The relation between morphology and electrical

properties in organic semiconductors

Alexandre Lira Foggiatto

July 2019
The relation between morphology and electrical
properties in organic semiconductors

Alexandre Lira Foggiatto

Doctoral Program in Applied Physics

Submitted to the Graduate School of

Pure and Applied Sciences
in Partial Fulfillment of the Requirements
for the Degree of Doctor of Philosophy in
Engineering

at the
University of Tsukuba
Abstract

Organic solar cells have achieved great attention in the past years, due to the low-cost production and
environmentally friendly materials. The efficiency has been increased higher than 10% in the past years.
However, this improvement has been achieved by the development of new materials and structural en-
gineering. However, many fundamental issues are still unclear, thus the improvement of organic devices
relays on the understanding of fundamental concepts in organic semiconductors.
Organic semiconductors are usually more disordered than the inorganic ones due to the van der Waals
interactions between the molecules. The defects in organic films may cause gap states that limit the
efficiency of organic solar cells. Molecules with high permanent dipole are more susceptible to this effect,
the misalignment of the molecules may induce gap states that affect the charge carrier injection barrier
and band bending.
In this work, the relation between morphology and electrical properties is examined for materials used
in the active layer of organic solar cells. Due to the high permanent dipole of SubPc, the alignment of
the molecules in the film can modify energy-level alignment, causing the huge open circuit voltage loss in
the SubPc/6T active layer devices.
Photoelectron and absorption spectroscopy techniques are combined to study how modification in the
morphology of the acceptor layer can reduce the energy loss. An accumulation of charge and band bending
at the donor-acceptor interface were observed in the SubPc/6T heterojunction that could be suppressed
when the SubPc film was annealed or when it was exchanged to a molecule with lower permanent dipole
(Cl6 SubPc). By using high-sensitivity ultraviolet photoelectron spectroscopy, It was observed that the
annealed SubPc film and Cl6 SubPc film have a narrower width of the highest occupied molecular orbital
(HOMO) states, which suggest a controlling of defects and gap states that reduce the states available
to the charge be transferred. Thus the thermodynamic equilibrium is achieved with lesser charges to be
transferred.

1
Contents

1 Introduction 4
1.1 Energy demand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2 Solar cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.3 Organic semiconductors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.4 Organic solar cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.5 Topics and goals of this work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.5.1 Energy-level alignment (ELA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.5.2 Density of gap states (DOGS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.6 Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

2 Materials and characterization methods 17

2.1 Used materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.1.1 Organic materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.1.2 Inorganic materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.1.3 Organic material purification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.2 Spectroscopy techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.2.1 Ultraviolet photoelectron spectroscopy (UPS) . . . . . . . . . . . . . . . . . . . . . 21
2.2.2 X-ray photoelectron spectroscopy (XPS) . . . . . . . . . . . . . . . . . . . . . . . . 22
2.2.3 Near edge X-ray absorption fine structure (NEXAFS) . . . . . . . . . . . . . . . . 24
2.3 Morphology characterization methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

3 Electronic proprieties at donor-acceptor interface 29

3.1 SubPc/6T/substrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.1.1 SubPc/6T/MoO3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.1.2 SubPc/6T/Cs2 CO3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.1.3 SubPc/6T/SiO2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.2 SubPc tailing states control summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.3 Cl6 SubPc/6T/substrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.4 High-