversion of the precursor to AZO via hydrolysis was achieved by
heating the samples to 140 °C for 5 min. The AZO precursor was synthesized as optimized
and reported earlier based on zinc acetate and aluminum nitrate in ethanol. The ZnO layer
was annealed on a hot plate at 80 °C for 5 min. Subsequently, interfacial modification layers
doctor bladed on top of the kinds of metal oxides layer with a thickness of ≈ 10 nm and then
dried at 100 °C for 10 min, respectively. Then, the active layer, such as pDPP5T-2:PCBM or
P3HT:PCBM, was deposited on top of substrate. Finally 10 nm MoO3 and 100 nm Ag were
c
63
deposited sequentially under 5 × 10-6 Torr by thermal evaporation through a shadow mask to
form an active area of 10.4 mm2.
Electron-only Device Fabrication
All device substrates were prepared as for the photovoltaic substrates. The EEL and active
layer were prepared as described in “Photovoltaic Device Fabrication”. 15 nm Ca and 85 nm
Ag were deposited sequentially under 6 × 10-6 Torr by thermal evaporation.
4.3. Characterization
Current density-voltage (j-V)
j-V characteristics of solar cells were measured using a source measurement unit from BoTest.
Illumination was provided by a solar simulator (Oriel Sol 1A, from Newport) with AM1.5G
spectra at 100 mW cm-2 which was calibrated by a certified silicon solar cell. The light
intensity was modulated with a series of neutral color density filters, allowing changing the
intensity from 100 to 0.4 mW cm-2. All devices were tested in air.
External quantum efficiency (EQE)
EQE spectra were recorded with an Enli Technology (Taiwan) EQE measurement system
(QE-R), and the light intensity at each wavelength was calibrated with a standard single-
crystal Si photovoltaic cell. The EQE in section 4.1 was measured with a Cary 500 Scan UV–
Vis–NIR Spectrophotometer under mono-chromatic illumination, which was calibrated with
a mono-crystalline silicon diode.
Film thickness Measurement
The thickness of thin films was measured using a profilometer (Tencor). A profilometer scans
a surface with a small needle and registers height differences. The lateral resolution is the in
the μm range due to the diameter of the tip of the needle, but the height resolution is in the
nm range. The thickness of a film can be measured by removing a part of the film via
scratching and measuring the difference from surface of the film to the bottom of the gap.
Atomic Force Microscopy (AFM)
64
In this thesis, AFM was used to study the surface morphology and roughness of thin films.
And topographical measurements were performed with AFM (Veeco Model D3100) in
tapping mode.
Scanning electron microscope (SEM)
The cross-section as well as surface images were prepared using a focused ion beam (FEI
Helios NanoLab 660) operating at 30 kV and subsequentially imaged with the electron beam
of the same instrument using an acceleration voltage of 2 kV.
Photoluminescence spectroscopy (PL)
The room temperature up-conversion PL spectra were recorded by a spectrofluorometer
(JASCO, FP-8500), the PL excitation wavelength was set to 320 nm.
Transmission electron microscopy (TEM)
Samples for cross-sectional TEM were prepared using the in situ lift-out technique in a
focused ion beam (FIB) instrument, an FEI Helios NanoLab 660 DualBeam system. TEM
investigations were performed on an aberration-corrected FEI Titan Themis³ 300 TEM with a
high brightness field emission gun (X-FEG) operated at 200 kV equipped with a Super-X
detector for energy-dispersive X-ray spectroscopy (EDXS).
Surface energy
The contact angle and the surface of the bladed and annealed films of pristine materials and
their binary and “ternary” blends were measured using a contact angle instrument from
Dataphysics (model OCA20). With the values of the initial contact angles, surface energies
were calculated using the SCA20-U software and the Owens–Wendt and Kaelble method.
X-ray photoelectron spectroscopy (XPS)
PES measurements were taken using a Kratos Axis Ultra PES system with a base pressure of
ca. 9x10-10 Torr. XPS measurements were carried out using a monochromatic Al ka excitation
source (1486.3 eV) and high-resolution core level spectra were acquired at a pass energy of
20 eV. XPS spectra are processed using the Vision 2 software package provided by Kratos
Analytical.
65
Ultraviolet photoelectron spectroscopy (UPS)
For UPS measurements, a He discharge lamp was utilized as the excitation source (He I =
21.22 eV) and the sample was biased at -10.0 V to enhance the yield of electrons at the
secondary electron cutoff (SECO.) The UPS photoemission data was plotted on a binding
energy scale with respect to (w.r.t.) the Fermi level energy, which was calibrated to a sputter-
cleaned Au foil (i.e., the Fermi energy of the Au foil is set to zero on the x-axis). We assumed
that all substrates are in electronic equilibrium with the instrument, implying that the Fermi
energy of all samples is at the same kinetic energy of the clean Au foil. The work function (Φ)
of all samples was calculated by subtracting the spectral width (Fermi energy, EF, – the low
kinetic energy edge, LKE) from the He I photon energy (21.22 eV), according to:
Φ = 21.22- (EF - LKE) equation (4-1)
The work function of a clean Au foil was periodically verified to be 5.1 eV, ensuring that the
instrument is properly calibrated. Ionization potentials were determined by analyzing the first
photoemission feature with lower kinetic energy than EF. C60 (triply purified via vacuum train
sublimation) was thermally evaporated from a boron nitride Knudsen cell in UHV (base
pressure ca. 1 x 10-9 Torr) and the thickness was monitored with a freshly calibrated QCM
that was calibrated via AFM.
Both UPS and XPS measurements were made at multiple spots while maintaining a constant
z-position and takeoff angle (0°) relative to the surface normal.
Drift–diffusion modeling
Besides the simulation of the j–V characteristics of the devices with coarse phase separation,
we also modeled the two devices with active layer made from finely dispersed
polymer:fullerene mixture. The simulation of these devices can be done employing the
metal–insulator–metal (MIM) approach. This consists in treating the blend as one intrinsic
semiconductor material whose LUMO and HOMO energy levels correspond to the LUMO
and HOMO levels of the acceptor and of the donor material, respectively. Based on this
approach, a numerical simulation code was developed, the details of which can be found in
Ref (Table 4-4).
66
The MIM model used in this thesis contains drift and diffusion of charge carriers, and the
effect of space charge on the electric field in the device. It should be mentioned that the
influence of disorder on the transport of carriers was only taken into account through the
magnitude and field/temperature dependence of the mobility, leaving the equations
themselves unaltered. The resulting basic equations described transport through
semiconductors are solved self-consistently. Recombination was described as a bimolecular
process, with the rate given by Langevin. The rate of generation of bound electron-hole pairs
was assumed to be homogeneous throughout the device. As the devices considered in this
study were very thin (100 nm), the assumption of uniform generation of electron-hole pairs
did not give rise to serious inconsistencies
Table 4-4. Overview of the parameters used in the fit to the data shown in Figure 5-19.
Parameter Symbol Numerical value
absolute temperature T 295K
device length/thickness L 100nm
effective band gap Egapeff 1.10eV
Electron mobility μn 1.0×10-7 m2/Vs
hole mobility μp 2.0×10-8 m2/Vs
Generation rate of bound pairs Ge−h 6.25×1027 m-3s-1
Eff. Density of states Ncv 2.5×1025 m-3
Dielectric constant ⟨ε⟩ 3.0
Decay rate kf 1.0×106 s-1
e/h pair distance a 1.0nm
surface recombination of electrons at anode Sp -1.0×10-6 m/s
surface recombination of holes at cathode Sn -1.0×10-6 m/s
67
Space charge limited current (SCLC) method
Single carrier devices in chapter 5 were fabricated and the dark current-voltage characteristics
measured and analyzed in the space charge limited (SCL) regime. The reported mobility data
of pristine and blended films were average values over the eighteen devices of each sample
for a range of thicknesses. While we calibrated the hole and electron mobilities by fitting the
current-voltage curves, SCLC of devices is described by180:
𝐽𝑆𝐶𝐿 =9
8𝜀0𝜀𝑟𝜇
𝑉2
𝐿3 equation (4-2)
Where JSCL is the current density, ε0 is the permittivity of free-space, εr is the relative
dielectric constant of the active layer, μ is the charge carrier mobility, L is the thickness of the
tested layers and V in is the voltage dropped across the sample. The field dependent SCLC
expression yielded a reasonably good fit to the measured j-V curves of single-carrier devices.
68
69
Chapter 5
Overcoming Electrode-Induced Losses by
Tailoring a Quasi-Ohmic Contact via
Alkali Hydroxide Layers
Interface engineering plays a vital important role for improving the efficiency and lifetime of
organic solar cells. Tailoring a Schottky barrier to quasi-ohmic contact can overcome interface
losses. In this chapter, we fabricated high-efficiency inverted organic solar cell via inserting an
ultrathin alkali hydroxide layer. The functionality of alkali hydroxides as interface modification
layers was studied and the impact on device performance was investigated.
The idea of inserting thin layers of Ba(OH)2 originally was presented by L. P. Lu et al.181
for
OELD applications. Rationalizing the fundamental mechanism of dipole assisted interface
alignment I was stimulated to study that concept in great detail for organic solar cells. Based on
green and environmentally safe processing, I further decided to look into alternatives to Ba(OH)2
which can be processed by non-hazardous solvents like deionized water and isopropyl alcohol
instead of 2-methoxyethanol.
My focus in this research was the device optimization as a function of the interlayer processing
conditions. All device preparations, processing, characterization incl determining device
performance, lifetime, contact barrier as well as the impact of light soaking was done solely by
me. At a very early stage of the investigations it however became clear that detailed interface
investigations need to be undertaken, triggering me to start a cooperation with Prof. Neal R.
Armstrong at the University of Arizona. Prof. Armstrong investigated the interfacial energetics
of these electron collecting contacts on layer systems either prepared in Erlangen or prepare din
Arizona following our recipes using UPS measurements. These studies confirmed the presence
of a large interface dipole but also surprisingly showed a new interface state between the Fermi
energy and the C60 HOMO for alkali hydroxide-modified AZO contacts. The interpretation of
70
these energetic states on the device performance was done in cooperation with Neal Armstrong
and his assistant Dr. Clayton Shallcross.
Finally, and in order to relate the interface energetics to the observed device losses,
representative devices were simulated with a numerical device model based on a 1D drift–
diffusion approach by Prof. L.J.A. (Jan Anton) Koster at University of Groningen.
Part of this chapter has been published (reproduced with permissions):
H. Zhang, T. Stubhan, N. Li, M. Turbiez, G. J. Matt, T. Ameri, C. J. Brabec, J. Mater. Chem.
A, 2014,2, 18917-18923
H. Zhang, R. C. Shallcross, N. Li, T. Stubhan, Y. Hou, W. Chen, T. Ameri, M. Turbiez, N. R. Armstrong and C. J. Brabec, Adv. Energy Mater. 2016, 1502195
5.1 Solution-processed alkali hydroxide interlayers in organic
solar cells
In this section, a route to solve one of the severe challenges in large scale printing green
interface layer of organic solar cells is demonstrated. The performance of inverted OSCs can
be significantly improved by tailoring a quasi-ohmic contact via solution-processed alkali
hydroxide (AOH) interlayers on top of n-type metal oxide layers. Furthermore, the strong
“light-soaking” effect is no longer observed in devices with a solution-processed alkali
hydroxide interface. Solution processed low-cost AOH interlayer increased the efficiency of
the inverted device by dominantly reducing the energy barrier for electron extraction from
PC61BM. The drastic improvement in device efficiency and the simplicity of fabrication by
solution processing suggest AOH as a promising and practical route to reduce interface
induced recombination losses at the cathode of organic solar cells.
5.1.1 A solution-processed Ba(OH)2 modified AZO for inverted OSCs
In the pioneering years, it has been found that interfacial losses and inefficient electron
collection can be overcome by inserting ultrathin buffer layers with low work function,
71
including calcium Ca, Ba or LiF, between the extraction electrode and the active layer of
OSCs. In order to overcome poor electron extraction from OSCs with a metal oxide
interfacial layers, solution-processed Ba(OH)2 is inserted between an n-type AZO and the
active layer (pDPP5T-2:PCBM) in inverted OSCs to tailor the contact barrier between the n-
type metal oxides and the lowest unoccupied molecular orbital (LUMO) of PCBM.
Figure 5-1a shows the inverted device architecture used for study. The j-V characteristics of
inverted OSCs based on different electron transport layers without and with Ba(OH)2 are
shown in Figure 5-1b, and summarized in Table 5-1. The control device with bare AZO as
an electron transport layer exhibits an VOC of 0.50 V, a JSC of 12.31 mA cm-2, a FF of 53.90%
and a PCE of 3.33%. The series resistance is 1.05 Ω cm2 and the shunt resistance is 11.58 KΩ
cm2. When Ba(OH)2 layer is inserted between AZO and the active layer, the JSC increased
from 12.31 mA cm-2 to 15.79 mA cm-2, the VOC from 0.50 V to 0.55 V and the FF from 53.90%
to 69.18%. The PCE increases significantly up to 6.01%, and the improvement is mainly due
to a higher JSC and an increased FF. Figure 5-1b shows that the photocurrent density of the
device with Ba(OH)2 is higher than for the one without Ba(OH)2 under illumination. This is
in agreement with the assumption that the Ba(OH)2 layer managed to reduce the extraction
barrier. In parallel, Figure 5-1c shows that the injected current density in the dark (0.7 V –
2.0V) is higher with than without Ba(OH)2 and the series resistance decreased from 1.05 Ω
cm2 to 0.88 Ω cm2. This is in agreement with a reduced injection barrier.
b a
72
Figure 5-1. a) Schematic of the inverted organic solar cells structure in this chapter. b) j-V characteristics
of inverted organic solar cells without and with the Ba(OH)2 layer. c) Corresponding logarithmic plot of
dark j-V characteristics. d) EQE spectra of inverted organic solar cells without and with the Ba(OH)2 layer.
d) j-V characteristics of typical devices with different concentrations of Ba(OH)2. Reproduced with
permission from ref. 37. Copyright © 2014 The Royal Society of Chemistry.
Figure 5-1d shows the EQE spectra of both devices. The JSC values calculated from the EQE
data are 11.78 mA cm-2 for the device without Ba(OH)2 and 15.19 mA cm-2 for the device
with Ba(OH)2, both being in good agreement with the experimental Jsc values. The EQE
spectrum in Figure 5-1d points out that Ba(OH)2 dominantly improves the response from
short wavelength regime (350nm ~ 550nm), i.e., where PC61BM is absorbing. Here, the
quantum efficiency is nearly doubled from 28% to 50%, while the EQE in the long
wavelength regime (550 nm ~ 850 nm) is increasing from app 50% - 60 %. The higher EQE
values of the device with Ba(OH)2 layer in the PCBM absorption region demonstrate that
AZO/Ba(OH)2 more efficiently extracts electrons from PCBM, thus successfully reducing
recombination at or around the interface between the AZO and active layer. Several factors
may play a role there – a reduction of the extraction barrier does suppress bulk recombination
due to a lower local charge carrier density. Further, and since Ba(OH)2 has a deep HOMO181,
a confinement of holes in the active layer would reduce surface recombination.
c d
73
Table 5-1. Key values of the j-V characteristics of inverted OSCs without and with the Ba(OH)2 layer.
iOSCs VOC
(V)
JSC
(mA cm-2
)
FF
(%)
PCE
(%)
RS
(Ω cm2)
RSh
(kΩ
cm2)
Without Ba(OH)2 0.50 -12.31 -11.78a 53.90 3.33 1.05 11.58
With Ba(OH)2 0.55 -15.79 -15.19a 69.18 6.01 0.88 376.41
a) Current density values derived from the EQE measurement.
5.1.2 Green solution-processing of the AOH layer for inverted OSCs
Large area processing of OSCs requires green and environmentally safe processing.
Therefore green solvents are desired for each single layer for the whole roll-to-roll printing
procedures. Toxic solvents, such as 2-methoxyethanol and methanol37, 157, 182, have been
widely used for processing the interface layer, such as polyelectrolytes (PEs)157, Ba(OH)237
and ethanolamine183, due to their good solubility and compatibility to coating processes.
However, their toxicity makes them inacceptable for commercial production. Their ban in
mass production in industrial countries with strict environmental health and safety regulations
evidences the need for more environmentally friendly solvents. This was our main driving
force to look into alternatives to Ba(OH)2 which can be processed by non-hazardous solvents
like deionized water and isopropyl alcohol instead of 2-methoxyethanol.
In the following we demonstrate that the alkali-hydroxides are such a class of interface
materials. Figure 5-2 shows the j-V characteristics of inverted devices based on an AZO /
NaOH interface processed from different solvents (2-methoxyethanol, deionized water and
isopropyl alcohol), and a summary of the device performance is tabulated in Table 5-2. With
NaOH dissolved in deionized water and isopropyl alcohol, respectively, the JSC (ca. -15.56
mA cm-2), VOC (ca. 0.56 V), FF (ca. 69.04 %) and PCE (ca. 6.00 %) match well with that of
devices based on NaOH processed from 2-Methoxyethanol. Alkali hydroxides indeed are
compatible to processing from environmentally friendly solvents without paying any penalty
in performance.
Table 5-2. Key values of the j-V characteristics of inverted devices based on NaOH dissolving into
different solvents for solution-process.a
74
Different solvent VOC
(V)
JSC
(mA cm-2
) FF (%)
PCE (%)
avg (best)
2-methoxyethanol 0.56 -15.71±0.70 66.00±1.05 5.80 (6.10)
deionized water 0.56 -15.56±0.61 69.04±1.01 6.00 (6.08)
isopropyl alcohol 0.56 -15.49±0.22 69.06±1.37 5.98 (6.01) a Each value represents the average from five cells.
Figure 5-2. j-V characteristics of inverted OSCs based on NaOH dissolving into different solvents for
solution-process under illumination of an AM 1.5G solar simulator (100 mW/cm2), ZnO as ETL.
Reproduced with permission from ref.184. Copyright © 2016 Wiley-VCH Verlag GmbH & Co. KGaA.
Following the demonstrations of green and environmentally safe processing, in this part, we
extend this solution-processed approach to more alkali hydroxide (LiOH, NaOH and KOH)
interlayers for modification of n-type metal oxide (AZO and ZnO) layers for inverted
pDPP5T-2:PCBM devices. The j-V characteristics of inverted OSCs based on different ETL
without and with alkali hydroxides are shown in Figure 5-3, and a summary of the device
performance is tabulated in Table 5-3. For the devices with AZO as ETL, insertion of an
alkali hydroxide layer resulted in an increase in VOC from 0.50 V to 0.58 V, JSC from 12.33
mA cm-2 to 14.56 mA cm-2 (LiOH), 15.01 mA cm-2 (NaOH), 15.14 mA cm-2 (KOH) and FF
from 52.29% to 70.04% (LiOH), 68.20% (NaOH), 67.75% (KOH), thereby increasing the
PCE from 3.26% to 5.9% (LiOH), 5.91% (NaOH) and 5.90% (KOH), respectively. Similarly,
an improvement in performance was also observed for devices using ZnO as ETL when the
alkali hydroxide layer was inserted between ZnO and the active layer (see Figure 5-3c). The
same distinct trend was again observed: for AZO or ZnO devices with the alkali hydroxide
layers from LiOH to KOH, VOC achieved 0.58V, which is a full Voc of normal architecture
75
cells, and JSC increased slightly. JSC increased because of enhanced electron extraction at the
cathode, and VOC reached 0.58V because of lower cathode work function and a more
complete Fermi level pinning185, which will be discussed further in the section 5.2.
Figure 5-3. a) j-V characteristics of inverted organic solar cells without interlayer (black) and with thin
layers of LiOH (red), NaOH (green) and KOH (blue) under illumination of an AM 1.5G solar simulator
(100 mW/cm2), AZO as ETL. Inset: corresponding logarithmic plot of dark j-V characteristics. b) EQE
spectra of inverted organic solar cells corresponding to (a). c) j-V characteristics of inverted organic solar
cells without interlayer (black) and with thin layers of LiOH (red), NaOH (green) and KOH (blue) under
illumination of an AM 1.5G solar simulator (100 mW/cm2), ZnO as ETL. Inset: corresponding logarithmic
plot of dark j-V characteristics. d) EQE spectra of inverted organic solar cells corresponding to (c).
Reproduced with permission from ref.184. Copyright © 2016 Wiley-VCH Verlag GmbH & Co. KGaA.
Table 5-3. Key values of the j-V characteristics of inverted OSCs without and with the alkali hydroxide
layer.a
a
c d
76
OSCs
VOC
(V)
JSC
(mA cm-2
)
FF (%)
PCE (%)
avg (best)
RS
(Ω cm2)
RSh
(kΩ cm2)
AZO 0.50 -12.33±0.86 52.29±1.86 3.26 (3.70) 2.06 75.46
AZO/LiOH 0.58 -14.56±1.26 70.04±1.56 5.90 (6.13) 1.21 533.66
AZO/NaOH 0.58 -15.01±1.09 68.20±1.26 5.91 (6.24) 1.01 664.33
AZO/KOH 0.58 -15.14±1.56 67.75±2.20 5.90 (6.26) 1.41 572.41
ZnO 0.56 -13.41±0.51 64.23±0.20 4.86 (5.10) 1.54 207.97
ZnO/LiOH 0.58 -14.84±0.14 71.21±0.47 6.01 (6.09) 1.50 741.13
ZnO/NaOH 0.58 -15.69±0.26 66.23±0.60 5.95 (6.01) 0.97 656.31
ZnO/KOH 0.58 -15.71±0.22 69.10±1.00 6.16 (6.29) 1.07 554.77
a Each value represents the average from five cells.
The external quantum efficiency (EQE) curves of devices with various cathodes are shown in
Figure 5-3b,d. The integrated EQE is consistent with the JSC of the various devices
mentioned above. For AZO as ETL, the photocurrents integrated from the EQE data are
12.21, 14.36, 14.42 and 14.43 mA/cm2 for the control, LiOH, NaOH, and KOH, respectively.
Similarly, with ZnO as ETL, the photocurrent integrated from the EQE data are 13.28, 14.75,
15.09 and 15.18 mA/cm2 for the control, LiOH, NaOH, and KOH, respectively. The higher
EQE values of devices with alkali hydroxide layers are dominantly in the absorption regimes
between 425–550 nm and 625–800 nm simply reflect a more efficient electron extraction at
the cathodes upon applying the alkali hydroxide layer, and maybe small changes in the
microstructure of active layer at the interface.
5.1.3 The effect of an AOH layer on the device contact barrier
A low work function cathode is needed to match the LUMO of the acceptor (i.e. PCBM) to
suppress interfacial losses upon extraction of electrons. To elucidate a quasi-ohmic contact of
an alkali hydroxide layer on the interface formation of OSCs, electron-only devices with the
following device configuration: ITO/MOX/with or without AOH/pDPP5T-2:PCBM/Ca/Ag
were fabricated and the corresponding j-V characteristics are depicted in Figure 5-4.
77
We distinguished between two transport regimes for electron only devices. First, there is the
low voltage regime which should resemble ohmic transport in the absence of a contact barrier
respectively contact resistivity (Figure 5-4a and b). The high voltage regime is more
indicative for the bulk transport mechanisms and is discussed in Figure 5-4c and d. Electron
current density in the low voltage regime is significantly increased for injection from the
alkali hydroxide interlayer cathode, but only slightly increased when injected from the Ca/Ag
cathode. This asymmetric improvement is most directly explained by a double sided electron
barrier at the PCBM / ZnO interface, which is a larger injection barrier than extraction barrier.
As shown in Figure 5-4a, electron-only devices without the alkali hydroxide interlayer
present diode like j-V characteristics, resembling those of a Schottky type barrier contact for
the electron injection from the MOX into the active layer.12 For devices with alkali hydroxide
interlayers from LiOH to KOH, the slopes of the j-V characteristics increase in the low
voltage regime, which suggests an effectively decreased contact resistivity. We observe that
the trend in contact resistivity, from LiOH to KOH, scales with the enhanced dipole moment
at the interface due to the increasing electronegativity of the alkali metals. Next, we scrutinize
the high voltage regime, where the injection current from the alkali hydroxide / MOx device
results in space-charge limited current (SCLC) behavior186. Moreover, it is completely
symmetrical to the injection behavior from the Ca/Ag electrode187, which is known to form
an ohmic contact to fullerenes (see Figure 5-4c and d). Summarizing the single electron
device studies, we find significantly enhanced electron injection and extraction properties of
a ZnO or AZO interface upon coverage with a thin alkali hydroxide layer. The strength of the
effect scales with the electronegativity of the alkali atom. The effect appears to be
explainable by invoking reduction of a double sided energetic barrier at the MOx / fullerene
interface.
78
Figure 5-4. Current density versus voltage characteristics of (a) ITO/AZO/(without or with alkali
hydroxide layer)/pDPP5T-2:PCBM/Ca/Ag electron-only devices and (b) ITO/ZnO/(without or with alkali
hydroxide layer)/pDPP5T-2:PCBM/Ca/Ag electron-only devices. Logarithmic plot of current density
versus voltage characteristics of (c) ITO/AZO/(without or with alkali hydroxide layer)/pDPP5T-
2:PC61BM/Ca/Ag and (d) ITO/ZnO/(without or with alkali hydroxide layer)/pDPP5T-2:PC61BM/Ca/Ag
electron-only devices with SCLC curve fit. AOH stands for alkali hydroxide layer in the figure.
Reproduced with permission from ref.184. Copyright © 2016 Wiley-VCH Verlag GmbH & Co. KGaA.
Figure 4-12 adds further support for the proposed mechanism for charge collection as a result
of alkali hydroxide modification of the metal oxide contact. Photoluminescence (PL) studies
suggest that alkali hydroxides contribute in the passivation of defect sites on the AZO and
ZnO surface. PL spectra of AZO and ZnO show that the 510 nm green emission peak, which
is typically attributed to the oxygen vacancy emission of ZnO,188 is quite significantly
quenched upon coating with an alkali hydroxide.
a b
c d
79
Figure 5-5. a) PL spectra of glass/AZO with and without alkali hydroxides at 320 nm excitation. b) PL
spectra of glass/ZnO with and without alkali hydroxides at 320 nm excitation. Reproduced with permission
from ref.184. Copyright © 2016 Wiley-VCH Verlag GmbH & Co. KGaA.
5.1.4 The impact of the AOH layer on light soaking
The current generation of inverted OSCs with n-type MOX semiconductors (TiOx, ZnO, AZO)
as an electron transport layer (ETL) frequently displays so-called light soaking effects, which
is best described by the appearance of an S-shaped current−voltage curve near the open
circuit point.149, 150 As discussed in the Chapter 2, the origins of light-soaking effects in OSCs
are multiply and were correlated to trap filling in the metal oxide layer; chemisorbed oxygen
on the metal oxide surface that creates carrier depletion; space charge limited current effects
in the metal oxide layer or to an energy barrier at the metal oxide/organic active layer
interface.149 The light induced charge generation in the MOx is coming along with a filling of
the defective trap sites and is enhancing conductivity (photoconductivity) in parallel to
reducing injection barriers. Light soaking therefore requires exposure to light with UV light
(λ ≤ 400 nm) and gradually enhances solar cell efficiency over illumination time towards the
value of maximum performance.
Normally, the conductivity of highly crystalline ZnO degrades under environmental
conditions with time due to the presence of traces of oxygen and water, while UV treatment
can release this oxygen189 and restore the original electronic properties. Therefore, the strong
O-Alkali atom bond, especially the O-K bond, can enhance the interfacial layer´s resistance
a b
80
against oxygen and humidity and are expected to result in a reduced necessity for light
soaking.
Figure 5-6. Logarithmic plots of dark j-V characteristics of devices before (red) and after (blue) exposed to
light with UV components (λ ≤ 400 nm) in the AM 1.5G illumination, (a) ZnO, (b) ZnO/KOH, (c) AZO
and (d) AZO/KOH as ETL. Reproduced with permission from ref.184. Copyright © 2016 Wiley-VCH
Verlag GmbH & Co. KGaA.
As shown in Figure 5-6, time zero (t0) light soaking is indeed overcome for the KOH / metal
oxide (ZnO and AZO) interface for pDPP5T-2:PCBM based inverted organic solar cells.
Both, Rs and Rsh remain constant before and after light soaking. In contrast to devices solely
based on ZnO and AZO, the KOH/MOX devices show almost identical dark j-V
characteristics before and after AM 1.5G illumination (note that the UV part of the AM1.5
spectrum was not blocked). Pristine MOX cathodes (ZnO and AZO) require illumination of
the AM 1.5G solar spectrum for 1 min before the dark j-V characteristics show a significant
change in Rs and Rsh. Again, we want to confirm that these findings are in agreement with
the formation of an O-A bond, reducing the otherwise oxygen and humidity susceptible
dangling bonds. Therefore, no obvious “light-soaking” is observed in devices with alkali
a b
c d
81
hydroxide/ metal oxides ETLs. We anticipate that the final OSC product will not rely on UV
light treatment due to the interface treatment by solution-processed alkali hydroxides.
5.1.5 The environmental stability of devices with AOH layers
Interface stability is a key requirement for outdoor operation of organic solar cells. Among
the reasons for solar cell degradation are impurity induced shunts, photobleaching of organic
semiconductors via oxygen and water as well as corrosion and delamination of the metal
contacts due to the same reagents. Follow the understands of the strong O-Alkali atom bond,
especially the O-K bond, can enhance the interfacial layer´s resistance against oxygen and
humidity, the air-stability of the un-encapsulated inverted organic solar cells with or without
the alkali hydroxide interlayer was periodically measured under humidity of 55% and 25 °C
for 800 hours (see Figure 5-7).
Oxygen and moisture at the surface of ZnO are known to alter the electronic properties and
energetics at the ZnO/fullerene interface and thus may lead to an interfacial barrier over the
time of the degradation. Indeed, devices without alkali hydroxide modification interlayer
retain only 30% of their original PCE after being exposed to ambient conditions. Devices
with alkali hydroxide modified ZnO exhibit better long-term stability and devices with alkali
hydroxide modified interlayers maintain up to 80% of the initial efficiency for KOH/ZnO, 70%
for NaOH/ZnO, and 40% for LiOH/ZnO. The changes in normalized JSC, VOC and FF over
the period of keeping in air are shown in Figures 5-7a-d. As shown in Figures 5-7a-d, the
regular decay is mainly found in a drop of JSC and FF from LiOH to KOH due to the relative
activity and doping strength of the alkali metals. VOC remained almost unchanged during
ambient condition degradation. Similar results were observed in devices with AZO as ETL
(Figures 5-7e-h).
82
Figure 5-7. Air-stability test of the unencapsulated structures of ZnO (a-d) or AZO (e-h)-based device
with or without alkali hydroxides stored for 800 hours under ambient conditions (humidity of 55% and
a
c d
b
h
f
g
e
83
25 ℃ ). (a,e) Normalized VOC, (b,f) Normalized JSC, (c,g) Normalized FF, (d,h) Normalized PCE.
Reproduced with permission from ref.184. Copyright © 2016 Wiley-VCH Verlag GmbH & Co. KGaA.
5.2 The mechanism for the interfacial energetics of AOH-
modified AZO cathode contacts
After having explored the macroscopic effect of an alkali hydroxide layer on the device
performance, we next investigate the microscopic mechanism being responsible for the
reduction of the electron barrier. In this section, Cooperated with Prof. Neal R. Armstrong at
University of Arizona, we investigate the interfacial energetics of these electron collecting
contacts with a prototypical electron acceptor (C60) using ultraviolet photoemission
spectroscopy (UPS) measurements, which reveal the presence of a large interface dipole and
a new interface state between the Fermi energy and the C60 HOMO for alkali hydroxide-
modified AZO contacts. These novel interfacial gap states are a result of ground state
electron transfer from the metal hydroxide-functionalized AZO contact to the adsorbed
molecules, which are hypothesized to be electronically hybridized with the contact. These
interface states tail all the way to the Fermi energy, providing for a highly n-doped (metal-
like) interfacial molecular layer. This study exemplifies the importance of interfacial contact
doping/hybridization as a major tool for improving efficiency and stability of OSCs, as
required for scalable fabrication of printed alkali hydroxide layers, which is predicted to
enable cheap, reliable and practical routes to stable cathode contacts to fullerene based
OSCs.
5.2.1 The work function of MOx cathode was measured using the Kelvin
Probe Method
The influence of alkali hydroxides on the work function of MOx cathode was measured using
the Kelvin Probe Method (KPM) before and after depositing the various alkali hydroxide
layers. The work function of AZO was found to decrease from 4.40 eV to 3.91 eV (LiOH),
84
3.88 eV (NaOH) and 3.84 eV (KOH). Similarly, the work function of ZnO was found to
decrease from 4.30 eV to 3.88 eV (LiOH), 3.85 eV (NaOH) and 3.83 eV (KOH). Huang et al
studied ITO – alkali interfaces and concluded that alkali atoms are capable of forming an
interface dipole to metal oxides (M–O–A complex), where O and A stand for oxygen and
alkali metal species, respectively. Analogously, we suggest that an ultrathin layer of O-A is
formed at the AZO or ZnO surface with the alkali ions towards the active layer, resulting in a
reduction in the work function of the MOX. We speculate that O-A replaces the hydroxyl
group on the AZO or ZnO surface by surface chemical reaction: –OH + A+ → –O–A + H+,
similar to the mechanism discussed by Huang et al for alkali carbonates on ITO.147, 148
5.2.2 The interfacial energetics of KOH-modified AZO cathode contacts
with fullerene
In order to elucidate the interfacial energetics of the bare AZO and AOH-modified AZO
contacts with PCBM, ultraviolet photoemission spectroscopy (UPS) was carried out for a
model system consisting of evaporated C60 on both AZO and AZO/KOH electron collecting
contacts (Figure 5-8a and b). Deposition of the KOH interlayer on top of AZO reduces the
work function of the AZO contact (3.7 eV) by 0.9 eV to 2.8 eV. This decrease in work
function may be attributed to the aforementioned interface dipole formation between the
oxide contact and the alkali metal cation; the observed change in the surface potential of the
interface dipole formation may also contribute to a “push-back” effect94, 185, which can
account for a decrease in work function of the contact. Furthermore, we note that the work
function values measured by UPS for AZO and AZO/KOH are lower than the KPM
measurements. While the UPS values are in fact lower when compared to KPM, the relative
decrease in the AZO work function upon deposition of the KOH layer is significant (ca. 0.6
eV or greater) in both cases and, thus, demonstrates the validity of the AOH layers as low
work function interfacial modifiers.
85
Figure 5-8. UPS (He I) spectra [with respect to (w.r.t.) the Fermi energy] and energy level diagrams for
C60 deposited on bare AZO and KOH-modified AZO. a) Normalized secondary electron cutoff of bare
AZO and thin layers of C60 subsequently deposited onto the AZO surface (left panel); substrate valence
band and C60 HOMO region as a function of C60 thickness (right panel). b) Normalized secondary electron
cutoff of bare AZO/KOH and thin layers of C60 subsequently deposited onto the AZO/KOH surface (left
panel); substrate valence band and C60 HOMO region as a function of C60 thickness – magnification of the
spectra near the Fermi energy shows a clear interface state for the first ca. 5 nm of C60 (right panel). c)
Energy band alignment between C60 and AZO without (left panel) and with (right panel) KOH treatment.
Reproduced with permission from ref.184. Copyright © 2016 Wiley-VCH Verlag GmbH & Co. KGaA.
Upon deposition of C60 onto the AZO contact, there is a continuous increase (ca. 0.6 eV) in
the work function of the sample with subsequent C60 deposition that saturates around 5 nm
(Figure 5-8a, left panel). The valence band region shows an increase in intensity of the C60
HOMO with increasing thickness (Figure 5-8a, right panel); the onset of the HOMO feature
86
(ca. 2.1 eV) denotes the hole injection barrier (h), which, along with the work function, can
be used to determine the ionization potential of the C60 layer. Unlike the work function of the
sample, a monatomic shift in h is not observed with increasing C60 thickness, which rules out
band bending in the C60 layer; therefore, we attribute the shift in the work function and, thus,
vacuum level to an interface dipole formed by electron transfer from the AZO contact to the
C60 layer, providing for doped C60 anions in the vicinity of the contact that equilibrate the
Fermi energies of the two materials. Core level spectra for the substrate photoelectrons (i.e.,
Zn 2p 3/2) do not shift throughout the C60 layer deposition sequence; therefore, there is no
band bending observed in the AZO contact (Figure 5-10). For the core level spectra for the
substrate, it will be discussed further in the section 5.2.3.
Similar to the AZO/C60 interface, however more pronounced, we observe a continuous and
substantial increase (ca. 1.3 eV) in the work function of the AZO/KOH contact with
increasing C60 thickness that saturates after ca. 5 nm (Figure 5-8b, left panel). The increased
magnitude of the work function shift can be ascribed to the much lower work function of the
KOH-modified AZO contact. Again, we observe the evolution of the C60 HOMO and with
increasing thickness near the Fermi energy of the spectrum (Figure 5-8b, right panel). We
also observe a relatively constant h, and, like the bare AZO contact, there is no significant
band banding in the C60 layer. In contrast to the bare AZO contact, we observe a well-
resolved ionization feature that grows in and then disappears after ca. 5 nm between the C60
HOMO feature and the Fermi energy (see 5x magnified spectra in Figures 5-8a and b and
semi-log plot in Figure 5-9). This relatively low intensity ionization feature can be ascribed
to an interface or gap state that forms only near the contact. UPS work by Schulz et al. on the
interfacial energetics between soft-sputtered ZnO and C60 has shown a very similar ionization
feature, which, with the aid of DFT calculations, is ascribed to a hybrid interface state with
charge transfer character that is formed due to electron transfer from ZnO to C60190. The
interfacial states observed via UPS by Schultz et al. for monolayer C60 on ZnO have been
correlated with significantly increased electron delocalization times (relative to bulk C60),191
providing direct evidence for interfacial hybridization between negatively charged C60
molecular orbitals with states on the ZnO surface. Thus, we conclude that there is
87
hybridization (i.e., orbital mixing) between C60 anions and surface states of the AZO/KOH
electron-selective contact that afford observable filled states that exponentially tail to the
Fermi energy (Figure 5-10).
In contrast to the bare AZO contact, we do observe a small shift (ca. 0.3 eV) toward the
Fermi energy (i.e., toward lower binding energy) for the Zn 2p 3/2 core level photoelectrons
that originate from AZO layer that is beneath the KOH layer for the first ca. 5 nm of C60
deposition (Figure 5-10). This core level shift for the underlying contact is indicative of
upward band bending in the AZO layer due to charge transfer to the C60 layer through the
insulating KOH layer, which incidentally does not show any signs of band bending.
Figure 5-9. Logarithmic plots of high kinetic energy region (near Fermi) as a function of C60 thickness
satellites, the spectra were smoothed by applying a 5-pt boxcar averaging function. Reproduced with
permission from ref.184. Copyright © 2016 Wiley-VCH Verlag GmbH & Co. KGaA.
5.2.3 Core level X-ray photoelectron spectra for the AZO and AZO/KOH
substrate signals upon deposition of C60
In order to better understand the growth type (i.e., layer-by-layer or island growth) and the
situation of band bending, we take a closer look at the attenuation of the substrate PES
signals due to deposition of C60 (Figure 5-10d). The layer-by-layer growth model predicts
88
that the slope is directly related to the thickness and inversely related to the inelastic mean
free path (IMFP) at normal takeoff angles by the general attenuation equation192:
equation (5-1)
where I0 is the photoelectron intensity in the absence of C60 (i.e., the “bulk” substrate), I is the
intensity at a given C60 thickness, d is the C60 thickness in nm and λn is the IMFP in nm;
therefore, the slope of the linear regression is the reciprocal of the IMFP, providing values of
ca. 4.3 nm and 6.3 nm for Zn 2p 3/2 and K 2p 3/2 photoelectrons originating from AZO and
AZO/KOH substrates, respectively. These IMFP values are significantly larger than one
would expect when consulting the empirical relationship between the IMFP and the kinetic
energy (KE) of the photoelectrons for KEs greater than 150 eV192:
equation (5-2)
where Bn is an empirical fitting coefficient specific for photoelectrons traveling through a
matrix of “organic” material, and E is the KE of the photoelectrons in eV (e.g., 464.4 eV and
1192.1 eV for Zn 2p 3/2 and K 2p 3/2 photoelectrons that originate from the AZO and
AZO/KOH contact, respectively.) Assuming layer-by-layer growth, the equation predicts
IMFP values of ca. 1.9 eV and 3.0 eV for the Zn 2p 3/2 and K 2p 3/2 photoelectrons
originating from the AZO and AZO/KOH substrates, respectively. These predicted values are
less than one half of the values calculated from Figure 5-10d using the measured attenuation
of the substrate core level signals; therefore, we conclude that the C60 molecules deposit via
island growth rather than conformal/layer-by-layer growth, providing for bare substrate
regions between C60 islands that afford observation of the substrate signals at thicknesses that
would normally not be observed for layer-by-layer growth. This island growth mode for C60
on these substrates explains why we observe the interface state for thicknesses that are
significantly greater than a monolayer of C60 (0.82 nm)193.
- ln II0
=d
ln
ln = BnE1/2
89
Figure 5-10. Core level X-ray photoelectron spectra for the AZO and AZO/KOH substrate signals upon
deposition of C60 (see legend for C60 thickness in nm). (a) Zn 2p 3/2 peak from the AZO substrate. (b) Zn
2p 3/2 peak from the AZO/KOH substrate (inset shows the peak binding energy vs. C60 thickness). (c) K
2p spectra from the AZO/KOH substrate. (d) Attenuation of the substrate core level signal for AZO (Zn 2p
3/2; black) and KOH (K 2p 3/2; red). Reproduced with permission from ref.184. Copyright © 2016 Wiley-
VCH Verlag GmbH & Co. KGaA.
5.2.4 A semi-quantitative energy level alignment between the contacts and
the C60 layer
Using these UPS and XPS results, we can infer a semi-quantitative energy level alignment
between the contacts and the C60 layer (Figure 5-8c). In the absence of inverse
photoemission spectroscopy, we extrapolate the electron affinity of the C60 layer and the
conduction band of the AZO contact using the transport band gap of 2.4 eV194 and optical
90
band gap of 3.7 eV195, respectively. The AZO contact appears to make for an acceptable
electron extracting contact (Figure 5-8c, left panel); however, the addition of the KOH layer
significantly lowers the AZO work function to the extent that a greater degree of charge
transfer must take place in order to align the Fermi energy with the adsorbed C60 layer,
providing for filled hybridized interface states that tail all the way to the Fermi energy
(Figure 5-10). Intersection of these the hybridized interfacial molecular states with the Fermi
level implies that they are degenerately n-doped and metallic. While the C60 LUMO appears
to form a barrier to electron extraction at the AZO/KOH/C60 interface, the presence of metal-
like interface states affords quasi-Ohmic contact and barrier-free electron extraction. Schulz
et al. demonstrated that the presence of these hybrid interface states proved to enhance
electron collection, which further supports the enhanced power conversion efficiency of the
OSCs with the AOH interlayers.
Since we explain the observed UPS results with via interfacial charge transfer models (i.e.,
interface dipole and hybridization), the effect should only occur within the first monolayer
(0.82 nm)193 of C60 deposition; however, we observe energetic changes that extend up to ca. 5
nm from the contact interface. We explain this discrepancy by determining that the C60
molecules grow as three-dimensional islands rather than conformal layers. By analyzing the
attenuation of the substrate core level signals (e.g., Zn 2p 3/2 and K 2p 3/2 for AZO and
AZO/KOH, respectively) with increasing C60 thickness, we are able to rule out layer-by-layer
growth for C60 on these relatively rough surfaces (see Figure 5-10). Therefore, we observe
the interface specific effects at elevated nominal thicknesses until the spaces between the C60
islands are eventually filled and the film coalesces. We note that the observed thickness
values are used to construct the energetic diagram in Figure 5-8c; however, it is understood
that the interface dipole and hybridization likely only occur in the first adsorbed monolayer.
The consequences of both effects, reduction of the ZnO and AZO work function on the one
hand and metallic doping of the first fullerene layer on the other hand are also well reflected
in an enhanced JSC and FF of the OSCs and a reduced contact resistivity of the electron-only
devices in the section 5.1.
91
5.3 Interface-induced Suppression of the Bulk Recombination in
Organic Bulk Heterojunction Solar Cells
In this section, we studied the electrical properties of the AOH-modified diode devices to
understand the charge accumulation, extraction and recombination at AOH-modified
contacts. In addition, the distinguished recombination kinetics by measuring the j-V
characteristics of devices at various illumination intensities was investigated. Remarkably, a
reduction of the biomolecule recombination strength was observed in case of KOH-modified
devices. Combined with a numerical device model based on a 1D drift–diffusion approach by
Prof. L.J.A. (Jan Anton) Koster, the majority carrier density at the respective interface gets
reduced when inserting an interface modification layer. The bulk bimolecular recombination
with the photo generated minority carriers gets reduced. And, although this is a bulk
recombination, this effect is only occurring at the interface, therefore reducing the current
rather than the voltage. We consider the effect is interface-induced suppression of the bulk
recombination in organic bulk heterojunction photovoltaics.
5.3.1 The electrical properties of the AOH-modified diode devices
5.3.1.1 Ideality factor and reverse saturation current density
There are two processes dominating the ideality factor (n) in OSCs, both related to
recombination phenomena at different time scales. It is known that excitons generated under
illumination in the active layer must diffuse to the donor/acceptor (D-A) interface (sub
nanosecond regime).62, 42 After dissociation, the individual carriers are transported in the
respective donor and acceptor phases to the electrodes (nanosecond – microsecond regime).
The first process is exciton diffusion, where geminate bulk recombination is dominant; the
second process is charge transport to and charge accumulation at the respective
organic/electrode interfaces, with 1st order or 2nd order recombination (trap induced or even
space charge induced) being dominant.11, 45, 62 Both processes do induce the loss of charge
carriers and enlarge n, thus reducing the FF. Devices with n >> 1 typically suffer from high
electron-hole recombination, which is of 2nd or higher order.11, 45, 62
92
To understand recombination status in OSCs, in this case we analyzed the dark current
characteristics as shown in Figure 4-2b, using the equivalent circuit model.58, 59 In this model,
the j-V characteristics can be derived by the following equation:
𝐽 = 𝐽0 {𝑒𝑥𝑝 [𝑞(𝑉−𝐽𝐴𝑅𝑠 )
𝑛𝑘𝑇] − 1} + (𝑉 − 𝐽𝐴𝑅𝑠)/𝐴𝑅𝑠ℎ − 𝐽𝑝ℎ(𝑉) equation (5-3)
where J0, q, n, Rs, Rsh, k, T, and A is reverse saturation current density, elementary charge,
ideality factor, series resistance, shunt resistance, Boltzmann constant, absolute temperature,
and the active area of the device, respectively. By measuring the dark current density while
sweeping the bias voltage, the diode characteristics including saturation current density (J0),
diode ideality factor (n), series resistance, and shunt resistance can be evaluated. When the
linear fitting is extrapolated versus 0 V, J0 can be calculated from the y-axis intercept while the
diode ideality factor is determined by the slope of the line, as shown in Figure 5-11b.
For devices without Ba(OH)2, an ideality factor of n =1.87 is typically indicative for higher
order bulk recombination or for enhanced recombination at an interface due to charge
accumulation. Surface charge accumulation functions as a major FF limitation, either via
surface recombination or via enhanced bulk recombination due to delayed extraction. The
significant reduction of the diode ideality factor upon insertion of a Ba(OH)2 layer strongly
suggests, that charge accumulation at an interface in combination with an extraction barrier is
a dominant loss mechanism in the devices without Ba(OH)2.
The lower J0 value (4.70 × 10-7 mA cm-2) of the device with Ba(OH)2 stands for a significantly
reduced junction leakage current density. Typically, OSC BHJ devices have more than one
diode, with the BHJ as the most prominent one formed between the polymer and the fullerene,
and with one or two planar heterojunction diodes at the semiconductor / charge extraction
layer. A leakage current is associated with each of them. The significant reduction of the
leakage current due to insertion of a Ba(OH)2 layer strongly suggests this interface as the
problematic one. Charge extraction layers with insufficient selectivity lead to various
phenomena, among them the so-called photoshunt, photoconductivity or enhanced leakage
currents. The dominant factor to suppress the reverse saturation current is than ascribed to the
93
formation of a more selective contact between AZO and the semiconductor. Here, this is in
agreement with an increase of the injection barrier for minority carriers, i.e. holes from ITO
into the HOMO level of P3HT. In other words, the insertion of a Ba(OH)2 layer facilitates
electron injection while hampering in parallel hole injection.
Figure 5-11. a) Mott-Schottky capacitance plots for the data of voltage the device without and with
Ba(OH)2 layer. Inset: the intercept with voltage axis yields the built-in voltage. b) Semilog plot of dark j-V
characteristics curves of inverted OSCs without and with the Ba(OH)2 layer .The dashed lines show the
extrapolated line for determining saturation current density from intercept and diode ideality factor from
the slope. Reproduced with permission from ref. 37. Copyright © 2014 The Royal Society of Chemistry.
5.3.1.2 Capacitance
Impedance spectroscopy is a nondestructive tool for the determination of the density of free
charge carriers and thus non-geminate recombination losses in BHJ solar cells. In order to
further investigate the recombination at AOH-modified contacts, we performed capacitance
versus voltage (C-V) measurements for solar cells without and with Ba(OH)2, as shown in
Figure 5-11a. Two important parameters can be given directly from C-V plots. The built-in
voltage (Vbi) is defined by the intersection of the 1/C2 curve and the horizontal bias axis. On
the other hand, the upper-limit of the dopant (carrier) density (N) is determined from the
slope of Mott - Schottky plot182, 196 using N = −2(ⅆC(x)2/ⅆV)−1/qε0εrA2, where q accounts
for the elementary charge, ε0 is the dielectric constant of vacuum, εr is the relative dielectric
constant of the semiconductor (assuming εr of 3.5), and A corresponds to the device active
area. This analysis provides an appropriate comparison of the differences in Vbi and N
between devices without and with Ba(OH)2. We find that Vbi is slightly increased in the
a b
94
devices with an inserted Ba(OH)2 layer, resulting in an enhanced electrical field supporting
charge carrier extraction. Devices without Ba(OH)2 showed a Vbi of approximately 0.51 ±
0.01 V and a maximum dopant density of around 4.97 × 1018 cm-3 . Compared to the devices
with Ba(OH)2 layer, the Vbi increased to 0.55 ± 0.01 V and the maximum value of N
increased as well to approximately 6.55 × 1018 cm-3. The increased dopant density (N)
suggests an increase in the surface charge density, either by an absolute higher number or by
a reduced thickness of the charged regime. Based on previous reports on other alkali-based
interfaces, this increase in the surface charge density can be attributed to interactions of Ba2+
ions with the surface of the active layer, giving rise to interfacial ionic charge doping.
Combined with the microscopic mechanism of the interfacial energetics of alkali hydroxide-
modified MOx cathode contacts, Intersection of the hybridized interfacial molecular states
with the Fermi level implies that they are degenerately n-doped and metallic. The
arrangement of Ba2+ ions is suggested to dominantly promote a reduction of the electron
extraction barrier, enhancing extraction of carriers from PCBM to AZO. Noteworthy to
mention, that a reduction of the barrier is further expected to enhance electron injection from
AZO into PC61BM as well. Depending on the nature of the interaction between Ba2+ and
AZO, an interface dipole and hybridization is expected to form.
With the higher built-in voltage and lower charge accumulation at the AOH-modified contact,
recombination in OSCs can be effectively suppressed. Using electron-only diodes and the
space-charge limited current (SCLC) model, in a pristine device without Ba(OH)2, the
average electron mobility was calculated to be 3.42 × 10-3 cm2 V-1 s-1. Interestingly, the
device with Ba(OH)2 showed the similar average electron mobility of 4.70 × 10-3 cm2 V-1 s-1.
It suggests that the presence of Ba(OH)2 actually has no effect on the electron mobility of the
bulk active layer in OSCs. Therefore, the suppressed recombination is not from the bulk of
devices.
5.3.2 The recombination kinetics by measuring the j-V characteristics of the
AOH modified devices
The simplest and most common method to distinguish between bimolecular recombination
95
and trap-assisted recombination is probing the VOC and JSC dependence on incident light
intensity, as discussed in the section 2 theory part. We studied the recombination kinetics by
measuring the j-V characteristics of devices at various illumination intensities without or with
KOH, as shown in Figure 5-12.
VOC follows the empirical expression 𝑉𝑂𝐶 =1
𝑒(𝐼𝐸𝑑 − 𝐸𝐴𝑎) −
𝜂𝑘𝑇
𝑒ln (
𝐽𝐵𝑅
𝐽𝑆𝐶) and
JSC follows the expression 𝐽𝑆𝐶 = 𝛽(𝐼)α.
In AZO-based polymer OSCs, α strongly depends on interlayers contact conditions. Figure
5-12a shows the short circuit current versus intensity for devices without and with KOH
layers. Fitting these data yields α = 0.94 for the control device and up to α = 0.97 for the
device with KOH.
Figure 5-12. a) Measured JSC of devices without interlayer (black) and with KOH (blue) interlayer plotted
against light intensity on a logarithmic scale. Fitting a power law (solid lines) to these data yields α. b)
Measured Voc of devices without interlayer (black) and with KOH (blue) interlayer as a function of light
intensity, together with linear fits to the data (solid lines).
The limitations in the open-circuit voltage of OSCs are attributed to the charge density within
the active layer and thus the splitting of the quasi-Fermi levels. VOC is then a measure of the
amount of recombination in the device. VOC depends on the saturation current of the devices
and the light-generated current. At open-circuit voltage, the photo-generated charge carriers
recombine within the devices. Thus, the detailed information of various mechanisms can be
obtained by measuring the VOC as a function of the light intensity.
Figure 5-12b shows the VOC versus light intensity relationship for the AZO devices without
and with KOH. The slope for the control device without potassium hydroxide layers is 2𝑘𝑇/𝑞,
a b
96
implying that bimolecular recombination and monomolecular recombination is dominant
even at open circuit point. After inserting KOH, the slope of the devices reduced to 1.13
𝑘𝑇/𝑞. The significant reduction of the slopes due to insertion of potassium hydroxide layers
strongly suggests KOH significantly reduces bimolecular and trap-assisted recombination at
organic semiconductor / AZO interface, and recombination kinetics is governed by a
combination of trap-assisted and bimolecular type. Since KOH might reduce the density of
interfacial traps between the organic semiconductor and AZO contact for quasi-ohmic
electron extraction. Moreover, the modified contact/active layer interface is efficiently
injecting electron density into the surface located fullerenes, thus creating a thin, up to ca. 5
nm thick, n-doped interfacial layer that is hybridized with the contact. These results imply
that the KOH layer effectively reduces the density of traps between the acceptor and Ag
contact, and hence bimolecular and trap-assisted interface recombination is suppressed.
Through the theoretical derivation in the section 2.3, we further note that, although this
bimolecular recombination mechanism is located or induced by an interface, it will behave
very different from classical surface recombination. In difference to surface recombination
which mainly impacts VOC, this interface related recombination mechanism mainly will
reduce JSC. Therefore, we could not explain this kind of extraction barrier in terms of reduced
surface recombination. This is very close to the p-i-n scheme, and that mechanism should be
one clear benefit of inserting semiconducting interface layer with little / no charge generation.
5.3.3 1D drift-diffusion modeling
We use numerical drift diffusion simulation to study the interface modification and
investigate how the physical recombination mechanisms is affected by interface
modification layer as a blocking layer, charge extraction close to the interface, and band
bending.
97
Figure 5-13. a) j-V characteristics of inverted polymer OPVs without (black) and with (blue) the KOH
layer. b) Calculations of the predicted solar-cell response with (blue) and without (black) blocking layers.
In section 5.1, our studies on KOH modified AZO electron extraction layers gave excellent
insight into the microscopic processes occurring at such interfaces. KOH was proven to
simultaneously reduce the extraction barrier between PCBM and AZO and dope the surface
fullerene layer . Such n-doped - intrinsic interface layers are regarded as ideal interfaces for
field driven devices. We demonstrated that the performance of OPVs can be significantly
improved through making ohmic cathode contacts via KOH to treat AZO layer. In the j-V
characteristics of OSCs (Figure 5-13a), we can find, without KOH to modify the electrode,
OSCs exhibited insufficient Voc and JSC. Most interestingly, the dominant impact of interface
engineering resulted in a significant higher photocurrent. This particular correlation – namely
the increase in the photocurrent upon interface engineering – was also reported by other
studies. The connection between interface engineering and photocurrent enhancement is not
trivial at first glance and one would have expected to see surface recombination as the
dominant loss. In diffusion controlled devices, surface recombination is typically seen by a
major loss in VOC accompanied by a minor loss in JSC. Field driven devices are different in
that aspect where surface recombination at the electrodes typically becomes relevant in the
limit when VOC approaches the built-in or flat-band voltage. Under flat-band conditions, the
minority carrier concentration at the contacts does increase with illumination, but the voltage
cannot increase any further as it becomes fixed by the contacts or limited from the built-in
field. Petersen et al. recently predicted that the relevance of surface recombination in organic
solar cells will depend on the presence of an extraction barrier. We probe recombination in
a b
98
the devices with 1D drift-diffusion modeling to investigate the effects of surface
recombination on the open-circuit voltage. As shown in Figure 5-13b, without the blocking
layers, JSC reaches 78.9 A/m2. Therefore there is a 21% recombination. Introducing a hole
blocking layer and an electron blocking layer, JSC increases to 86.2 A/m2, so a 9%
improvement in current. However, there are no losses in VOC in this simulation. The
recombination is reduced at the modified electrode due to bimolecular recombination in the
bulk rather than surface recombination, we assume.
This numerical model hinges on the description of the active blend layer of organic solar cells
as one effective medium (the so-called metal-insulator-metal model), described in detail
elsewhere. The numerical program considers diffusion and drift of free carriers in one
dimension. Also, drift and diffusion of charge carriers, the effect of charge density on the
electric field and bimolecular recombination are included. From the modeling of current-
voltage characteristics (Figure 5-13b), it is found that the bimolecular recombination strength
is significantly reduced, and is governed by the slowest charge carrier.
Figure 5-14. a) The device with (solid lines) or without (dashed lines) blocking layer at short-circuit,
shows the electron and hole current densities (red and blue, respectively), (b) the device with (red lines) or
without (black lines) blocking layer at short-circuit shows the recombination rate.
In the simulations an active layer thickness of 100 nm is assumed. The relevant energy
difference between the HOMO level of the donor and of the LUMO level of the acceptor is
the effective band gap, assumed to be 𝐸𝑔𝑎𝑝𝑒𝑓𝑓 = 1.10 eV. Specifically, we assume interface
modified treatment as a blocking layer and consider an extraction barrier in the real devices
without the blocking layer. We simulated a device with charge mobilities of 1.0×10-7 m2/Vs
a b
99
and 2.0×10-8 m2/Vs for electrons and holes, respectively. We chose the light intensity
(uniform spatial profile) such that the current would be 100 A/m2 in the absence of
recombination. The hole blocking layer is 17 nm thick and the electron blocking layer is 8 nm
thick, which are the optimum thicknesses. However, as the recombination losses are not so
large, the difference is also not so large, but it's there. The effect we see here is indeed bulk
recombination, albeit close to the electrodes. Figure 5-14b shows fits to the recombination
rate as a function of position to be compared with a generation rate of 6.25×1027 m-3s-1, under
short-circuit conditions. The cathode is on the left (x=0 for normal, x=-17 nm for blocking
layers). The blocking layers reduce the recombination as the electron density is much reduced
in the active layer. What we find is indeed a second order effect: the hole blocking layer
(electron blocking layer) reduces the majority carrier density in the active layer. As a result
fewer minority carriers recombine. In the simulations, the electron density at the cathode is
fixed to a high value (see Figure 5-14a). In the reference device without blocking, this means
that there are very many electrons in the first couple of nm in the active layer. These
recombine with the photogenerated holes close to the cathode. Inserting a blocking layer
changes the electron density in the active layer. The cathode still induces as many electrons,
but this time in the blocking layer and not in the active layer. As there are no (= hardly any)
holes in the blocking layer, there is no recombination. The blocking layer simply makes the
photo-generated holes don't get too close to the enormous electron density induced by the
cathode. Similar reasoning applies to the anode. We assume the extraction barrier would
correspond to slow surface recombination of the majority carriers. Comparing with the
reference device without blocking layer, a device with reduced majority surface
recombination at the contacts yields a much higher VOC, but there's no significant change in
JSC. We speculate that there are some many majority carriers that, even if they're slow, still
enough can exit the device in order to maintain the original majority densities at the contacts.
Therefore, we could not explain this kind of extraction barrier in terms of reduced surface
recombination. This is very close to the p-i-n scheme, and that mechanism should be one
clear benefit of inserting semiconducting interface layer with little / no charge generation.
Clearly these results are in good agreement with the experimental data, showing that the
majority carrier density at the respective interface gets reduced when inserting an interface
100
modification layer. That way, the bulk bimolecular recombination with the photo generated
minority carriers gets reduced. And, although this is a bulk recombination, this effect is only
occurring at the interface, therefore reducing the current rather than the voltage.
5.4 Conclusions
In summary, we showed that alkali hydroxide coatings on MOx (ZnO or AZO) are a very
general concept to form quasi-ohmic contacts between the MOx and fullerenes, which may
enable readily scalable printing of solar cells with air-stable, inexpensive contacts that exceed
the performance of other interface modifiers that have been previously used to achieve low
effective work functions. In the case of pDPP5T-2:PC61BM this leads to significantly
improved device performance, both, in terms of efficiency and lifetime. The insertion of the
alkali hydroxide layer switches the electronic properties of the cathode from barrier-limited
electron harvesting, to a quasi-ohmic electron extraction contact. In addition, the strong
“light-soaking” effect is no longer observed in devices with a solution-processed alkali
hydroxide interface.
Multiple microscopic mechanisms can be proposed to explain the superior performance of
MOx/AOH modified interfaces. First, alkali hydroxide treatment reduces the work function
of the metal oxide due to the formation of a surface dipole with the oxide contact. Second, the
modified contact/active layer interface is efficiently injecting electron density into the surface
located fullerenes, thus creating a metal-like n-doped interfacial molecular layer that is
hybridized with the contact. Lastly, alkali-hydroxide coatings seem to reduce the density of
surface defects of both the ZnO and AZO contacts
We studied the electrical properties of the AOH-modified diode devices. Charge
accumulation at an interface in combination with an extraction barrier was a dominant loss
mechanism in the devices without AOH-modified diode devices, and then enhanced junction
leakage current density. In addition, the distinguished recombination kinetics by measuring
the j-V characteristics of devices at various illumination intensities was investigated.
Remarkably, a reduction of the biomolecule recombination strength was observed in case of
101
KOH-modified devices. Charges accumulation can be greatly reduced the extraction time, via
modifications of the electrode/active layer interface. Combined with a numerical device
model based on a 1D drift–diffusion approach by Prof. L.J.A. (Jan Anton) Koster, we
developed the effect of perfect contacts with blocking layer leading to a reduced the majority
carrier density at the respective interface in OPVs. The effect of reduced interfacial majority
carrier at the electrodes was simulated. Because of a reduced the majority carrier density, the
bulk bimolecular recombination with the photo generated minority carriers gets reduced. And,
although this is a bulk recombination, this effect is only occurring at the interface, therefore
reducing the current rather than the voltage. In other words, Interface-induced suppressed the
bulk recombination in organic bulk heterojunction solar cells.
Overall, the sum of these effects allows forming a uni-polar, quasi-ohmic contact between a
MOx (ZnO or AZO) and fullerenes. The impacts of such a loss free contact on the device
performance are manifold: Enhanced charge collection, reduced 2nd order recombination at
the interface, absence of light soaking as well as enhanced environmental stability. Alkali
hydroxide layers exemplify the importance of unipolar but quasi-ohmic contacts as an
essential prerequisite for designing production relevant electrodes, and open a promising and
practical route to cathode interfacial modification via solution-processed techniques.
102
103
Chapter 6
Roll to Roll Compatible Fabrication of
Inverted Organic Solar Cells
To realize the widespread applications of organic solar cells, high performance, low cost and
larger scale devices are needed. Thus, reasonable design rules and routes are required to define.
Generally, solution processing and low temperature interface layer is a key to realize the goal.
The strategy of blending self-organization interfacial materials (PCBDAN or PEI) in donor-
acceptor BHJ blends originally was presented by D. Ma et al.197
and H. Kang et al.198
However,
the devices were manufactured by spin-coating under inert conditions; spin-coating is a batch
process and as such incompatible with the visionary concept of fully printed or coated organic
solar modules. In this approach PCBDAN or PEI had to be dissolved in a polar solvent, such as
2-methoxyethanol, before mixing with the photoactive solutions. Addition of the poor (polar)
solvents to the photovoltaic semiconductor ink may complicate the microstructure formation of
the active layer. Based on large-area high-speed roll-to-roll processing, I further decided to
design a simple approach to printing efficient devices compatible to roll-to-roll processing.
Meanwhile, a self-organized small-molecule Phen-NaDPO was provided by Prof. Xuhui Zhu
from South China University of Technology. Different from the strongly hydrophilic interface
materials such as PCBDAN and PEI, Phen-NaDPO exhibits good solubility in weakly polar
solvents, such as chlorobenzene (CB) and o-dichlorobenzene (ODCB). Consequently, no
additional polar solvent is needed in processing the active layer, thus facilitating the large-scale
roll-to-roll production with high reproducibility and reliability.
My focus in this research was the design of a simple approach to printing inverted organic solar
cells with a self-organized charge selective cathode interface layer. Finally, in order to reduce
large area production cost and helpful for commercialization of OSCs, the all-solution process
deposition of interface layers rather than evaporation of that are a way to be considered. I have
104
developed an annealing-free solution-processed s-MoOX:PEG hybrid composites as HTLs in
inverted OSCs. In this chapter, experimental design and all device preparations, processing,
characterization was done solely by me.
Part of this chapter has been published (reproduced with permissions):
H. Zhang, W. Tan, S. Fladischer, L. Ke, T. Ameri, N. Li, M. Turbiez, E. Spiecker, X. Zhu, Y. Cao and C. J. Brabec, J. Mater. Chem. A, 2016, DOI: 10.1039/C6TA00391E
H． Zhang, S. Chen, V. Sgobba, T. Ameri, M. Turbiez, C. J. Brabec, Deposition of annealing-free
solution-processed molybdenum oxide / Poly(ethylene glycol) Hybrids in inverted organic solar cells
by doctor-blading, will be submitted soon
6.1 A simple approach to printing inverted organic solar cells
with a self-organized charge selective cathode interface layer
Generally speaking spin-coating is a batch process and as such incompatible with the
visionary concept of fully printed or coated organic solar modules. As stated in the Chapter 5,
one of the relevant problems in the printed OSCs technology is the formation of a selective
and barrier free extraction contact at the cathode. Therefore, a simple printing approach is
further watched with concern as it may restrict large-area high-speed roll-to-roll
processing.8, 199
In this section, we successfully demonstrate a simple approach to printing efficient, inverted
organic solar cells with a self-organized charge selective cathode interface layer based on
the small-molecule Phen-NaDPO. Different from previous studies, Phen-NaDPO molecules
are blended into a pDPP5T-2/PCBM blend and processed by doctor blading in air (Figure
6-1a). We observe a spontaneous, surface energy driven migration of Phen-NaDPO towards
the ZnO interface and a subsequent formation of electron selective and barrier free
extraction contacts. In the presence of 0.5 wt% Phen-NaDPO, a PCE of 5.4% is achieved for
inverted device based on an ITO/ZnO cathode. The self-organization of Phen-NaDPO
105
through spontaneous vertical phase separation is mainly attributed to its high surface energy
and strong interaction with the cathode material.
6.1.1 The performance of solar cells with Phen-NaDPO as the additive
In some case, most of the EELs are sequentially processed from solution, along with other
function layers, and are fairly thickness-sensitive. Therefore, this thickness-sensitivity is
further watched with concern as it may restrict large-area high-speed roll-to-roll processing.
In order to overcome this issue, the inverted organic solar cells (ITO/ZnO/pDPP5T-
2:PC61BM:Phen-NaDPO/MoO3/Ag) involving a small amount of Phen-NaDPO as the
additive were fabricated (Figure 6-1b, left) and The relevant energy levels are illustrated in
Figure 6-1c. As a versatile cathode modification material, Phen-NaDPO possesses multiple
attractive attributes, combining facile synthesis and purification, a high glass transition
temperature (Tg ≈ 116 °C) and electron mobility (μe = ∼10−4 to 10−3 cm2 V−1 s−1@8 × 105 V
cm−1) with a low-lying HOMO level of −6.1 eV and LUMO level of −2.74 eV.200
a
b
106
Figure 6-1. a) Conceptual diagram for the active layer deposition with a self-organizing EEL by doctor-
blading as well as the chemical structures of the pDPP5T-2, PC61BM and Phen-NaDPO. b) Schematic
illustration of the device configuration of the OSCs with Phen-NaDPO as an additive in the active layer
(left) and as an interlayer between the active layer and ZnO (right). The donor (pDPP5T-2) to acceptor
(PCBM) ratio was kept to 1:1.5 by weight fraction. The active layer thickness was optimized to ~120 nm.
c) Schematic energy level diagram of the functional materials before contact formation (left) and after
contact formation (right). After contact formation, the interaction of Phen-NaDPO with the ZnO substrate
resulted in lowering the vacuum level of the cathode. Reproduced with permission from ref.201. Copyright
© 2016 The Royal Society of Chemistry.
Introducing a small amount of Phen-NaDPO into the pDPP5T-2:PCBM blend results in
considerably increased PCEs as compared to the reference device without the EEL (Figure 6-
2a and Table 6-1). For instance, the reference device without the EEL exhibited a PCE of
4.0% with an open circuit voltage (VOC) of 0.56 V, a short circuit current density (JSC) of
12.12 mA cm2 and a fill factor (FF) of 59.59%. By adding 0.5 wt % Phen-NaDPO into the
semiconductor ink, JSC and FF were increased to 13.51 mA cm2 and 69.06%, respectively,
yielding a PCE of 5.22% while VOC remained unaffected. Upon further increasing the Phen-
NaDPO concentration to 2 wt% JSC continues to rise, but the FF started to drop more rapidly,
resulting in an overall lower PCE of 4.79%.
The external quantum efficiency (EQE) spectra of the solar cells with 0-2 wt % Phen-NaDPO
are presented in Figure 6-2b. The photocurrents calculated by integrating the EQE spectra
with the AM1.5 solar spectrum reproduce the trend observed from j-V analysis and are fully
consistent with the experimental values. The EQE spectra reveal that incorporating the small
amount of Phen-NaDPO as additive in the blend dominantly enhances the response in the
c
107
short wavelength regime (350–450 nm), i.e. where PCBM is absorbing, as well as the long
wavelength regime (600–800 nm). It is noteworthy that the identical trend was observed for
the otherwise identical solar cells with Ba(OH)2 as EEL.37 The analogue with Ba(OH)2
suggests a similar functionality of Phen-NaDPO, namely a more efficient electrode extraction
resulting in reduced 2nd order recombination at the interface. We note that there may be an
effectively self-organized Phen-NaDPO layer on the ZnO substrate.
Figure 6-2. a) j-V characteristics of inverted OSCs (ITO/ZnO/”ternary” active layer/MoO3/Ag) with
different Phen-NaDPO content under simulated AM1.5 illumination (100 mW/cm2). b) EQE spectra of
inverted OSCs corresponding to (a). Reproduced with permission from ref.201. Copyright © 2016 The
Royal Society of Chemistry.
Table 6-1. Summary of the j-V characteristics of inverted OSCs (ITO/ZnO/”ternary” active
layer/MoO3/Ag) with varying Phen-NaDPO content under simulated AM1.5 illumination (100 mW/cm2).a Phen-NaDPO
(wt %)
VOC
(V)
JSC
(mA cm-2
) FF (%)
PCE (%)
avg (best)
0 0.56 -12.12±0.56 59.59±0.50 4.00 (4.30)
0.25 0.56 -13.06±1.01 70.04±0.56 5.11 (5.46)
0.50 0.56 -13.51±0.30 69.06±0.98 5.22 (5.40)
1.00 0.56 -13.83±0.26 66.85±0.85 5.10 (5.17)
2.00 0.56 -14.63±0.76 58.58±1.00 4.79 (4.98)
a Each value represents the average from five cells.
a b
108
Figure 6-3. j-V characteristics of the OSCs (a. ITO/pDPP5T-2:PC61BM:Phen-NaDPO/MoO3/Ag) and (b.
ITO/AZO/pDPP5T-2:PC61BM:Phen-NaDPO/MoO3/Ag) with different Phen-NaDPO contents in the
ternary blend under illumination of an AM 1.5G solar simulator (100 mW/cm2). Reproduced with
permission from ref.201. Copyright © 2016 The Royal Society of Chemistry.
Following a universal strategy, we found that self-organization from “ternary” composites in
the active layer works well for bare ITO and ITO/AZO substrates (See Figure 6-3). However,
it should be noted that the modified ITO electrodes are constantly performing inferior to the
analogue ITO/ZnO or ITO/AZO cathodes, mainly because they are failing to give full VOC.
We speculate that uncomplete electrode coverage may cause the VOC losses via surface
recombination at the electrode.
6.1.2 The performance of solar cells with Phen-NaDPO as a thin interlayer
In order to gain a clear evidence for the observed self-organization, devices with the
architecture ITO/ZnO/Phen-NaDPO/pDPP5T-2:PCBM/MoO3/Ag were fabricated, where
Phen-NaDPO solution was separately bladed as EEL atop ZnO. The j-V characteristics of the
inverted OSCs with different thick Phen-NaDPO layers are summarized in Figure 6-4 and
Table 6-2. The best PCE of 4.80% was achieved at a blading speed of 10 mm/s with a Voc of
0.56 V, JSC of 13.94 mA cm−2 and FF of 61.6%. The interlayer thickness was ca. 8 nm,
subject to possible variation after doctor blading the active layer. Doctor blading methanol for
reference purposes appeared to have no significant influence on the PCE. These results show
that self-organization from a ternary formulation indeed is a plausible scenario as it leads to
generally better PCEs as a dedicated Phen-NaDPO layer on top of ZnO.
a b
109
Regarding the work function of substrates with Phen-NaDPO, Phen-NaDPO was already
shown as a versatile cathode modification layer for Ag, ITO and HOPG due to its strong
interaction with the substrates.200 We studied this effect for various metal oxide substrates
using Kevin probe spectroscopy. As expected, coating a thin layer of Phen-NaDPO decreases
the work functions for ITO, ITO/AZO and ITO/ZnO to 4.01, 3.92 and 3.83 eV, respectively.
Bare ITO cathodes are strongly lacking selectivity and are thus more prone to recombination
losses as compared to ITO/ZnO or ITO/AZO. We therefore varied the Phen-NaDPO layer
thickness for ITO/Phen-NaDPO/pDPP5T-2:PCBM/MoO3/Ag devices as a function of work
function (Figure 6-5). With increasing Phen-NaDPO layer thickness the work function
decreases from 4.80 to 4.01 eV before leveling out around 4.17 eV. Apparently, VOC, JSC and
FF follow the trend of the work function, underpinning the formation of a more and more
charge selective interface.
Table 6-2. Summary of the j-V characteristics of inverted OSCs with differently thick Phen-NaDPO layers
(doctor-blading speed: 5-15 mm/s) under simulated AM1.5 illumination (100 mW/cm2); Methanol was
doctor-bladed on top of the ZnO as reference.a
Devices configuration VOC
(V)
JSC
(mA cm-2
) FF (%)
PCE (%)
avg (best)
ZnO 0.56 -12.12±0.56 59.59±0.50 4.00 (4.30)
ZnO/Phen-NaDPO (5mm/s) 0.56 -13.56±1.07 59.04±1.56 4.48 (4.87) ZnO/Phen-NaDPO
(10mm/s) 0.56 -13.94±0.99 61.60±1.03 4.80 (5.08)
ZnO/Phen-NaDPO
(15mm/s) 0.56 -14.19±0.46 59.75±0.95 4.75 (4.96)
ZnO/MeOH 0.56 -13.26±0.28 56.45±1.14 4.19 (4.30) a Each value represents the average from five cells.
a b
110
Figure 6-4. a) j-V characteristics of inverted OSCs with a different thickness of Phen-NaDPO coated onto
ZnO and under simulated AM1.5 illumination (100 mW/cm2); Methanol was doctor-bladed on top of ZnO
as reference. b) EQE spectra of the inverted OSCs corresponding to (a). Reproduced with permission from
ref.201. Copyright © 2016 The Royal Society of Chemistry.
Figure 6-5．a) j-V characteristics of the OSCs (ITO/Phen-NaDPO/pDPP5T-2:PC61BM/MoO3/Ag) with a
different thickness of Phen-NaDPO under simulated AM 1.5G illumination (100 mW/cm2). b)
Corresponding photovoltaic parameters as a function of the work function for the different Phen-NaDPO
layer thickness. Reproduced with permission from ref.201. Copyright © 2016 The Royal Society of
Chemistry.
6.1.3 The distribution of Phen-NaDPO in the “ternary” active layer
Scanning transmission electron microscopy
To directly evidence the distribution of Phen-NaDPO at the substrate we investigated the
device cross-section by transmission electron microscopy (TEM). Energy dispersive X-ray
spectroscopy (EDXS) in the scanning transmission electron microscopy (STEM) mode
reveals the composition of the single layers and interfaces in OSCs. In this case the
phosphorous signal of Phen-NaDPO should be used to determine the material distribution as
this element is not present in the pDPP5T-2:PCBM blend. STEM EDXS analyses have been
performed at the ZnO/active layer interface of the device with a small amount (~ 2 wt %) of
Phen-NaDPO in the pDPP5T-2:PCBM blend, showing the elemental composition at different
locations of the cross-section (see Figure 6-6a). Note that the small amount of P is observed
around the ZnO interface rather than in the “ternary” blend composite. However, considering
the phosphoric acid ligands of ZnO from NanoGrade, we further investigated the device with
a b
111
a small amount (~ 2 wt %) of Phen-NaDPO in the pDPP5T-2:PCBM blend at the bare ITO
substrate (see Figure 6-6b). In this case no P signal was detected by EDXS, concluding that
the amount of P in the “ternary” blend composite is below the detection limit of this EDXS
system. Therefore we decided to indirectly prove the accumulation of Phen-NaDPO at the
substrate and analyzed the surface energy at the top and bottom surface of the semiconductor.
Figure 6-6. (a) Cross-sectional scanning transmission electron microscopy (STEM) image of the
ZnO/active layer interface. EDX spectra outline the chemical composition at different locations. (b) Cross-
sectional STEM image showing the device architecture, ITO/“ternary” active layer/MoO3/Al, (left). EDX
spectra of “ternary” BHJ composites (right, red) and the interface between ITO and “ternary” blend active
layer (right, blue). Reproduced with permission from ref.201. Copyright © 2016 The Royal Society of
Chemistry.
a
b
112
Surface energy
Self-organization is influenced by the blend composition, substrate and surface energy.197, 202,
203 Therefore we further decided to indirectly prove the accumulation of Phen-NaDPO at the
substrate and analyzed the surface energy at the top and bottom surface of the semiconductor.
The surface energies of ITO, pDPP5T-2, PCBM, Phen-NaDPO, pDPP5T-2:PCBM and
pDPP5T-2:PCBM:Phen-NaDPO were determined by measuring the droplet´s contact angle
for three liquids (ultrapure water, diiodomethane and ethylene glycol) using the Owens–
Wendt and Kaelble method (Figure 6-7a). The ITO and Phen-NaDPO surfaces exhibit rather
high surface energies (γITO = 62.32 mN/m and γPhen-NaDPO = 49.81 mN/m) originating from
their terminal hydroxyls, phosphine oxide and phenanthroline moieties. In contrast, pristine
pDPP5T-2 and PCBM show surface energies of 28.34 and 34.78 mN/m, respectively. The
surface energies of the composite films pDPP5T-2:PCBM and pDPP5T-2:PCBM:Phen-
NaDPO are found to be 28.37 and 26.61 mN/m, respectively, which are almost identical to
that of the pristine pDPP5T-2. The pDPP5T-2 and PCBM exhibit low surface energies due to
their hydrophobic diketopyrrolopyrrole segment and alkyl side chains. Similar values were
reported for other blends previously.197, 198
To further evidence the accumulation of Phen-NaDPO at the bottom contact, we peeled off
the substrate to expose the bottom contact, i.e. the bottom surface of the “ternary” blend film
(schematically illustrated in Figure 6-7b). As shown in Figure 6-7c, the top surface of the
“ternary” blend film is of hydrophobic nature with a water contact angle of ≈ 99.0°, similar to
the value for the top surface of the binary blend film (water contact angle ≈ 100.6°). Pristine
ITO and Phen-NaDPO are hydrophilic (water contact angle ≈ 24.1° and 58.8° respectively).
The bottom surface of the “binary” film is hydrophobic with a contact angle of ≈ 108.2°
which is significantly reduced to 87.8° upon addition of Phen-NaDPO for the “ternary” blend
film. These measurements clearly reveal that the “ternary” system is minimizing its free
energy by assembling the high-surface energy components (i.e. Phen-NaDPO) at the
substrate while the low-surface energy components (i.e. pDPP5T-2 and PCBM) accumulate
at the surface. This is further in excellent agreement with the observation that the surface
energy of pDPP5T-2:PCBM:Phen-NaDPO is nearly identical to the one of pDPP5T-2:PCBM,
suggesting that no hydrophilic Phen-NaDPO is existent at the surface of the film.
113
Figure 6-7d illustrates the self-organized molecules of Phen-NaDPO at the ITO interface,
which show potential interactions with the surface metal ions and the residual hydroxyl
groups, leading to reduction of the work function of the ITO substrate. Furthermore, the low
HOMO of Phen-NaDPO may benefit to reduce the electron-hole recombination at the
cathode.
Figure 6-7. a) The calculated surface energies of materials studied in this work. b) A schematic of the
peeling process of the “ternary” active layer blend film to expose the front contact. c) Photographs of water
droplets on the front / back substrate surfaces coated with films of “binary”, “ternary” active layer blend
film, the ITO, pDPP5T-2, PC61BM and Phen-NaDPO films, respectively. d) Illustration of the self-
a
b
c
top bottom
d
114
organized molecules of Phen-NaDPO at the ITO interface, which show potential interactions with the
surface metal ions and the residual hydroxyl groups. Reproduced with permission from ref.201. Copyright
© 2016 The Royal Society of Chemistry.
6.2 Deposition of annealing-free MoOx/PEG Hybrids in inverted
organic solar cells by doctor-blading
As previously stated, developing a stable solution-process for large area coatings with high
average efficiencies and reliability is still one of the major challenges for the photovoltaic
technology. In this section, we successfully employed annealing-free solution-processed s-
MoOX:PEG hybrid composites as HTLs in inverted organic solar cells, being deposited via a
doctor-blading processes under environmental conditions. Inverted devices with such s-
MoOX:PEG as HTLs show a performance comparable to those with evaporated MoOX.
Furthermore, s-MoOX:PEG improves the device stability in air. Excellent wetting of s-
MoOX:PEG solution on the active layer leads to highly uniform layers with complete surface
coverage and superior hole selectivity. As an alternative to PEDOT:PSS and evaporated
MoOX, s-MoOX:PEG is established as a highly promising hole transport material for efficient
and stable inverted organic solar cells that is fully compatible to roll-to-roll processing.
6.2.1 Comparison of s-MoOX, PEDOT:PSS and e-MoOX as HTLs in
inverted OSCs
The inverted device architecture is shown in Figure 6-8d. In this section, we also used a
blend of pDPP5T-2 and PCBM as our model for photovoltaic devices. Polymers containing
diketopyrrolopyrrole (DPP) units in the main chain attract more and more attention in the
research community owing to their ideal low bandgap, annealing-free and air-stability for
roll-to-roll productions. The j-V characteristics of inverted OSCs based on different HTLs
are shown in Figure 6-8a, and a summary of the device performance is tabulated in Table 6-
3. Two kinds of devices #1 and #2 were compared embedding evaporated or pristine
solution-processed MoOX as HTLs, respectively. The s-MoOX film was deposited on the
115
active layer of inverted OSCs at room-temperature by doctor-blading in air. As shown in
Table 6-3 and Figure 6-8a, the device #2 using optimized s-MoOX as HTL showed an open
circuit voltage (VOC) of 0.55 V, a JSC of 12.71 mA cm‑2, a FF of 55.55% and a PCE of 3.88%.
Compared to the device #1 with e-MoOX, the solar cells with the s-MoOx thin film have 8%
losses of JSC and 12% losses of FF, thus performance is lower than that of device #1.
PEDOT:PSS is the most commonly used hole transport layer in the OSCs. Devices #6 was
constructed as another control devices employing PEDOT:PSS as HTL under identical
conditions at room-temperature by doctor-blading in air. The PCE of inverted OSCs with the
optimized PEDOT:PSS is only 1.55%, which does not match with the PCEs of device #1
with e-MoO3 and #2 with s-MoOX. Since annealing of the PEDOT:PSS is found to have a
significant effect on device. PEDOT:PSS is a highly hygroscopic material retaining a large
fraction of water (10-15wt%) ， which need annealing to remove water and optimize
morphology.204, 205 However for annealing-free polymer materials such like pDPP5T-2,37, 206
cannot sustain PEDOT:PSS annealing temperature (120℃). Therefore, PEDOT:PSS is
incompatibility as HTL for the inverted OSCs.
a b
c d
116
Figure 6-8. a) j-V characteristics of inverted organic solar cells with various hole transport layers under
illumination of an AM 1.5G solar simulator (100 mW cm-2). b) Corresponding logarithmic plot of dark j-V
characteristics. c) EQE spectra of inverted organic solar cells corresponding to (a). PEDOT:PSS was
doctor-bladed on top of the active layer for comparison. d) Schematic of the inverted organic solar cells
structure in this work.
6.2.2 s-MoOX/PEG hybrids as HTL in inverted OSCs
Because of different drying kinetics of MoO3 films on the active layer, OSCs solution
processed by doctor-blading are still difficult to achieve comparable morphology with those
processed by spin-coating.177 The homogeneity of MoOX films during processing from a
homogeneous coating ink on the active layer of inverted OSCs should be controlled by
material properties and processing conditions and determines the efficiency of OSCs.
To achieve further performance improvement of our room-temperature s-MoOx to match
with e-MoOX, the s-MoOX/PEG hybrids were employed as the HTL in inverted pDPP5T-
2:PCBM OSCs, which did not require any post-treatment. The tuning of concentration of
PEG in s-MoOX/PEG hybrids was thoroughly investigated as a function of the effect of
hybridization on the device performance. The j-V characteristics for typical inverted OSCs
based on different concentration of PEG in s-MoOX/PEG hybrids as HTL were shown in
Figure 6-8a, and a summary of the devices #3-#5 performance were tabulated in Table 6-3.
From the j-V curves, it was found that the best performing OSC was based on the s-
MoOX:2.5wt% PEG hybrids as HTL which matched with that of the device based on e-MoOx.
As PEG contents increased in the s-MoOX:PEG hybrids, the device performance was
dramatically changed. The Jsc and FF were significantly enhanced to 13.99 mA cm‑2 and
64.00% with a PCE of 4.92% at 2.5 wt% of PEG in the s-MoOX:PEG hybrids. When the PEG
contents is up to 5wt% in the hybrids, JSC and FF dropped to 11.46 mA cm-2 and 59.05%,
respectively. The Jsc increased upon concentration of PEG in s-MoOX:PEG hybrids, in
accordance with the change in the EQE spectra showed in Figure 5-8c. The photocurrent
integrated from the EQE spectra are 11.50, 11.71, 13.41 and 10.55 mA cm-2 for the devices
with 0, 1, 2.5 and 5 wt % PEG in s-MoOX:PEG hybrids, respectively. The higher EQE values
of the devices with s-MoOX:2.5wt% PEG hybrids as HTL were dominantly in the absorption
regimes between 425–550 nm and 625–800 nm. The increased of EQE spectral response was
117
due to successfully reducing recombination at or around the interface between the HTL and
active layer, thus increase charge collection. Maybe several factors play a role there – (1) a
reduction of the extraction barrier, (2) a change of the HTL morphology, as will be described
later. From the dark j-V characteristics in Figure 5-8b, it was found that the leakage current
of the device with s-MoOX:2.5wt% PEG hybrids as HTL is considerably restrained,
indicating that the recombination of carriers was suppressed. With the increasing of the PEG
contents in the s-MoOX:PEG hybrids, the injected current density of devices with s-
MoOX:5wt% PEG hybrids in the dark (0.5 V – 2.0V) is decreased due to the insulating
nature of PEG.207 PEG overdose resulted in lowered conductivity of s-MoOX:PEG film, thus
reducing the charge transportation.
Table 6-3. Summary of the j-V characteristics of inverted organic solar cells with various hole transport
layers under illumination of an AM 1.5G solar simulator (100 mW cm-2).a
Devices hole transport layer VOC
(V)
JSC
(mA cm-2
) FF (%) avg PCE (%)
#1 e-MoOX 0.55 -13.84±0.30 63.10±1.50 4.80 #2 s-MoOX 0.55 -12.71±0.57 55.55±3.56 3.88 #3 s-MoOX:1wt% PEG 0.55 -13.07±0.23 61.60±2.93 4.43 #4 s-MoOX:2.5wt% PEG 0.55 -13.99±0.46 64.00±0.95 4.92 #5 s-MoOX:5wt% PEG 0.55 -11.46±0.38 59.05±2.04 3.72 #6 PEDOT:PSS 0.53 -9.30±0.68 31.28±3.87 1.55
a Each value represents the average from five cells.
6.2.3 Surface coverage and the morphology of s-MoOX/PEG film on active
layer
Surface coverage and the morphology of HTLs film on active layer are crucial for
determining resultant performance and stability of inverted OSCs.208 To investigate the
homogeneity and surface roughness of HTLs on the active layers, the AFM images obtained
from these e-MoOX, s-MoOX and s-MoOX:PEG films were shown in Figure 6-9. The e-
MoOx film shows a smooth surface with root-mean-square (RMS) roughness value of 1.79
nm, while the absence of color contrast in the phase images suggests a good and
homogeneous overlay of e-MoOx on the active layer. Figure 6-9 compared the surface
118
morphologies of s-MoOX and s-MoOX:PEG films on the top of active layer. For images
recorded over 3 µm × 3 µm, we find that RMS decreases from 28.5 nm to 17.7 and 11.5 nm
upon addition of 2.5wt% and 5.0wt% PEG, respectively. Addition of PEG in the s-MoOX
under the experimental conditions described here also resulted in decreased roughness and
the formation of dot-like nanoscale features.
Figure 6-9. AFM topographical height (left) and phase (right) images (size: 3 μm × 3 μm) of (a,b) after
deposition of e-MoOX on the active layer, c,d) after deposition of s-MoOX on the active layer, (e,f) after
deposition of s-MoOX:2.5wt% PEG on the active layer, (g,h) after deposition of s-MoOX:5wt% PEG on the
active layer.
In the phase images (Figure 6-9d), we can see that the entire surface of active layer is not
covered. Large aggregates of s-MoOX on the surface of active layer reaching heights of over
200 nm lead to short-circuited photovoltaic devices. A too rough surface morphology may
induce poorer contact between the HTL and Ag. This is in agreement with losses in Jsc and
FF for the s-MoOx HTL device. When the addition of 2.5wt% PEG into s-MoOX 6see Figure
119
5-9e and f), the surface of the s-MoOX:PEG film shows a more uniform distribution of s-
MoOX, and the surface of active layer is covered by a compact s-MoOX:PEG film, which is
beneficial to the charge transportation and collection, thus enhancing the Jsc and FF. It is
obvious from the result that the surfactant PEG can prevent the s-MoOX aggregation to bulk
material even in films during the doctor-blading processes in air. As PEG contents increased
in the s-MoOX:PEG hybrids, s-MoOX:PEG film is more homogeneous and compact
obviously. However, higher amounts of PEG in the s-MoOX:PEG lead to the decreased
efficiency of devices.
As previously stated in Section 6.1, materials as additives can turn the surface energy of the
film. Measurements of the water contact angle (θ) were performed on the surface of s-MoOX,
s-MoOX:PEG, PEG and active layer, and the images were collected with a digital camera. As
shown in Figure 6-10, the pDPP5T-2:PCBM surface (θ≈100.8°) was largely hydrophobic. In
contrast, the surface of s-MoOX (θ≈5°) was extremely hydrophilic. Due to the moderately
hydrophilic property of PEG surface (θ≈41.1°), upon concentration of PEG in s-MoOX:PEG
hybrids, values of water contact angle increase to 10.9°, 16.9° and 20.0° for surfaces of s-
MoOX:1wt% PEG, s-MoOX:2.5wt% PEG and s-MoOX:5wt% PEG, respectively. The
decreased hydrophilic property of s-MoOX:PEG surface can enhance the wettability of the s-
MoOX on the surface of the active layer, which benefit to the wet film uniform Spreading.209
Then, the wet s-MoOX:PEG film is dried on the active layer to form a homogeneous and
compact HTL, with the heat of a doctor-blading equipment. Figure 6-11 illustrates the
Conceptual diagram of the Surface coverage and the morphology for the s-MoOX film on
active layer deposition with or without PEG by doctor-blading in air.
120
Figure 6-10. Photographs of water droplets on the surfaces of various films
Figure 6-11. Conceptual diagram of the surface coverage and the morphology for the s-MoOX film on
active layer deposition with or without PEG by doctor-blading in air
6.2.4 The environmental stability of devices with s-MoOX/PEG
In the last part we explore the impact of the various HTLs on the environmental stability. The
air-stability of the un-encapsulated inverted OSCs with s-MoOX:PEG were compared and
periodically measured under humidity of 55% and 25 ℃ for 2500 hours (see Figure 6-12).
Oxygen and moisture at the surface of HTL are known to alter the electronic properties and
energetics at the HTL/active layer and HTL/metal electrode interface and thus may lead to an
interfacial barrier over the time of the degradation.208, 210, 211 Moreover, ambient water
entering the solar cell can be absorbed by the hygroscopic water based HTLs leading to
further degradation of the HTL interface.212 The inverted devices with e-MoOX maintain
more or less 40% of their original PCE after being exposed to ambient conditions. In parallel
devices with s-MoOX retain 50% of their original PCE, whereas the devices with s-
121
MoOX:PEG exhibit better long-term stability and maintain up to 60% of the initial efficiency.
The changes in normalized JSC, VOC and FF over the period of keeping in air are shown in
Figure 6-12. The regular decay is mainly found in a drop of JSC and FF due to a porosity of
the s-MoOX film on the active layer, meaning it can deliver water from the environment to
the active layer causing rapid degradation of the device. With addition of PEG in the s-MoOX,
a compact s-MoOX:PEG HTL prevent water from passing through the pores of the e-MoOX
or s-MoOX film, due to the slight hydrophobic property of PEG and suppressing s-MoOX to
form large aggregates. VOC remained almost unchanged during ambient condition
degradation. This comparison demonstrates the more stability of our s-MoOX:PEG compared
to the e-MoOX and s-MoOX used HILs.
Figure 6-12. Degradation trends of the photovoltaic parameters over storage time for unencapsulated
devices with different HILs in ambient air under ambient conditions (humidity of 55% and 25℃). (a)
Normalized VOC, (b) Normalized JSC, (c) Normalized FF, (d) Normalized PCE. The average and
deviations are calculated from 5 devices.
a b
c d
122
6.3 Conclusions
In summary, we have successfully proved that the self-assembly and self-formation of
electron-extraction layers from “ternary” blend solutions is compatible with roll to roll
processing strategies. A simple approach of fabricating efficient inverted OSCs with a self-
organizing small-molecule electron extraction layer Phen-NaDPO by doctor-blading was
demonstrated on ITO, ITO/ZnO and ITO/AZO. A PCE of 5.4% was achieved for the inverted
device based on an ITO/ZnO cathode. The photovoltaic performance remained quite stable
for a rather wide range of Phen-NaDPO concentrations in the active layer blend, thus offering
tremendous prospects for a stable and reliable process development. The self-organization of
Phen-NaDPO through vertical phase separation is attributed to its high surface energy and
strong interactions with the cathode material. The present study shows that utilizing self-
organization processes of proper interfacial materials from “ternary” active layer
formulations indeed is a promising strategy towards simplified roll-to-roll processing of
organic solar cells.
For a stable process for large area coatings with all solution-processed at room-temperature,
we have successfully employed annealing-free solution-processed s-MoOX:PEG hybrid as the
HTL in inverted organic solar cells during the doctor-blading processes in air. Excellent
wetting of s-MoOX:PEG solution on the active layer leads to uniform s-MoOX:PEG film with
complete surface coverage and superior hole selectivity for facilitating hole transport from
active layer to the Ag electrode, due to the moderately hydrophilic property of PEG. Inverted
devices with such s-MoOX:PEG as HTLs have performance comparable to those with
evaporated MoOX. Furthermore, s-MoOX:PEG improved the device stability in air. A
compact s-MoOX:PEG HTL prevent water and oxygen from the environment to destroy the
active layer. As an alternative to PEDOT:PSS and evaporated MoOX, s-MoOX:PEG is a
promising hole transport material for highly efficient and stable inverted organic solar cells
that can be used on flexible substrates via roll-to-roll processing. This strategy also offers a
promising and practical approach to improve the device stability on an industrial fabrication
level.
123
Chapter 7
Interfacial Engineering
for Perovskite Solar Cells
Along with the development of perovskite solar cells, significant progress has been realized in
perovskite photovoltaic devices with efficiencies over 20%. However, understanding of the
interface contacts of perovskite-organic and organic-electrode are deficient.
To minimize the contact barrier, the interface between the active layer and the electrodes should
be ohmic. J. H. Seo et al.143
demonstrated that it is possible to incorporate conjugated
polyelectrolyte layers to improve OSCs with very good efficiencies. Rationalizing interface dipole
of ultrathin conjugated polyelectrolyte I was stimulated to study that concept in great detail for
perovskite solar cells. Inserting an ultrathin polyelectrolyte layer, either PEIE or P3TMAHT
(provided by Prof. Ullrich Scherf from IfP, Bergische Universität Wuppertal), between the Ag
electrode and the PCBM layer, I have fabricated fairly efficient perovskite solar cells. The
suitability of the interface opens the opportunity to apply the interface design rules for the
organic solar cell to the perovskite solar cell technology. In this research, all device preparations,
processing, characterization was done solely by me.
In addition, spiro-MeOTAD is expensive and requires a relatively complex doping strategy,
which questions the reproducibility and standing time of this ink at industrial scales. Based on
an industrial point of view, I further decided to look into alternatives to spiro-MeOTAD, which is
stable and cheaper for perovskite solar cells at industrial scales. Using a novel water-free
PEDOT, which was provided by Heraeus, we fabricated a fairly simple device architecture of
ITO/LT-TiO2/Perovskite/Water-free PEDOT/Au, and a power conversion efficiency of up to
11.75% was demonstrated, along with 800 hours environmentally stability in air without any
encapsulation. In this research, I and colleague, Y. Hou, conceived and designed the
experiments. All device preparations, processing, characterization was done by both of us. Y. H.
carried out the optical characterizations and device structure optimization.
124
Part of this chapter has been published (reproduced with permissions):
H. Zhang, H. Azimi, Y. Hou, T. Ameri, T. Przybilla, E. Spiecker, M. Kraft, U. Scherf, and C. J. Brabec, Chemistry of Materials 2014 26 (18), 5190-5193
Y. Hou†, H. Zhang†, W. Chen, S. Chen, C.O. Ramirez-Quiroz, H. Azimi, A. Osvet, G. J. Matt, E. Zeira, N. Kausch-Busies, W. Lövenich, C. J. Brabec, Adv. Energy Mater., 2015 5 (15), . doi: 10.1002/aenm.201500543 († These authors contributed equally to this work)
7.1 A brief introduction of perovskite solar cells
Technically, perovskite is a type of mineral that was first found in the Ural Mountains. In
2009, Miyasaka et al.213 reported the first attempt using perovskite as sensitizers incorporated
into dye-sensitized solar cells, which represents a revolutionary step in perovskite
photovoltaic technology. Organolead halide perovskite materials (see Figure 7-1), ABX3 (A
= CH3NH3 or NHCHNH3, B = Pb, and X = Br, Cl, or I) are the subject of extensive
investigations. This kind of perovskite materials offer a broad range of attractive features
such as a direct optical bandgap, a low exciton bonding energy and long diffusion length, a
broad range of light absorption, excellent carrier transport and crystallinity.214-216 Recently,
significant progress has been realized in organolead halide perovskite photovoltaic devices
with efficiencies over 20%31, attracting tremendous attention in the photovoltaic industry.
However, the fabrication of most efficient perovskite solar cells is typically performed based
on employing a high-quality condensed TiO2 layer as electron transporting layer, which
requires high-temperature processing (450°C) for a long time (1 hour).217, 218 Until now, the
fabrication of high performance perovskite solar cells typically relies on either expensive
organic hole transport materials and or high-temperature sintered TiO2 electron transport
layers, which greatly increases the materials and manufacturing costs of the final devices.
Regarding the large scale fabrication of perovskite solar cells, it is among the biggest
challenges to be able to combine the advantages of low-temperature solution processing on
one hand with low-cost abundant raw materials on the other hand into efficient and stable
performances within a single solar cell.
With a development direction of efficient perovskite solar cells, devices can be also
125
fabricated using a hybrid planar heterojunction (PHJ), in which the perovskite layer is
sandwiched between a hole-transport layer, such as PEDOT:PSS, and an electron-transport
layer, such as PCBM. The advantages of this structure are the simplicity as well as low-
temperature solution-processability. Usually, the barrier at the contact interface between a
Fermi level of various electrode metals (e.g. Ag, Au) and the LUMO of the organic material
(PCBM) still exists in organic optoelectronic devices, leading to poor electron injection and
extraction.143, 219 The charge injection and extraction at the metal-organic semiconductor
interface has a significant impact on the electrical properties of the semiconductor devices.
To minimize the contact barrier, the interface between the metal electrode and the PCBM
layer should be a quasi-ohmic contact. The suitability of the interface opens the opportunity
to apply the interface design rules for the organic solar cell to the perovskite solar cell
technology. That strategy has led to efforts in interfacial engineering of hybrid
organic/inorganic perovskite PHJ solar cells. High fill factors were demonstrated with e.g. the
use of thermal vapour deposited LiF220, bathocuproine (BCP)221, 222 or fullerene (C60)221 on
top of the PCBM layer. All these systems improved the contact properties and enhanced the
device efficiency. However, the thermal vapour deposition of an interface layer is the
contradiction to the concept of large scale fabrication and cost-effective solution processing.
The identical strategy from organic solar cells should be an alternative pathway to engineer
the interface of hybrid organic/inorganic perovskite PHJ solar cells. 143, 222
Figure 7-1.Structure of a generic Perovskite crystal. Reproduced from ref 223.
126
In addition, as previously stated, the expensive HTM, the commonly used doped spiro-
MeOTAD, which greatly increase the materials and manufacturing cost, is one of the
limitations. As another development direction of efficient perovskite solar cells, the
development of a low-cost HTM, which is stable in its solution form, is therefore among the
most important issues to evolve this solar cell technology from lab to industry.
7.2 Polyelectrolyte Interlayers for Perovskite PHJ Solar Cells
As previously stated, to minimize the contact barrier, the interface between the metal
electrode and the PCBM layer should be a quasi-ohmic contact. In addition, The identical
strategy from organic solar cells should be an alternative pathway to engineer the interface
of hybrid organic/inorganic perovskite planar heterojunction (PHJ) solar cells.143, 222
In this section, highly efficient hybrid organic/inorganic perovskite PHJ solar cells (ITO /
PEDOT:PSS / CH3NH3PbI3-XClX / PCBM / polyelectrolyte interlayer /Ag) are fabricated
based on the interface layers solution-processed at low temperature. Relative to the control
device, the power conversion efficiency (PCE) increased significantly from 8.53% for the
control device to 12.01% (PEIE) and 11.28% (P3TMAHT) via incorporation of a
polyelectrolyte interlayer. The improvement in PCE for devices is chiefly assigned to the
effective influence of polyelectrolyte interlayers on reducing the work function of the
subsequently deposited metal electrode, thereby lowering the electron-injection barriers.
7.2.1 The device structure of the perovskite PHJ solar cell
Figure 7-2a shows the molecular structures of the interlayer materials, PEIE and P3TMAHT,
and the device configuration of the hybrid organic/inorganic perovskite PHJ solar cell. The
polyelectrolyte interlayers were subsequently deposited by spin-casting from PEIE (0.2% w/v)
and P3TMAHT (0.01% w/v) solutions in methanol. A cross-sectional scanning electron
microscopy (SEM) image of the device without the Ag electrode is shown in Figure 7-2b.
The cross-section was prepared using a focused ion beam (FEI Helios NanoLab 660)
127
operating at 30 kV and subsequential imaged with the electron beam of the same instrument
using an accelerating voltage of 2 kV. The perovskite layer has an average film thickness of
approximately 180 nm, and an average film thickness of approximately 250 nm was
determined by profilometer measurements. The smooth CH3NH3PbI3-XClX film covers the
surface of the PEDOT:PSS layer perfectly (see Figure 7-2d). Subsequently, the 65-nm-thick
PCBM layer is able to deposit on CH3NH3PbI3-XClX surface. It can be seen that a 65 nm
PCBM layer does perfectly cover the entire CH3NH3PbI3-XClX surface, which profits from a
smooth perovskite film.
Figure 7-2. a) Schematic of the hybrid organic/inorganic perovskite PHJ solar cell structure and chemical
structures of the PEIE and P3TMAHT. b) Cross-sectional SEM image showing the device structure of the
hybrid organic/inorganic perovskite PHJ solar cell without Ag electrode. The thickness of PEDOT:PSS,
CH3NH3PbI3-XClX, and PCBM layers are 50, 180, and 65 nm, respectively. c) Schematic energy level
diagrams of devices with and without the polyelectrolyte interlayer under flat band conditions. d) A
photograph of the smooth perovskite film. Reproduced with permission from ref.224, Copyright © 2014,
American Chemical Society.
When PEIE and P3TMAHT are respectively spin-coated on top of Ag electrode, the work
function of Ag decreases from 4.70 eV to 3.97 eV (PEIE) and 4.13 eV (P3TMAHT) by the
formation of a surface dipole, which is attributed to the “push-back” or “cushion” effect,143,
a b
c
d
128
157, 219, 225 according to the results of Kelvin probe microscopy (KPM) measurements. The
interfacial modification of PEIE or P3TMAHT/Ag interfaces was already reported for
organic photoelectric devices. The energy level diagrams of devices with and without the
polyelectrolyte interlayer under flat band conditions are illustrated in Figure 7-2c. The
interface dipoles (PEIE or P3TMAHT) with a negative charge toward the metals and the
corresponding positive charge toward the PCBM layer result in the lowering of the vacuum
level of the metal.143, 219 Under irradiation, free charge carriers are generated in the
CH3NH3PbI3-XClX layer. The oppositely charged holes and electrons can be extracted and
transferred by the PEDOT:PSS and the PCBM, respectively. The ITO anode collects holes
and the Ag cathode collects the electrons. The modifying work function of Ag and PCBM
interface minimizes electrical losses upon injection or extraction of electrons.
7.2.2 Electron injection efficiency at PCBM/Ag interface
The interfacial energy level of PEIE or P3TMAHT/Ag interfaces should match with PCBM,
resulting in effective electronic extraction in OSCs. To study the effects of electron injection
efficiency at PCBM/Ag interface, we fabricated electron-only devices with the following
device configuration: ITO/ZnO/PCBM/ polyelectrolyte interlayers/Ag. As shown in Figure 7-
3d, the electron current density significantly increased (up to 3 times) for PEIE or P3TMAHT
based devices when injected from the polyelectrolyte interlayer/Ag electrode, but remained
the same when injected from the ITO/ZnO electrode. This asymmetric improvement confirms
that the increase in electron current density is due to improved injection efficiency of the Ag
electrode. It is also clear that the polyelectrolyte interlayers improved the electron injection
current for Ag electrode, compared to the control devices. This is also in agreement with a
reduced injection barrier at PCBM/Ag interface. In parallel, in a methanol-treated device, the
j-V characteristics remain nearly symmetric, with the current injected from the Ag electrode
only being slightly higher than that injected from ZnO/ITO. The effects of methanol
treatment are shown in a slightly enhanced charge-transport property, a slightly accelerated
and enlarged charge injection. This result confirms that, the substantial positive effect of
methanol treatment on injection efficiency was already reported for organic solar cells.
129
7.2.3 The device performance of the perovskite PHJ solar cell
The j-V characteristics of perovskite PHJ solar cells based on different polyelectrolyte
interlayers are shown in Figure 7-3, and a summary of the device performance is tabulated in
Table 7-1. The control device without a polyelectrolyte interlayer exhibits an open circuit
voltage (VOC) of 0.849 V, a short circuit current density (JSC) of 16.00 mA cm-2, a fill factor
(FF) of 61.29% and a power conversion efficiency (PCE) of 8.53%. The series resistance (RS)
is 7.96 Ω cm2 and the shunt resistance (RShunt) is 0.90 KΩ cm2. As shown in Figure 7-3a, the
control device suggested a strong s-shaped j-V curve under illumination, which means that a
contact barrier prevents electron injection at the PCBM/Ag interface. Such an injection
barrier primarily implies an increased series resistance. The PCE increases significantly from
8.53% for the control device to 12.01% (PEIE) and 11.28% (P3TMAHT) when the
polyelectrolyte interlayer is inserted between PCBM layer and Ag electrode. The
improvement is mainly due to an increased FF. Relative to the control device, the
incorporation of the polyelectrolyte interlayer gave rise to an increase in Jsc from 16.00
mA/cm2 to 17.32 mA/cm2 (PEIE) and 17.10 mA/cm2 (P3TMAHT). Similarly, Voc enhanced
from 0.849 V to 0.899 V (PEIE) and 0.899 V (P3TMAHT) and FF from 61.29% to 77.10%
(PEIE) and 74.10% (P3TMAHT). The series resistance decreased from 7.96 Ω cm2 to 1.00 Ω
cm2 (PEIE) and 0.99 Ω cm2 (P3TMAHT), obviously due to the reduction of the injection
barrier between the PCBM layer and the Ag electrode, and is in line with an improved FF. To
study the impact of methanol processing on the CH3NH3PbI3-XClX/PCBM stack, we
fabricated another type of reference device, which involved deposition of methanol on the top
of PCBM layer by a sequence of steps similar to those for the polyelectrolyte interlayer
devices described above. As shown by the green line in Figure 7-3 and the data summarized
in Table 7-1, methanol-treated devices show negligible impact with Voc and Jsc values
remaining unchanged and FF values slightly increases. This clearly indicated that the
improvement in performance after PEIE or P3TMAHT insertion is due to the presence of the
thin polyelectrolyte interlayers; methanol processing has no effect on the performance of
devices. Figure 7-3b shows that the dark current densities with the thin polyelectrolyte
interlayers were significantly suppressed in the regime from -1 to 0.8 V, consistent with a
130
reduced leakage current density and an increased shunt resistance (Rsh) (0.90, 8.18, and 3.14
kΩ cm2 for the control, PEIE, and P3TMAHT, respectively). In the regime from 0.8 to 2 V,
the injected current density is higher with than without the polyelectrolyte interlayers. This is
in agreement with a reduced injection barrier upon inserting a polyelectrolyte interlayer.
The external quantum efficiency (EQE) curves of devices with and without different
polyelectrolyte interlayers are shown in Figure 7-3c. The improved PCEs of the devices with
PEIE or P3TMAHT layers are also consistent with the higher EQE values. The photocurrent
integrated from the EQE data are 15.2, 15.1, 16.7 and 16.5 mA/cm2 for the control, methanol
treated, PEIE, and P3TMAHT, respectively. The hybrid organic/inorganic perovskite PHJ
solar cells exhibit a spectral response from the visible to near-infrared wavelength regime
(300 to 800 nm) with a broad and flat peak around 70∼80% at approximately 400∼750 nm.
The higher EQE values of the device with PEIE or P3TMAHT layer in the visible to near-
infrared wavelength regime suggest that the polyelectrolyte interlayers more efficiently
collect electrons of the PCBM/Ag electrode, on account of successfully reducing energy
barrier at or around the interface between the PCBM and Ag resulting from the generation of
a surface dipole. Both effects, the surface dipole layers as well as the better protected
interface (reduced shunt) are in excellent agreement with the experimental findings.
Table 7-1. Key values of the j-V characteristics of hybrid organic/inorganic perovskite PHJ solar cell with
and without polyelectrolyte interlayers under illumination of an AM 1.5G solar simulator (100 mW/cm2).
Devices cathode
configuration
VOC
(mV)
JSC
(mA cm-2
)
FF
(%)
PCE (%) RS
(Ω
cm2)
RSh
(kΩ
cm2)
avg (best)
PCBM/Ag 849±1 -16.00±0.48 60.29±2.30 8.53 (8.82) 7.96 0.90 PCBM/methanol/Ag 849±1 -16.10±0.67 61.75±6.25 7.97 (8.92) 0.14 0.32
PCBM/PEIE/Ag 899±1 -17.32±0.31 77.10±1.27 12.01 (12.36) 1.00 8.18 PCBM/P3TMAHT/Ag 899±1 -17.10±0.42 74.10±1.34 11.28 (11.88) 0.99 3.14
131
Figure 7-3. a) j-V characteristics of the hybrid organic/inorganic perovskite PHJ solar cells without
interlayer (black) and with thin layers of PEIE (red) and P3TMAHT (blue) under illumination of an AM
1.5G solar simulator (100 mW/cm2). Methanol (green) was spin-cast on top of the PCBM layer. b)
Corresponding logarithmic plot of dark j-V characteristics. c) EQE spectra of perovskite solar cells
corresponding to (a); the integrated short circuit current is 15.2, 15.1, 16.7 and 16.5 mA/cm2, respectively.
d) Current density versus voltage characteristics of ITO / ZnO / PCBM (65nm) / with and without
polyelectrolyte interlayers /Ag electron-only devices. Inset: corresponding logarithmic plot of current
density versus voltage characteristics of electron-only devices. Reproduced with permission from ref.224,
Copyright © 2014, American Chemical Society.
7.2.4 Surface coverage of the perovskite films
It is well known that surface morphology of the perovskite layer do play a crucial role in
determining the ultimate device performance. Figure 7-4 shows the surface morphologies
obtained by atomic force microscopy (AFM). The surface of the pristine CH3NH3PbI3-XClX
layer is relatively smooth, with a root mean square (RMS) roughness of about 16.7 nm in an
area of 10 µm × 10 µm (see Figure 7-4a). After deposition of PCBM, the surface was much
a b
c d
132
smoother (rms roughness = 6.1 nm) and remained homogeneous, as shown in Figure 7-4b.
The perovskite film formed has a very low roughness, which allows full surface coverage
with a very thin PCBM layer (approximately 65 nm thick). Notably, full surface coverage of
the perovskite film by the PCBM layer is crucial to prevent leakage due to direct contact
between the metal and the perovskite film.226, 227 Further important point is that the full
surface coverage by the PCBM layer can prevent migration of the polar solvent such as
methanol to perovskite layer. We found this is important to ensure good reproducibility of the
device performance. To study the possible effect of the methanol solvent on perovskite film,
we fabricated devices with the following device structure: ITO/PEDOT:PSS/CH3NH3PbI3-
XClX/methanol/PCBM/Ag. We also found that the fabricated perovskite film turned dark
brown to light brown after treatment with methanol. Since CH3NH3I is re-dissolved in
methanol. The RMS roughness of the PCBM/CH3NH3PbI3-XClX layer after treatment with
methanol is 16.4 nm. The RMS roughness of a PEIE layer and a P3TMAHT layer deposited
on the PCBM / CH3NH3PbI3-XClX planar heterojunction films are 8.4 and 10.3 nm,
respectively.
Figure 7-4. Surface topographic AFM images images (size: 10 × 10 μm2) of a) the pristine CH3NH3PbI3-
XClX perovskite film, b) after deposition of PCBM on the CH3NH3PbI3-XClX perovskite film, c) the
PCBM/CH3NH3PbI3-XClX layer after treatment with methanol, d) after deposition of PEIE on the
a
d
c b
RMS = 16.7 nm RMS = 6.1 nm
RMS = 8.4 nm RMS = 10.3 nm
RMS = 16.4 nm
e
133
PCBM/CH3NH3PbI3-XClX layer and e) after deposition of P3TMAHT on the PCBM/CH3NH3PbI3-XClX
layer. Reproduced with permission from ref.224, Copyright © 2014, American Chemical Society.
7.3 Perovskite Solar Cell with a Water-Free PEDOT
From an industrial point of view, the current perovskite solar cell technology is still limited
in multiple aspects. The expensive hole transport materials (HTM), the commonly used doped
spiro-MeOTAD, which greatly increase the materials and manufacturing cost of the device, is
one of the limitations. In addition, spiro-MeOTAD requires a relatively complex doping
strategy by adding polar components into the solution for the benefit on an enhanced hole
transport ability.228
This not only increases the complexity of the fabrication steps, but also
questions the reproducibility and standing time of this ink at industrial scales. The
development of a low-cost HTM, which is stable in its solution form, is therefore among the
most important issues to evolve this solar cell technology from lab to industry.
In this section, a novel water-free PEDOT in toluene is introduced into perovskite solar cells
as a low-cost hole transporting material processed on top of perovskite thin films. Using a
fairly simple device architecture of ITO/LT-TiO2/Perovskite/Water-free PEDOT/Au, a power
conversion efficiency of up to 11.75% was demonstrated, along with 800 hours
environmentally stability in air without any encapsulation. Additionally, all the active layers
within this device are solution-processed at temperatures below 140 °C, which makes this
architecture compatible to roll-to-roll processing on plastic substrates.
7.3.1 Optical properties of perovskite films with water-free PEDOT
Figure 7-5a shows a device structure consisting of a low-temperature processed TiO2 (LT-
TiO2) as electron transport layer, a water-free PEDOT as a hole transport layer and a
sandwiched CH3NH3PbI3 layer. The work function of water-free PEDOT (Clevios HTL Solar
3) determined by Kelvin probe was found to be -5.0 eV, which is higher than that of
CH3NH3PbI3, allowing for efficient hole transfer from CH3NH3PbI3 to water-free PEDOT.
As seen from the energy diagram in Figure 7-5b, the energy levels of all layers are well
aligned which could allow efficient charge extraction.
134
Figure 7-5. a) Schematics cross-sectional view of the perovskite solar cell architecture: ITO/LT-
TiO2/perovskite/water-free PEDOT/Au and the corresponding process temperature and time of three active
layers; b) energy levels for LT-TiO2, perovskite, and the water-free PEDOT; c) UV–vis absorption spectra
of water-free PEDOT, perovskite, and perovskite/water-free PEDOT on quartz glass. Inset shows the UV–
vis absorption spectra of water-free PEDOT with a different scale. All the films used for the UV
measurement are made under exactly the same conditions as the devices; d) photoluminescence (PL)
spectra of perovskite films on top of LT-TiO2 and the one with incorporation of water-free PEDOT on top
of perovskite films. Inset shows the picture of 200 mL water-free PEDOT solution with ready availability.
Reproduced with permission from ref.229, Copyright © 2015 Wiley-VCH Verlag GmbH & Co. KGaA.
Figure 7-5c shows the UV–vis absorption of water-free PEDOT and perovskite films on
quartz substrates. We can see that there is negligible absorption of the water-free PEDOT in
the visible light region (inset of Figure 7-5c). The slight increase in the near-IR (NIR) regime
is attributed to the polaron absorption. In contrast, the perovskite films exhibit a strong
absorption in the visible light region. To further investigate the contribution of water-free
PEDOT to the overall absorption, the UV–vis spectra of CH3NH3PbI3/water-free PEDOT
stacked films were recorded for comparison. As it is shown in Figure 7-5c, the addition of
135
water-free PEDOT causes a slight enhancement around 700–850 nm, which can be clearly
attributed to the absorption by the PEDOT layer. The weak NIR absorption opens the
possibility of photons harvesting via reflection from the top Au electrode. We envisage that
PEDOT allows reflecting more unabsorbed photons from the Au electrode and pass through
the CH3NH3PbI3 layer again, which is in contrast to case of the HTM with strong absorption
in visible light region. In order to investigate whether water-free PEDOT layer could
efficiently extract photogenerated carriers from the perovskite absorber, we performed
photoluminescence (PL) measurements. From Figure 7-5d, we can find that the
CH3NH3PbI3 film shows a significant degree of PL quenching when a water-free PEDOT
layer is coated on top. This clearly highlights that water-free PEDOT is efficiently collecting
holes and/or electrons.
Figure 7-6. a) AFM topography of the surface of perovskite film and b) water-free PEDOT layer on top of
perovskite. The size of the AFM images is 5 × 5 μm2. The measured root mean square roughness is a) 14.2
nm and b) 3.1 nm; c) SEM image of the surface of perovskite layer; d) cross-sectional SEM image of a
complete perovskite device. Reproduced with permission from ref.229, Copyright © 2015 Wiley-VCH
Verlag GmbH & Co. KGaA.
7.3.2 Morphology of perovskite films with water-free PEDOT
136
It is well known that the film formation and surface morphology of the perovskite absorber
layer do play a crucial role in determining the ultimate solar cell performance. Perovskite
films showed the formation of CH3NH3PbI3 nanocrystals on the surface of the titanium
dioxide anode under high-magnification SEM images. From Figure 7-6a and b, we can find
that the surface roughness RMS of perovskite films is around 14.2 nm. The top PEDOT layer
further smoothens the surface down to a RMS value of 3.1 nm, which is perfectly sufficient
to provide a suitable interface for the top electrode.
7.3.3 Photovoltaic properties of perovskite solar cells with water-free
PEDOT
To probe the effect of layer thickness of water-free PEDOT (Clevios HTL Solar 3) on the
device performance, devices of various water-free PEDOT layer thickness were prepared by
spin-coating at different spin speeds (Table 7-2). As can be seen from Figure 7-7a, the
appropriate thickness of the hole transport layer is crucial to obtain high-performance devices.
With the increase of the water-free PEDOT thickness from 190 to 245 nm, the PCE and
the JSC enhance simultaneously. On the other hand, there is a slight loss in VOC;
however, VOC is still kept at above 1V.
Table 7-2. Key values of the j-V characteristics of the investigated perovskite solar cells with different
water-free PEDOT thicknesses under illumination of an AM 1.5G solar simulator (100 mW/cm2).
The thickness of
waterfree-PEDOT
VOC
(mV)
JSC
(mA cm-2
)
FF
(%)
PCE
(%)
RS
(Ω cm2)
RSh
(kΩ cm2)
190 nm 1032 -12.33 67.1 8.54 1.56 21.07 220 nm 1022 -14.40 61.5 9.05 3.33 17.42 245 nm 1001 -15.91 64.0 10.02 3.39 14.35
To investigate the reproducibility of the results, over 50 separate devices were fabricated and
tested using the optimized water-free PEDOT thickness. Figure 7-7b exhibits the devices
better reproducibility and performance. The j–V characteristics and EQE spectrum of the
best-performing device is shown in Figure 7-7c and d. From the j–V curve measured under
137
standard AM1.5G illumination with a scan direction from forward bias to short circuit,
the JSC, VOC, FF, and PCE were determined to be 15.93 mA cm−2, 1.02 V, 72.3%, and
11.75%. Additionally, an anomalous hysteresis in the j–V curves of planar perovskite solar
cells has been observed which agrees with other reported literatures.230, 231 The PCE of 9.7%
was recorded when using a scan direction from forward bias to short circuit with a scan rate
of 0.1 V s−1 while the opposite scan direction yielded an efficiency of 7.5%. The main
difference comes from FF, and is attributed mainly to the slow electron transfer from the
perovskite absorber to the n-type charge collection layer. 232
Figure 7-7. a) Efficiency variation of perovskite solar cells prepared by DMF:DMSO precursor with
different water-free PEDOT (Clevios HTL Solar 3) thicknesses; b) histogram of the solar cell efficiencies
obtained from 50 samples processed by mixed solvent GBL:DMSO and DMF:DMSO, respectively. c) j–
V characteristics measured under 100 mW cm−2 AM1.5G illumination (red line) and in the dark (black line)
for the highest-performing ITO/LT-TiO2/perovskite/water-free PEDOT (Clevios HTL Solar 3)/Au device
prepared by DMF:DMSO precursor. d) EQE spectrum of the highest-performing ITO/LT-
TiO2/perovskite/water-free PEDOT (Clevios HTL Solar 3)/Au device prepared by DMF:DMSO precursor.
Reproduced with permission from ref.229, Copyright © 2015 Wiley-VCH Verlag GmbH & Co. KGaA.
138
Considering the scale up of efficient perovskite solar cells, the stability is as important as
efficiency and costs. In fact, there are many factors that influence the stability of the
perovskite solar cells, such as instability of perovskite layer, reaction between perovskite and
electrodes, and so on. Here, we also confirmed that using water-free PEDOT and gold
electrodes are very effective means for enhancing the cell durability, which is maintained at
an almost constant level of ≈9% after 800 h. As shown in Figure 7-8b, a device based on
water-free PEDOT and Ag electrodes without encapsulation undergoes fairly fast degradation
under ambient conditions at room temperature of a humidity of about ≈25%. This is much
more unstable than cells based on water-free PEDOT and Au electrode under the same
conditions. The dramatically different lifetimes indicate that Ag is reacting much faster with
the perovskite layer than Au.
7.3.3 Defect states at the interface to the perovskite
In order to further examine whether water-free PEDOT has any additional impact on the
formation of defect states at the interface to the perovskite, we investigated the sub-bandgap
photocurrent generation for different hole extraction layers. In addition to spiro-MeOTAD,
we tested two water-free PEDOTs: one is the already described Clevios HTL Solar 3, while
the other one is a pH neutralized version of the water-free, toluene based PEDOT (Clevios
SEJ 272). The sub-bandgap photocurrent was investigated with the fairly sensitive Fourier-
transform photocurrent spectroscopy (FTPS) (Figure 7-8a). Consistent with previous
reports,233 for high qualitative perovskites we find a sharp optical edges at ≈1.6 eV and an
exponential, very steep decay for over four orders in magnitude. In good agreement with
previous studies for perovskite/sipro-MeOTAD layers, we find that the Urbach tail-like
absorption features are distributed within few 10 of meV233 proving the absence of optically
active defect states in perovskite absorber layer. However, we also observed a very weak but
distinct contribution at energies below 1.45 eV. For spiro-MeOTAD, these shallow defect
states are in the order of 10−5 or less. Ballif and co-workers had not reported such weak
absorption features, however, they measured FTPS in a planar geometry with no HTL
139
deposited on top of the perovskite.233 In order to understand the origin of the sub-bandgap
features, we further studied the FTPS spectra for the two water-free PEDOT HTLs. After
coating PEDOT (Clevios HTL Solar 3) on top of the perovskite, we found an apparent more
intense subgap signal in the regime between 1.0 and 1.5 eV. The intensity of the signal is
more than an order more intense than observed for the spiro-MeOTAD. Next, by replacing
the water-free acid PEDOT (Clevios HTL Solar 3) with a water-free neutral PEDOT (Clevios
SEJ 272), the Urbach-like subgap absorptions are significantly reduced to almost the level
found for spiro-MeOTAD. We note two observations: first, the deposition of an HTL on top
of the perovskite may contribute to the formation of subgap states at the interface of the
perovskite. Second, the density of subgap states seems to depend on the nature of the HTL,
with acid PEDOT showing the largest density of subgap states and spiro-MeOTAD showing
the lowest one. The single cell hero performance for the acid PEDOT is 11.7%, while a
maximum PCE of 14.2% is observed for the neutral form of the water-free PEDOT (Clevios
SEJ 272).
Figure 7-8. a) Comparison of the normalized FTPS spectra of solar cells applied with the water-free
PEDOT (Clevios HTL Solar 3), the water-free PEDOT (Clevios SEJ 272), and the spiro-MeOTAD. b)
Stability of water-free acid PEDOT devices with different electrode contacts was tested under ~ 25%
humidity atmosphere without encapsulation. Reproduced with permission from ref.229, Copyright ©
2015 Wiley-VCH Verlag GmbH & Co. KGaA.
b
140
7.4 Conclusions
In this chapter, we fabricated solution processed hybrid organic/inorganic perovskite PHJ
solar cells via incorporation of a polyelectrolyte interlayer by using low-temperature
processing. The insertion of a polyelectrolyte interlayer improved the PCE to 12.01% for the
PEIE device and 11.28% for the P3TMAHT device. We attribute the improvement in PCE
for devices as compared to devices without PEIE or P3TMAHT, to the formation of surface
dipoles. Both PEIE and P3TMAHT, effectively reduce the work function of the subsequently
deposited metal, thereby lowering the electron-injection barrier to PCBM. Thus, this study
provides a practical route for fabricating high efficiency perovskite thin film solar cells using
low temperature solution processing.
As a further step towards requirement of standing ink at industrial scales in future perovskite
photovoltaic industry, we introduced a water-free dispersion of PEDOT to replace spiro-
MeOTAD in perovskite solar cells. After the optimization of the water-free PEDOT (Clevios
HTL Solar 3) layer thickness, the device showed a maximum PCE of 11.75% and more than
800 h stability under ambient environmental conditions without packaging. As the impact of
the acid nature on the perovskite surface was a potential concern, we investigated a pH
neutralized PEDOT, formulated in Anisole (Clevios SEJ 272). Moreover, we found a neutral
water-free PEDOT (Clevios SEJ 272) could effectively reduce the density of the sub bandgap
defects. These results suggest that water-free, nonacidic PEDOT, a widely used scalable
conducting polymer dispersed in organic solvent, can become a very promising candidate for
industrial scale processing of perovskite solar cells.
141
Chapter 8
Summary and Outlook
This chapter summarizes the main achievements as presented in this thesis. Limitations and
future challenges for organic solar cells are supplied as an outlook.
142
8.1 Summary
OSCs technology has undergone a rapid development in the past two decades, and significant
progress is required to develop the OSCs technology towards industrial standards. From an
industrial point of view, the current OSCs technology is still limited in multiple aspects,
which are three important factors of process, stability and power conversion efficiency.
Introduction of innovative interface engineering technologies via new materials and processes
make OSCs more efficient, reliable, cost effective and environmentally benign.
In the first part of this thesis, we focused on the development and understanding of interface
contact between the organic semiconductors (PCBM) and electrodes (MOX) for OSCs.
Chapter 4.1 studied the effects of an alkaline earth hydroxide Ba(OH)2 layer on the
improvement of device performance. We employed a Ba(OH)2 layer to tune the barrier
between the conduction band of AZO and the LUMO of PCBM in the active layer. We found
that the improvement in device performance originated from the potential surface doping by
diffusion of Ba2+ ions through the surface of active layer, the reduced charge recombination,
decreased energy barrier for electron extraction and transport via inserting an interfacial
dipole layer and the increased built-in potential from a correspondingly increase of surface
charge density. Chapter 4.2 further investigated the interfacial energetics of the alkali
hydroxide KOH functionalized AZO contacts with a prototypical electron acceptor (C60)
using UPS measurements, which reveal the presence of a large interface dipole and a new
interface state between the Fermi energy and the C60 HOMO for alkali hydroxide-modified
AZO contacts. These novel interfacial gap states are a result of ground state electron transfer
from the metal hydroxide-functionalized AZO contact to the adsorbed molecules, which are
hypothesized to be electronically hybridized with the contact. These interface states tail all
the way to the Fermi energy, providing for a highly n-doped (metal-like) interfacial molecular
layer. In order to understand a deep recombination mechanistic of the active layer/contact
with interface modification, Chapter 4.3 studied the recombination kinetics by measuring the
j-V characteristics of devices at various illumination intensities with various interface
treatment. Combined with a numerical device model based on a 1D drift–diffusion approach,
143
the majority carrier density at the respective interface was reduced by inserting an interface
modification layer. We considered the effect of interface modification is interface-induced
suppression of the bulk recombination in organic bulk heterojunction photovoltaics. The
research works summarized in this thesis should contribute in this progression. In i), alkali
hydroxides, which are cost-effective and can simply be fabricated in ambient atmosphere, can
tune the interface barrier to improve the efficiency and lifetime of OSCs. In ii), we reveal a
new electronically hybridized interface state to enable a deep understanding of the interfacial
energetics at modified cathode contacts. In iii), we proposed a new recombination
mechanistic “interface-induced suppression of the bimolecular recombination in organic bulk
heterojunction solar cells” to further understand the recombination related losses at modified
cathode contacts.
In the second part of this thesis, we introduced two elegant approaches to printing efficient
devices compatible to roll to roll processing under ambient conditions. In Chapter 5.1, a
simple method was introduced to print efficient, inverted OSCs using a self-organized, small-
molecule cathode interfacial material Phen-NaDPO. This simple one-step solution processing
led to a de facto bilayer with a bottom cathode interfacial layer and a top BHJ photoactive
layer is advantageous for high-volume roll-to-roll printing. In Chapter 5.2, we successfully
developed an s-MoOX/PEG ink as HTL to fabricate all solution processed inverted OSCs via
doctor-blading in air, which showed a performance comparable to those with evaporated
MoOX. This solution processed s-MoOX/PEG hybrid does not require any annealing process
and therefore can be easily transferred to industrial production processes. The research works
in this thesis should contribute in this progression. In i), the present results highlight that
properly designed self-organized cathode interfacial material processed from a “ternary”
active layer is fully compatible with the requirements for roll-to-roll fabrication of inverted
organic solar cells. In ii), as an alternative to PEDOT:PSS and evaporated MoOX, s-
MoOX:PEG is a promising hole transport material for highly efficient and stable inverted
organic solar cells that can be used on flexible substrates via roll-to-roll processing.
In the third part of this thesis, we focused on the development of organic–metal interfaces
based on perovskite solar cells. We demonstrate that fairly efficient perovskite solar cell can
144
be fabricated by inserting an ultrathin polyelectrolyte layer based on the interface layers
solution-processed at low temperature, either PEIE or P3TMAHT, in Chapter 6.1. The PCE
increases from 8.53% to 12.01% (PEIE) and 11.28% (P3TMAHT) for the solution-processed
polyelectrolyte-modified interfaces. In Chapter 6.2, we introduced a water-free dispersion of
PEDOT to replace spiro-MeOTAD in perovskite solar cells, and the device showed a
maximum PCE of 11.75% and more than 800 h stability under ambient environmental
conditions without packaging. We also found that the density of subgap states at the interface
of the perovskite/HTL seems to depend on the nature of PEDOT. This water-free PEDOT can
become a very promising candidate for industrial scale processing of perovskite solar cells.
The research works in this thesis should contribute in this progression. In i), the suitability of
the interface opens the opportunity to apply the interface design rules of organic solar cells to
the perovskite photovoltaic technology. In ii), the development of large-scale organic
materials, which are cost-effective and can be fabricated at low temperature in ambient
atmosphere, as interface layers or modification layer is in favour of the concept of large scale
fabrication and cost-effective solution processing to evolve this solar cell technology from lab
to industry.
8.2 Outlook
The results and findings demonstrated in this thesis are a broad investigation of the electronic
and chemical structure of organic semiconductor interfaces, and polymer interfaces in
particular. The goal of this work is to achieve control of these interfaces in order to optimize
injection, extraction and transport of charge carriers into, from and across, respectively and to
develop more powerful interfacial materials to pave the way for commercialization of OSCs
via roll-to-roll processing.
Although the development of interface layer is in an advanced stage, there is not a reliable
interface engineering fundamental understanding of interface science for device manufacturer.
The physicochemical mechanisms that govern the energy level alignment at organic/electrode
145
and organic/organic interfaces are equally important. In Chapter 4, we further understand the
effect of alkali metal salt on interfacial energetics at organic/cathode interface, which reveal a
highly n-doped (metal-like) interfacial molecular layer. In order to develop and refine the
fundamental interfacial modification mechanism, an assumed research on n-type and p-typre
doped interfacial molecular layers would offer a clear answer to the important question on the
essential difference between the contact of cathode and anode. More work is needed to
answer a lot of questions for interface processes, including the dynamics of charge transport
and separation at such interfaces and the interfacial recombination process of charge carriers.
As discussed in Chapter 5, the proper materials choice and carefully engineering for interface
layer can solve many problems existing in commercialization of OSCs. Important points that
should be addressed in the future are the better contacting schemes, the reduced number of
layers, the simplified technological process, the reduced production costs and higher
efficiency and lifetime of the printed solar cells. For example, incorporating the tandem
concept into printing techniques requires developing a new tandem structure that stacks with
a minimum number of layers, which means connecting two sub single-junction cells with a
recombination layer that combines p-type and n-type interfacial materials. Nevertheless, the
great progress in OSCs gives the opportunity and confidence to develop more powerful and
proper interfacial materials and roll-to-roll compatible printing and coating methods, thus to
improve the industrialization of OSCs for sustainable future developments.
146
147
Appendix A
Bibliography
1. European Photovoltaic Industry Association (2013). "Global Market Outlook for
Photovoltaics 2013-2017".
2. https://en.wikipedia.org/wiki/Photovoltaics#cite_note-epia-2013-market-report-41.
3. C. J. Brabec, N. S. Sariciftci and J. C. Hummelen, Adv Funct Mater, 2001, 11, 15-26.
4. C. J. Brabec, S. Gowrisanker, J. J. M. Halls, D. Laird, S. J. Jia and S. P. Williams, Adv
Mater, 2010, 22, 3839-3856.
5. C. J. Brabec, Sol Energ Mat Sol C, 2004, 83, 273-292.
6. G. Dennler, M. C. Scharber and C. J. Brabec, Adv Mater, 2009, 21, 1323-1338.
7. R. Sondergaard, M. Hosel, D. Angmo, T. T. Larsen-Olsen and F. C. Krebs, Mater
Today, 2012, 15, 36-49.
8. F. C. Krebs, Sol Energ Mat Sol C, 2009, 93, 1636-1641.
9. W. Brütting, in Physics of Organic Semiconductors, Wiley-VCH Verlag GmbH & Co. KGaA, 2006, DOI: 10.1002/3527606637.ch, pp. 1-14.
10. S. Günes, H. Neugebauer and N. S. Sariciftci, Chem Rev, 2007, 107, 1324-1338.
11. A. Pivrikas, G. Juška, A. J. Mozer, M. Scharber, K. Arlauskas, N. Sariciftci, H. Stubb and R. Österbacka, Phys Rev Lett, 2005, 94, 176806.
12. S. M. Sze and K. K. Ng, Physics of semiconductor devices, John wiley & sons, 2006.
13. A. Köhler and H. Bässler, in Electronic Processes in Organic Semiconductors, Wiley-VCH Verlag GmbH & Co. KGaA, 2015, DOI: 10.1002/9783527685172.ch1, pp. 1-86.
14. A. K. K. Kyaw, X. W. Sun, C. Y. Jiang, G. Q. Lo, D. W. Zhao and D. L. Kwong, Appl
Phys Lett, 2008, 93.
15. http://oelectronics.weebly.com/organic-semiconductors.html.
148
16. V. Coropceanu, J. Cornil, D. A. da Silva, Y. Olivier, R. Silbey and J. L. Bredas, Chem
Rev, 2007, 107, 926-952.
17. L.-L. Chua, J. Zaumseil, J.-F. Chang, E. C.-W. Ou, P. K.-H. Ho, H. Sirringhaus and R. H. Friend, Nature, 2005, 434, 194-199.
18. H. Spanggaard and F. C. Krebs, Sol Energ Mat Sol C, 2004, 83, 125-146.
19. M. Volmer, Ann Phys-berlin, 1913, 345, 775-796.
20. H. Shirakawa, E. J. Louis, A. G. Macdiarmid, C. K. Chiang and A. J. Heeger, J Chem
Soc Chem Comm, 1977, DOI: Doi 10.1039/C39770000578, 578-580.
21. C. K. Chiang, C. R. Fincher, Y. W. Park, A. J. Heeger, H. Shirakawa, E. J. Louis, S. C. Gau and A. G. Macdiarmid, Phys Rev Lett, 1977, 39, 1098-1101.
22. P. Peumans, A. Yakimov and S. R. Forrest, J Appl Phys, 2003, 93, 3693-3723.
23. C. W. Tang, Appl Phys Lett, 1986, 48, 183-185.
24. N. S. Sariciftci, L. Smilowitz, A. J. Heeger and F. Wudl, Science, 1992, 258, 1474-1476.
25. G. Yu, J. Gao, J. C. Hummelen, F. Wudl and A. J. Heeger, Science, 1995, 270, 1789-1791.
26. S. P. Economopoulos, G. Itskos, P. A. Koutentis and S. A. Choulis, in Organic
Photovoltaics, Wiley-VCH Verlag GmbH & Co. KGaA, 2014, DOI: 10.1002/9783527656912.ch01, pp. 1-26.
27. C. H. Lee, G. Yu, D. Moses and A. J. Heeger, Phys Rev B, 1994, 49, 2396-2407.
28. M. Theander, A. Yartsev, D. Zigmantas, V. Sundstrom, W. Mammo, M. R. Andersson and O. Inganas, Phys Rev B, 2000, 61, 12957-12963.
29. N. C. Miller, E. T. Hoke and M. D. McGehee, in Organic Photovoltaics, Wiley-VCH Verlag GmbH & Co. KGaA, 2014, DOI: 10.1002/9783527656912.ch13, pp. 421-444.
30. G.-J. A. H. Wetzelaer, L. J. A. Koster and P. W. M. Blom, in Organic Photovoltaics, Wiley-VCH Verlag GmbH & Co. KGaA, 2014, DOI: 10.1002/9783527656912.ch11, pp. 343-376.
31. M. A. Green, K. Emery, Y. Hishikawa, W. Warta and E. D. Dunlop, Prog
Photovoltaics, 2016, 24, 3-11.
32. L. Y. Lu and L. P. Yu, Adv Mater, 2014, 26, 4413-4430.
33. Y. H. Liu, J. B. Zhao, Z. K. Li, C. Mu, W. Ma, H. W. Hu, K. Jiang, H. R. Lin, H. Ade
149
and H. Yan, Nat Commun, 2014, 5, 5293.
34. J. Zhao, Y. Li, A. Hunt, J. Zhang, H. Yao, Z. Li, J. Zhang, F. Huang, H. Ade and H. Yan, Adv Mater, 2016, 28, 1868-1873.
35. W. Ma, C. Yang, X. Gong, K. Lee and A. J. Heeger, Adv Funct Mater, 2005, 15, 1617-1622.
36. H.-C. Liao, C.-C. Ho, C.-Y. Chang, M.-H. Jao, S. B. Darling and W.-F. Su, Mater
Today, 2013, 16, 326-336.
37. H. Zhang, T. Stubhan, N. Li, M. Turbiez, G. J. Matt, T. Ameri and C. J. Brabec, J
Mater Chem A, 2014, 2, 18917-18923.
38. R. Steim, F. R. Kogler and C. J. Brabec, J Mater Chem, 2010, 20, 2499-2512.
39. T. Ameri, G. Dennler, C. Lungenschmied and C. J. Brabec, Energ Environ Sci, 2009, 2, 347-363.
40. J. Y. Kim, K. Lee, N. E. Coates, D. Moses, T.-Q. Nguyen, M. Dante and A. J. Heeger, Science, 2007, 317, 222-225.
41. G. Zhao, Y. He and Y. Li, Adv Mater, 2010, 22, 4355-4358.
42. H. Hoppe and N. S. Sariciftci, J Mater Res, 2004, 19, 1924-1945.
43. H. Ma, H. L. Yip, F. Huang and A. K. Y. Jen, Adv Funct Mater, 2010, 20, 1371-1388.
44. http://plasticphotovoltaics.org/lc/lc-polymersolarcells/lc-how.html.
45. C. Deibel and V. Dyakonov, Rep Prog Phys, 2010, 73, 096401.
46. C. Deibel, D. Mack, J. Gorenflot, A. Scholl, S. Krause, F. Reinert, D. Rauh and V. Dyakonov, Phys Rev B, 2010, 81, 085202.
47. P. E. Shaw, A. Ruseckas and I. D. W. Samuel, Adv Mater, 2008, 20, 3516-3520.
48. A. Haugeneder, M. Neges, C. Kallinger, W. Spirkl, U. Lemmer, J. Feldmann, U. Scherf, E. Harth, A. Gügel and K. Müllen, Phys Rev B, 1999, 59, 15346.
49. T. Stubinger and W. Brutting, J Appl Phys, 2001, 90, 3632-3641.
50. J.-M. Nunzi, Comptes Rendus Physique, 2002, 3, 523-542.
51. J. E. Kroeze, T. J. Savenije, M. J. Vermeulen and J. M. Warman, The Journal of
Physical Chemistry B, 2003, 107, 7696-7705.
52. L. Lüer, H.-J. Egelhaaf, D. Oelkrug, G. Cerullo, G. Lanzani, B.-H. Huisman and D. De Leeuw, Org Electron, 2004, 5, 83-89.
150
53. L. Onsager, Physical Review, 1938, 54, 554.
54. C. L. Braun, The Journal of chemical physics, 1984, 80, 4157-4161.
55. C. Deibel, A. Wagenpfahl and V. Dyakonov, physica status solidi (RRL)-Rapid
Research Letters, 2008, 2, 175-177.
56. A. Pivrikas, N. S. Sariciftci, G. Juška and R. Österbacka, Progress in Photovoltaics:
Research and Applications, 2007, 15, 677-696.
57. A. Wagenpfahl, D. Rauh, M. Binder, C. Deibel and V. Dyakonov, Phys Rev B, 2010, 82, 115306.
58. B. Y. Qi and J. Z. Wang, Phys Chem Chem Phys, 2013, 15, 8972-8982.
59. J. H. Lee, S. Cho, A. Roy, H. T. Jung and A. J. Heeger, Appl Phys Lett, 2010, 96.
60. E. L. Ratcliff, B. Zacher and N. R. Armstrong, J Phys Chem Lett, 2011, 2, 1337-1350.
61. M. M. Mandoc, W. Veurman, L. J. A. Koster, B. de Boer and P. W. M. Blom, Adv
Funct Mater, 2007, 17, 2167-2173.
62. G. Lakhwani, A. Rao and R. H. Friend, Annual Review of Physical Chemistry, Vol 65, 2014, 65, 557-581.
63. C. M. Proctor, M. Kuik and T. Q. Nguyen, Prog Polym Sci, 2013, 38, 1941-1960.
64. L. J. A. Koster, V. D. Mihailetchi and P. W. M. Blom, Appl Phys Lett, 2006, 88, 052104.
65. M. D. McGehee, Nat Photonics, 2009, 3, 250-252.
66. C. Deibel, T. Strobel and V. Dyakonov, Adv Mater, 2010, 22, 4097-4111.
67. J. M. Hodgkiss, S. Albert-Seifried, A. Rao, A. J. Barker, A. R. Campbell, R. A. Marsh and R. H. Friend, Adv Funct Mater, 2012, 22, 1567-1577.
68. D. S. Lemons and A. Gythiel, Am J Phys, 1997, 65, 1079-1081.
69. C. G. Shuttle, R. Hamilton, B. C. O'Regan, J. Nelson and J. R. Durrant, Proceedings
of the National Academy of Sciences of the United States of America, 2010, 107, 16448-16452.
70. D. Credgington and J. R. Durrant, J Phys Chem Lett, 2012, 3, 1465-1478.
71. L. J. A. Koster, M. Kemerink, M. M. Wienk, K. Maturova and R. A. J. Janssen, Adv
Mater, 2011, 23, 1670-1674.
72. V. Gupta, A. K. K. Kyaw, D. H. Wang, S. Chand, G. C. Bazan and A. J. Heeger, Sci
151
Rep-Uk, 2013, 3, 1965
73. S. R. Cowan, A. Roy and A. J. Heeger, Phys Rev B, 2010, 82, 245207.
74. M. Kuik, L. J. A. Koster, G. A. H. Wetzelaer and P. W. M. Blom, Phys Rev Lett, 2011, 107, 256805.
75. M. M. Mandoc, F. B. Kooistra, J. C. Hummelen, B. de Boer and P. W. M. Blom, Appl
Phys Lett, 2007, 91, 263505.
76. V. I. Klimov, A. A. Mikhailovsky, S. Xu, A. Malko, J. A. Hollingsworth, C. A. Leatherdale, H. J. Eisler and M. G. Bawendi, Science, 2000, 290, 314-317.
77. D. Bartesaghi, M. Turbiez and L. J. A. Koster, Org Electron, 2014, 15, 3191-3202.
78. J. Peet, J. Y. Kim, N. E. Coates, W. L. Ma, D. Moses, A. J. Heeger and G. C. Bazan, Nat Mater, 2007, 6, 497-500.
79. C. J. Brabec, M. Heeney, I. McCulloch and J. Nelson, Chem Soc Rev, 2011, 40, 1185-1199.
80. A. Kumar, G. Lakhwani, E. Elmalem, W. T. S. Huck, A. Rao, N. C. Greenham and R. H. Friend, Energ Environ Sci, 2014, 7, 2227-2231.
81. T. Kirchartz, B. E. Pieters, K. Taretto and U. Rau, Phys Rev B, 2009, 80, 035334.
82. S. Wheeler, F. Deledalle, N. Tokmoldin, T. Kirchartz, J. Nelson and J. R. Durrant, Phys Rev Appl, 2015, 4, 024020.
83. L. J. A. Koster, E. C. P. Smits, V. D. Mihailetchi and P. W. M. Blom, Phys Rev B, 2005, 72, 085205.
84. T. Kirchartz, F. Deledalle, P. S. Tuladhar, J. R. Durrant and J. Nelson, J Phys Chem
Lett, 2013, 4, 2371-2376.
85. A. Petersen, T. Kirchartz and T. A. Wagner, Phys Rev B, 2012, 85, 045208.
86. E. H. Rhoderick and R. H. Williams, Metal-semiconductor contacts, Clarendon Press Oxford, 1988.
87. W. Mönch, Semiconductor surfaces and interfaces, Springer Science & Business Media, 2013.
88. A. Kahn and N. Koch, in The Molecule–Metal Interface, Wiley-VCH Verlag GmbH & Co. KGaA, 2013, DOI: 10.1002/9783527653171.ch8, pp. 219-241.
89. H. Ishii, K. Sugiyama, E. Ito and K. Seki, Adv Mater, 1999, 11, 972-972.
152
90. X. Crispin, V. Geskin, A. Crispin, J. Cornil, R. Lazzaroni, W. R. Salaneck and J. L. Bredas, J Am Chem Soc, 2002, 124, 8131-8141.
91. J. Hwang, A. Wan and A. Kahn, Mat Sci Eng R, 2009, 64, 1-31.
92. S. Braun, W. Osikowicz, Y. Wang and W. R. Salaneck, Org Electron, 2007, 8, 14-20.
93. A. Crispin, X. Crispin, M. Fahlman, M. Berggren and W. R. Salaneck, Appl Phys Lett, 2006, 89, 213503.
94. S. Braun, W. R. Salaneck and M. Fahlman, Adv Mater, 2009, 21, 1450-1472.
95. Z. He, C. Zhong, S. Su, M. Xu, H. Wu and Y. Cao, Nat Photonics, 2012, 6, 591-595.
96. X. Crispin, Sol Energ Mat Sol C, 2004, 83, 147-168.
97. A. Wagenpfahl, C. Deibel and V. Dyakonov, Ieee J Sel Top Quant, 2010, 16, 1759-1763.
98. M. Jorgensen, K. Norrman, S. A. Gevorgyan, T. Tromholt, B. Andreasen and F. C. Krebs, Adv Mater, 2012, 24, 580-612.
99. J. U. Lee, J. W. Jung, J. W. Jo and W. H. Jo, J Mater Chem, 2012, 22, 24265-24283.
100. W. R. Mateker, J. D. Douglas, C. Cabanetos, I. T. Sachs-Quintana, J. A. Bartelt, E. T. Hoke, A. El Labban, P. M. Beaujuge, J. M. J. Frechet and M. D. McGehee, Energ Environ
Sci, 2013, 6, 2529-2537.
101. B. T. de Villers, C. J. Tassone, S. H. Tolbert and B. J. Schwartz, J Phys Chem C, 2009, 113, 18978-18982.
102. W. Tress, K. Leo and M. Riede, Adv Funct Mater, 2011, 21, 2140-2149.
103. W. C. Yang, Y. Yao and C. Q. Wu, Org Electron, 2013, 14, 1992-2000.
104. M. Glatthaar, M. Riede, N. Keegan, K. Sylvester-Hvid, B. Zimmermann, M. Niggemann, A. Hinsch and A. Gombert, Sol Energ Mat Sol C, 2007, 91, 390-393.
105. R. D. Xia, D. S. Leem, T. Kirchartz, S. Spencer, C. Murphy, Z. C. He, H. B. Wu, S. J. Su, Y. Cao, J. S. Kim, J. C. deMello, D. D. C. Bradley and J. Nelson, Adv Energy Mater, 2013, 3, 718-723.
106. H. Wirth, 2015, vol. Fraunhofer ISE.
107. J. Reinhardt, M. Grein, C. Buhler, M. Schubert and U. Wurfel, Adv Energy Mater, 2014, 4, 1400081.
108. H. Ehrenreich and F. Spaepen, Solid state physics, Academic Press, 2001.
153
109. F. Seitz and D. Turnbull, Solid state physics: advances in research and applications, Academic Press, 1961.
110. J. Ristein, Science, 2006, 313, 1057-1058.
111. M. T. Edmonds, M. Wanke, A. Tadich, H. M. Vulling, K. J. Rietwyk, P. L. Sharp, C. B. Stark, Y. Smets, A. Schenk, Q. H. Wu, L. Ley and C. I. Pakes, J Chem Phys, 2012, 136, 124701.
112. K. Rietwyk, M. Wanke, H. Vulling, M. Edmonds, P. Sharp, Y. Smets, Q.-H. Wu, A. Tadich, S. Rubanov and P. Moriarty, Phys Rev B, 2011, 84, 035404.
113. J. Ristein, Applied Physics A, 2006, 82, 377-384.
114. D. C. Gleason-Rohrer, B. S. Brunschwig and N. S. Lewis, J Phys Chem C, 2013, 117, 18031-18042.
115. Y. Q. Zhang, R. Acharya and X. A. Cao, J Appl Phys, 2012, 112, 013103.
116. C. C. Wu, C. I. Wu, J. C. Sturm and A. Kahn, Appl Phys Lett, 1997, 70, 1348-1350.
117. M. G. Mason, L. S. Hung, C. W. Tang, S. T. Lee, K. W. Wong and M. Wang, J Appl
Phys, 1999, 86, 1688-1692.
118. K. Sugiyama, H. Ishii, Y. Ouchi and K. Seki, J Appl Phys, 2000, 87, 295-298.
119. M. R. Lilliedal, A. J. Medford, M. V. Madsen, K. Norrman and F. C. Krebs, Sol Energ
Mat Sol C, 2010, 94, 2018-2031.
120. H. L. Yip, S. K. Hau, N. S. Baek, H. Ma and A. K. Y. Jen, Adv Mater, 2008, 20, 2376-2382.
121. C. H. Hsieh, Y. J. Cheng, P. J. Li, C. H. Chen, M. Dubosc, R. M. Liang and C. S. Hsu, J Am Chem Soc, 2010, 132, 4887-4893.
122. S. K. Hau, H. L. Yip, H. Ma and A. K. Y. Jen, Appl Phys Lett, 2008, 93, 233304.
123. S. K. Hau, Y. J. Cheng, H. L. Yip, Y. Zhang, H. Ma and A. K. Y. Jen, Acs Appl Mater
Inter, 2010, 2, 1892-1902.
124. W. Tress, Organic Solar Cells: Theory, Experiment, and Device Simulation, Springer International Publishing, 2014.
125. B. E. Sernelius, K.-F. Berggren, Z.-C. Jin, I. Hamberg and C. Granqvist, Phys Rev B, 1988, 37, 10244.
126. T. Stubhan, H. Oh, L. Pinna, J. Krantz, I. Litzov and C. J. Brabec, Org Electron, 2011, 12, 1539-1543.
154
127. O. Pachoumi, C. Li, Y. Vaynzof, K. K. Banger and H. Sirringhaus, Adv Energy Mater, 2013, 3, 1428-1436.
128. Z. G. Yin, Q. D. Zheng, S. C. Chen, D. D. Cai and Y. L. Ma, Adv Energy Mater, 2016, 6, 1501493.
129. Z. Y. Hu, J. J. Zhang and Y. J. Zhu, Sol Energ Mat Sol C, 2013, 117, 610-616.
130. A. K. K. Kyaw, X. W. Sun, D. W. Zhao, S. T. Tan, Y. Divayana and H. V. Demir, Ieee
J Sel Top Quant, 2010, 16, 1700-1706.
131. Z. Yin, Q. Zheng, S. C. Chen, D. Cai, L. Zhou and J. Zhang, Adv Energy Mater, 2014, 4, 1301404.
132. M. Ohyama, H. Kozuka and T. Yoko, Thin Solid Films, 1997, 306, 78-85.
133. T. Lenz, M. Richter, G. J. Matt, N. A. Luechinger, S. C. Halim, W. Heiss and C. J. Brabec, J Mater Chem C, 2015, 3, 1468-1472.
134. I. Irfan, S. Graber, F. So and Y. L. Gao, Org Electron, 2012, 13, 2028-2034.
135. M. G. Helander, Z. B. Wang, M. T. Greiner, Z. W. Liu, J. Qiu and Z. H. Lu, Adv
Mater, 2010, 22, 2037-2040.
136. I. Irfan and Y. L. Gao, J Photon Energy, 2012, 2, 021213.
137. Irfan, H. J. Ding, Y. L. Gao, C. Small, D. Y. Kim, J. Subbiah and F. So, Appl Phys
Lett, 2010, 96, 243307.
138. J. Meyer, A. Shu, M. Kroger and A. Kahn, Appl Phys Lett, 2010, 96, 133308.
139. T. Stubhan, M. Salinas, A. Ebel, F. C. Krebs, A. Hirsch, M. Halik and C. J. Brabec, Adv Energy Mater, 2012, 2, 532-535.
140. J. A. Bardecker, H. Ma, T. Kim, F. Huang, M. S. Liu, Y. J. Cheng, G. Ting and A. K. Y. Jen, Adv Funct Mater, 2008, 18, 3964-3971.
141. J. Min, H. Zhang, T. Stubhan, Y. N. Luponosov, M. Kraft, S. A. Ponomarenko, T. Ameri, U. Scherf and C. J. Brabec, J Mater Chem A, 2013, 1, 11306-11311.
142. T. B. Yang, M. Wang, C. H. Duan, X. W. Hu, L. Huang, J. B. Peng, F. Huang and X. Gong, Energ Environ Sci, 2012, 5, 8208-8214.
143. J. H. Seo, A. Gutacker, Y. M. Sun, H. B. Wu, F. Huang, Y. Cao, U. Scherf, A. J. Heeger and G. C. Bazan, J Am Chem Soc, 2011, 133, 8416-8419.
144. A. K. K. Kyaw, D. H. Wang, V. Gupta, J. Zhang, S. Chand, G. C. Bazan and A. J. Heeger, Adv Mater, 2013, 25, 2397-2402.
155
145. Y. Udum, P. Denk, G. Adam, D. H. Apaydin, A. Nevosad, C. Teichert, M. S. White, N. S. Sariciftci and M. C. Scharber, Org Electron, 2014, 15, 997-1001.
146. J. Lee, Y. Park, D. Y. Kim, H. Y. Chu, H. Lee and L. M. Do, Appl Phys Lett, 2003, 82, 173-175.
147. J. S. Huang, Z. Xu and Y. Yang, Adv Funct Mater, 2007, 17, 1966-1973.
148. J. S. Huang, G. Li and Y. Yang, Adv Mater, 2008, 20, 415-419.
149. C. Y. Nam, J Phys Chem C, 2014, 118, 27219-27225.
150. S. Trost, K. Zilberberg, A. Behrendt, A. Polywka, P. Gorrn, P. Reckers, J. Maibach, T. Mayer and T. Riedl, Adv Energy Mater, 2013, 3, 1437-1444.
151. S. Wilken, J. Parisi and H. Borchert, J Phys Chem C, 2014, 118, 19672-19682.
152. J. Kim, G. Kim, Y. Choi, J. Lee, S. H. Park and K. Lee, J Appl Phys, 2012, 111, 114511.
153. C. C. Chueh, C. Z. Li and A. K. Y. Jen, Energ Environ Sci, 2015, 8, 1160-1189.
154. Z. G. Zhang, H. Li, Z. Qi, Z. W. Jin, G. Liu, J. H. Hou, Y. F. Li and J. Z. Wang, Appl
Phys Lett, 2013, 102, 143902.
155. Z. G. Zhang, H. Li, B. Y. Qi, D. Chi, Z. W. Jin, Z. Qi, J. H. Hou, Y. F. Li and J. Z. Wang, J Mater Chem A, 2013, 1, 9624-9629.
156. C. H. Duan, W. Z. Cai, B. B. Y. Hsu, C. M. Zhong, K. Zhang, C. C. Liu, Z. C. Hu, F. Huang, G. C. Bazan, A. J. Heeger and Y. Cao, Energ Environ Sci, 2013, 6, 3022-3034.
157. Y. H. Zhou, C. Fuentes-Hernandez, J. Shim, J. Meyer, A. J. Giordano, H. Li, P. Winget, T. Papadopoulos, H. Cheun, J. Kim, M. Fenoll, A. Dindar, W. Haske, E. Najafabadi, T. M. Khan, H. Sojoudi, S. Barlow, S. Graham, J. L. Bredas, S. R. Marder, A. Kahn and B. Kippelen, Science, 2012, 336, 327-332.
158. S. Woo, W. H. Kim, H. Kim, Y. Yi, H. K. Lyu and Y. Kim, Adv Energy Mater, 2014, 4, 1301692.
159. Z. C. He, B. Xiao, F. Liu, H. B. Wu, Y. L. Yang, S. Xiao, C. Wang, T. P. Russell and Y. Cao, Nat Photonics, 2015, 9, 174-179.
160. H. Choi, J. S. Park, E. Jeong, G. H. Kim, B. R. Lee, S. O. Kim, M. H. Song, H. Y. Woo and J. Y. Kim, Adv Mater, 2011, 23, 2759-2763.
161. D. W. Zhao, P. Liu, X. W. Sun, S. T. Tan, L. Ke and A. K. K. Kyaw, Appl Phys Lett, 2009, 95, 153304.
156
162. C. J. Brabec, S. E. Shaheen, C. Winder, N. S. Sariciftci and P. Denk, Appl Phys Lett, 2002, 80, 1288-1290.
163. F. Chen, Q. Chen, L. Mao, Y. X. Wang, X. Huang, W. Lu, B. Wang and L. W. Chen, Nanotechnology, 2013, 24, 484011.
164. Y. H. Jia, L. Y. Yang, W. J. Qin, S. G. Yin, F. L. Zhang and J. Wei, Renew Energ, 2013, 50, 565-569.
165. T. Xiao, W. P. Cui, M. Cai, W. Leung, J. W. Anderegg, J. Shinar and R. Shinar, Org
Electron, 2013, 14, 267-272.
166. W. R. Lindemann, T. Xiao, W. J. Wang, J. E. Berry, N. A. Anderson, R. S. Houk, R. Shinar, J. Shinar and D. Vaknin, Org Electron, 2013, 14, 3190-3194.
167. G. J. Wang, T. G. Jiu, C. M. Sun, J. Li, P. D. Li, F. S. Lu and J. F. Fang, Acs Appl
Mater Inter, 2014, 6, 833-838.
168. X. D. Li, W. J. Zhang, X. Y. Wang, F. Gao and J. F. Fang, Acs Appl Mater Inter, 2014, 6, 20569-20573.
169. Z. K. Tan, Y. Vaynzof, D. Credgington, C. Li, M. T. L. Casford, A. Sepe, S. Huettner, M. Nikolka, F. Paulus, L. Yang, H. Sirringhaus, N. C. Greenham and R. H. Friend, Adv Funct
Mater, 2014, 24, 3051-3058.
170. P. C. Jukes, S. J. Martin, A. M. Higgins, M. Geoghegan, R. A. L. Jones, S. Langridge, A. Wehrum and S. Kirchmeyer, Adv Mater, 2004, 16, 807-811.
171. A. M. Higgins, S. J. Martin, P. C. Jukes, M. Geoghegan, R. A. L. Jones, S. Langridge, R. Cubitt, S. Kirchmeyer, A. Wehrum and I. Grizzi, J Mater Chem, 2003, 13, 2814-2818.
172. K. Cnops, G. Zango, J. Genoe, P. Heremans, M. V. Martinez-Diaz, T. Torres and D. Cheyns, J Am Chem Soc, 2015, 137, 8991-8997.
173. D. M. Huang, S. A. Mauger, S. Friedrich, S. J. George, D. Dumitriu-LaGrange, S. Yoon and A. J. Moule, Adv Funct Mater, 2011, 21, 1657-1665.
174. Y. Yoon, H. J. Kim, C. H. Cho, S. Kim, H. J. Son, M. J. Ko, H. Kim, D. K. Lee, J. Y. Kim, W. Lee, B. J. Kim and B. Kim, Acs Appl Mater Inter, 2014, 6, 333-339.
175. W. M. Qiu, R. Muller, E. Voroshazi, B. Conings, R. Carleer, H. G. Boyen, M. Turbiez, L. Froyen, P. Heremans and A. Hadipour, Acs Appl Mater Inter, 2015, 7, 3581-3589.
176. X. L. Hou, Q. X. Li, T. Cheng, L. Yu, F. Z. Wang, J. Lin, S. Y. Dai, Y. F. Li and Z. A. Tan, J Mater Chem A, 2015, 3, 18727-18734.
177. H. T. Yang and P. Jiang, Langmuir, 2010, 26, 13173-13182.
157
178. J. You, L. Dou, K. Yoshimura, T. Kato, K. Ohya, T. Moriarty, K. Emery, C.-C. Chen, J. Gao, G. Li and Y. Yang, Nat Commun, 2013, 4, 1446.
179. J. D. Chen, C. H. Cui, Y. Q. Li, L. Zhou, Q. D. Ou, C. Li, Y. F. Li and J. X. Tang, Adv
Mater, 2015, 27, 1035-1041.
180. Z. C. He, C. M. Zhong, X. Huang, W. Y. Wong, H. B. Wu, L. W. Chen, S. J. Su and Y. Cao, Adv Mater, 2011, 23, 4636-4643.
181. L. P. Lu, D. Kabra and R. H. Friend, Adv Funct Mater, 2012, 22, 4165-4171.
182. H. Q. Zhou, Y. Zhang, J. Seifter, S. D. Collins, C. Luo, G. C. Bazan, T. Q. Nguyen and A. J. Heeger, Adv Mater, 2013, 25, 1646-1652.
183. B. R. Lee, E. D. Jung, Y. S. Nam, M. Jung, J. S. Park, S. Lee, H. Choi, S. J. Ko, N. R. Shin, Y. K. Kim, S. O. Kim, J. Y. Kim, H. J. Shin, S. Cho and M. H. Song, Adv Mater, 2014, 26, 494-500.
184. H. Zhang, R. C. Shallcross, N. Li, T. Stubhan, Y. Hou, W. Chen, T. Ameri, M. Turbiez, N. R. Armstrong and C. J. Brabec, Adv Energy Mater, 2016, DOI: 10.1002/aenm.201502195, n/a-n/a.
185. H. B. Wang, P. Amsalem, G. Heimel, I. Salzmann, N. Koch and M. Oehzelt, Adv
Mater, 2014, 26, 925-930.
186. Murgatro.Pn, J Phys D Appl Phys, 1970, 3, 151-156.
187. J. H. Li, J. S. Huang and Y. Yang, Appl Phys Lett, 2007, 90, 173505.
188. L. J. Brillson and Y. C. Lu, J Appl Phys, 2011, 109, 121301.
189. F. Verbakel, S. C. J. Meskers and R. A. J. Janssen, Appl Phys Lett, 2006, 89, 102103.
190. P. Schulz, L. L. Kelly, P. Winget, H. Li, H. Kim, P. F. Ndione, A. K. Sigdel, J. J. Berry, S. Graham, J. L. Bredas, A. Kahn and O. L. A. Monti, Adv Funct Mater, 2014, 24, 7381-7389.
191. L. L. Kelly, D. A. Racke, H. Kim, P. Ndione, A. K. Sigdel, J. J. Berry, S. Graham, D. Nordlund and O. L. A. Monti, Adv Mater, 2015, 28, 3960-3965.
192. M. P. Seah and W. A. Dench, Surf Interface Anal, 1979, 1, 2-11.
193. O. Janzen and W. Monch, J Phys-condens Mat, 1999, 11, L111-L118.
194. Z. L. Guan, J. B. Kim, H. Wang, C. Jaye, D. A. Fischer, Y. L. Loo and A. Kahn, Org
Electron, 2010, 11, 1779-1785.
195. F. K. Shan and Y. S. Yu, J Eur Ceram Soc, 2004, 24, 1869-1872.
158
196. B. B. Chen, X. F. Qiao, C. M. Liu, C. Zhao, H. C. Chen, K. H. Wei and B. Hu, Appl
Phys Lett, 2013, 102.
197. D. Ma, M. L. Lv, M. Lei, J. Zhu, H. Q. Wang and X. W. Chen, Acs Nano, 2014, 8, 1601-1608.
198. H. Kang, S. Kee, K. Yu, J. Lee, G. Kim, J. Kim, J. R. Kim, J. Kong and K. Lee, Adv
Mater, 2015, 27, 1408-1413.
199. M. Hösel, D. Angmo, R. R. Søndergaard, G. A. dos Reis Benatto, J. E. Carlé, M. Jørgensen and F. C. Krebs, Advanced Science, 2014, 1, 1400002.
200. W. Y. Tan, R. Wang, M. Li, G. Liu, P. Chen, X. C. Li, S. M. Lu, H. L. Zhu, Q. M. Peng, X. H. Zhu, W. Chen, W. C. H. Choy, F. Li, J. B. Peng and Y. Cao, Adv Funct Mater, 2014, 24, 6540-6547.
201. H. Zhang, W.-Y. Tan, S. Fladischer, L. Ke, T. Ameri, N. Li, M. Turbiez, E. Spiecker, X.-H. Zhu, Y. Cao and C. J. Brabec, J Mater Chem A, 2016, DOI: 10.1039/C6TA00391E.
202. K. Yao, L. Chen, X. Chen and Y. W. Chen, Chem Mater, 2013, 25, 897-904.
203. A. C. Baudouin, J. Devaux and C. Bailly, Polymer, 2010, 51, 1341-1354.
204. B. Friedel, P. E. Keivanidis, T. J. K. Brenner, A. Abrusci, C. R. McNeill, R. H. Friend and N. C. Greenham, Macromolecules, 2009, 42, 6741-6747.
205. Y. H. Kim, C. Sachse, M. L. Machala, C. May, L. Muller-Meskamp and K. Leo, Adv
Funct Mater, 2011, 21, 1076-1081.
206. N. Li, D. Baran, G. D. Spyropoulos, H. Zhang, S. Berny, M. Turbiez, T. Ameri, F. C. Krebs and C. J. Brabec, Adv Energy Mater, 2014, 4, 1400084.
207. J. W. Jung, J. W. Jo and W. H. Jo, Adv Mater, 2011, 23, 1782-1787.
208. M. Jorgensen, K. Norrman and F. C. Krebs, Sol Energ Mat Sol C, 2008, 92, 686-714.
209. C. N. Hoth, P. Schilinsky, S. A. Choulis and C. J. Brabec, Nano Lett, 2008, 8, 2806-2813.
210. K. Norrman, M. V. Madsen, S. A. Gevorgyan and F. C. Krebs, J Am Chem Soc, 2010, 132, 16883-16892.
211. K. Kawano, R. Pacios, D. Poplavskyy, J. Nelson, D. D. C. Bradley and J. R. Durrant, Sol Energ Mat Sol C, 2006, 90, 3520-3530.
212. K. Norrman, S. A. Gevorgyan and F. C. Krebs, Acs Appl Mater Inter, 2009, 1, 102-112.
159
213. A. Kojima, K. Teshima, Y. Shirai and T. Miyasaka, J Am Chem Soc, 2009, 131, 6050-6051.
214. G. Hodes, Science, 2013, 342, 317-318.
215. B. Conings, L. Baeten, C. De Dobbelaere, J. D'Haen, J. Manca and H. G. Boyen, Adv
Mater, 2014, 26, 2041-2046.
216. M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami and H. J. Snaith, Science, 2012, 338, 643-647.
217. N. J. Jeon, H. G. Lee, Y. C. Kim, J. Seo, J. H. Noh, J. Lee and S. I. Seok, J Am Chem
Soc, 2014, 136, 7837-7840.
218. S. Ryu, J. H. Noh, N. J. Jeon, Y. C. Kim, S. Yang, J. W. Seo and S. I. Seok, Energ
Environ Sci, 2014, 7, 2614-2618.
219. T. M. Khan, Y. H. Zhou, A. Dindar, J. W. Shim, C. Fuentes-Hernandez and B. Kippelen, Acs Appl Mater Inter, 2014, 6, 6202-6207.
220. J. Seo, S. Park, Y. C. Kim, N. J. Jeon, J. H. Noh, S. C. Yoon and S. I. Sang, Energ
Environ Sci, 2014, 7, 2642-2646.
221. Q. Wang, Y. C. Shao, Q. F. Dong, Z. G. Xiao, Y. B. Yuan and J. S. Huang, Energ
Environ Sci, 2014, 7, 2359-2365.
222. J. Y. Jeng, Y. F. Chiang, M. H. Lee, S. R. Peng, T. F. Guo, P. Chen and T. C. Wen, Adv
Mater, 2013, 25, 3727-3732.
223. M. Gratzel, Nat Mater, 2014, 13, 838-842.
224. H. Zhang, H. Azimi, Y. Hou, T. Ameri, T. Przybilla, E. Spiecker, M. Kraft, U. Scherf and C. J. Brabec, Chem Mater, 2014, 26, 5190-5193.
225. S. Stolz, M. Scherer, E. Mankel, R. Lovrincic, J. Schinke, W. Kowalsky, W. Jaegermann, U. Lemmer, N. Mechau and G. Hernandez-Sosa, Acs Appl Mater Inter, 2014, 6, 6616-6622.
226. J. W. Seo, S. Park, Y. C. Kim, N. J. Jeon, J. H. Noh, S. C. Yoon and S. I. Seok, Energ
Environ Sci, 2014, DOI: 10.1039/C4EE01216J.
227. Q. Wang, Y. Shao, Q. Dong, Z. Xiao, Y. Yuan and J. Huang, Energ Environ Sci, 2014, 7, 2359-2365.
228. J. H. Noh, N. J. Jeon, Y. C. Choi, M. K. Nazeeruddin, M. Gratzel and S. I. Seok, J
Mater Chem A, 2013, 1, 11842-11847.
229. Y. Hou, H. Zhang, W. Chen, S. Chen, C. O. R. Quiroz, H. Azimi, A. Osvet, G. J. Matt,
160
E. Zeira, J. Seuring, N. Kausch-Busies, W. Lovenich and C. J. Brabec, Adv Energy Mater, 2015, 5.
230. E. L. Unger, E. T. Hoke, C. D. Bailie, W. H. Nguyen, A. R. Bowring, T. Heumuller, M. G. Christoforo and M. D. McGehee, Energ Environ Sci, 2014, 7, 3690-3698.
231. H. S. Kim and N. G. Park, J Phys Chem Lett, 2014, 5, 2927-2934.
232. K. Wojciechowski, S. D. Stranks, A. Abate, G. Sadoughi, A. Sadhanala, N. Kopidakis, G. Rumbles, C. Z. Li, R. H. Friend, A. K. Y. Jen and H. J. Snaith, Acs Nano, 2014, 8, 12701-12709.
233. S. De Wolf, J. Holovsky, S. J. Moon, P. Loper, B. Niesen, M. Ledinsky, F. J. Haug, J. H. Yum and C. Ballif, J Phys Chem Lett, 2014, 5, 1035-1039.
161
Appendix B
Publications and Presentations
Publications
2012-2016
1. Zhang, H.; Shallcross, R. C.; Li, N.; Stubhan, T.; Hou, Y.; Chen, W.; Ameri, T.; Turbiez, M.; Armstrong, N. R.; Brabec, C. J., Overcoming Electrode-Induced Losses in Organic Solar Cells by Tailoring a Quasi-Ohmic Contact to Fullerenes via Solution-Processed Alkali Hydroxide Layers. Advanced Energy Materials 2016, DOI:10.1002/ aenm.201502195 .
2. Zhang, H.; Tan, W.-Y.; Fladischer, S.; Ke, L.; Ameri, T.; Li, N.; Turbiez, M.; Spiecker, E.; Zhu, X.-H.; Cao, Y.; Brabec, C. J., Roll to roll compatible fabrication of inverted organic solar cells with a self-organized charge selective cathode interfacial layer. Journal of Materials Chemistry A 2016, DOI: 10.1039/C6TA00391E.
3. Ke, L.; Min, J.; Adam, M.; Gasparini, N.; Hou, Y.; Perea, J. D.; Chen, W.; Zhang, H.; Fladischer, S.; Sale, A.-C.; Spiecker, E.; Tykwinski, R. R.; Brabec, C. J.; Ameri, T., A Series of Pyrene-Substituted Silicon Phthalocyanines as Near-IR Sensitizers in Organic Ternary Solar Cells. Advanced Energy Materials 2016, DOI: 10.1002/aenm.201502355.
4. Hou, Y.*; Zhang, H.*; Chen, W.; Chen, S.; Quiroz, C. O. R.; Azimi, H.; Osvet, A.; Matt, G. J.; Zeira, E.; Seuring, J.; Kausch-Busies, N.; Lovenich, W.; Brabec, C. J., Inverted, Environmentally Stable Perovskite Solar Cell with a Novel Low-Cost and Water-Free PEDOT Hole-Extraction Layer. Advanced Energy Materials 2015, 5, (15), DOI: 10.1002/aenm.201500543. (*These authors contributed equally to this work)
5. Adams, J.; Spyropoulos, G. D.; Salvador, M.; Li, N.; Strohm, S.; Lucera, L.; Langner, S.; Machui, F.; Zhang, H.; Ameri, T.; Voigt, M. M.; Krebs, F. C.; Brabec, C. J., Air-processed organic tandem solar cells on glass: toward competitive operating lifetimes. Energy
& Environmental Science 2015, 8, (1), 169-176.
162
6. Azimi, H.; Ameri, T.; Zhang, H.; Hou, Y.; Quiroz, C. O. R.; Min, J.; Hu, M. Y.; Zhang, Z. G.; Przybilla, T.; Matt, G. J.; Spiecker, E.; Li, Y. F.; Brabec, C. J., A Universal Interface Layer Based on an Amine-Functionalized Fullerene Derivative with Dual Functionality for Efficient Solution Processed Organic and Perovskite Solar Cells. Advanced
Energy Materials 2015, 5, (8).
7. Zhang, H.; Azimi, H.; Hou, Y.; Ameri, T.; Przybilla, T.; Spiecker, E.; Kraft, M.; Scherf, U.; Brabec, C. J., Improved High-Efficiency Perovskite Planar Heterojunction Solar Cells via Incorporation of a Polyelectrolyte Interlayer. Chemistry of Materials 2014, 26, (18), 5190-5193.
8. Zhang, H.; Stubhan, T.; Li, N.; Turbiez, M.; Matt, G. J.; Ameri, T.; Brabec, C. J., A solution-processed barium hydroxide modified aluminum doped zinc oxide layer for highly efficient inverted organic solar cells. Journal of Materials Chemistry A 2014, 2, (44), 18917-18923.
9. Li, N.; Baran, D.; Spyropoulos, G. D.; Zhang, H.; Berny, S.; Turbiez, M.; Ameri, T.; Krebs, F. C.; Brabec, C. J., Environmentally Printing Efficient Organic Tandem Solar Cells with High Fill Factors: A Guideline Towards 20% Power Conversion Efficiency. Advanced
Energy Materials 2014, 4, (11).
10. Min, J.; Zhang, H.; Stubhan, T.; Luponosov, Y. N.; Kraft, M.; Ponomarenko, S. A.; Ameri, T.; Scherf, U.; Brabec, C. J., A combination of Al-doped ZnO and a conjugated polyelectrolyte interlayer for small molecule solution-processed solar cells with an inverted structure. Journal of Materials Chemistry A 2013, 1, (37), 11306-11311.
2009-2012
11. Zhang, H.; Xu, M. F.; Cui, R. L.; Guo, X. H.; Yang, S. Y.; Liao, L. S.; Jia, Q. J.; Chen, Y.; Dong, J. Q.; Sun, B. Y., Enhanced performance of inverted organic photovoltaic cells using CNTs-TiOX nanocomposites as electron injection layer. Nanotechnology 2013, 24, (35).
12. Guo, X. H.; Yang, S. Y.; Cui, R. L.; Hao, J.; Zhang, H.; Dong, J. Q.; Sun, B. Y., Application of polyhydroxylated fullerene derivatives in hemoglobin biosensors with enhanced antioxidant capacity. Electrochemistry Communications 2012, 20, 44-47.
13. Hao, J.; Guan, L. H.; Guo, X. H.; Lian, Y. F.; Zhao, S. X.; Dong, J. Q.; Yang, S. Y.; Zhang, H.; Sun, B. Y., Interaction Between Fullerenes and Single-Wall Carbon Nanotubes: The Influence of Fullerene Size and Electronic Structure. Journal of Nanoscience and
Nanotechnology 2011, 11, (9), 7857-7862.
163
14. Zhao, S. X.; Zhang, J.; Dong, J. Q.; Yuan, B. K.; Qiu, X. H.; Yang, S. Y.; Hao, J. A.; Zhang, H.; Yuan, H.; Xing, G. M.; Zhao, Y. L.; Sun, B. Y., Scanning Tunneling Microscopy Investigation of Substrate-Dependent Adsorption and Assembly of Metallofullerene Gd@C-82 on Cu(111) and Cu(100). Journal of Physical Chemistry C 2011, 115, (14), 6265-6268.
15. Zhao, S. X.; Zhang, J.; Guo, X. H.; Qiu, X. H.; Dong, J. Q.; Yuan, B. K.; Ibrahim, K.; Wang, J. O.; Qian, H. J.; Zhao, Y. L.; Yang, S. Y.; Hao, J.; Zhang, H.; Yuan, H.; Xing, G. M.; Sun, B. Y., Structural change of metallofullerene: an easier thermal decomposition. Nanoscale 2011, 3, (10), 4130-4134.
Presentations
1. A Conjugated Polyelectrolyte as an Interfacial Dipole Layer Modified Aluminum Doped Zinc Oxide Layer for Highly Efficient Inverted Organic Solar Cells, 2
nd
International Congress Next Generation Solar Energy -From Fundamental to
Applications (Dec. 9-12, 2013) Erlangen, Germany (Poster Presentation).
2. Solution-Processed Barium Hydroxide as an Interfacial Dipole Layer Modified Aluminum Doped Zinc Oxide Layer for Highly Efficient Inverted Organic Solar Cells, E-
MRS 2014 Spring Meeting & Exhibit (May 26-30, 2014) Lille, France (Oral Presentation).
3. Interface engineering from Organic Solar cells to Perovskite Solar Cells via Incorporation of a Polyelectrolyte Interlayer, 2015 International Conference on Molecular
Electronic Materials and Devices (January 5-8, 2015) Hong Kong, China (Oral
Presentation).
4. Making Ohmic Cathode Contacts to Inverted Organic Solar Cells and Perovskite Solar Cells via Solution-Processed Interfacial Materials, 11
th International Conference on
Organic Eletronics 2015 (June 15-17, 2015) Erlangen, Germany (Poster Presentation).
5. Energy level alignment of [6, 6]-phenyl C61 butyric acid methyl ester (PCBM) on Indium tin oxide (ITO) and modified ITO surfaces, 11
th International Conference on
Organic Eletronics 2015 (June 15-17, 2015) Erlangen, Germany (Poster Presentation).
164
6. Overcoming Light-soaking in Organic Solar Cells by Forming Quasi-Ohmic Cathode Contacts to Fullerenes via Solution-Processed Alkali Hydroxide, MRS 2015 Fall Meeting &
Exhibit (November 29-December 4, 2015) Boston, USA (Oral Presentation).
![arXiv · arXiv:1506.08615v1 [math.OC] 23 Jun 2015 Vom Fachbereich Mathematik der Technischen Universität Kaiserslautern zur Verleihung des akademischen Grades Doktor der Naturwissenschaft](https://static.fdocuments.in/doc/165x107/601213e81c1e716ac440b103/arxiv-arxiv150608615v1-mathoc-23-jun-2015-vom-fachbereich-mathematik-der-technischen.jpg)
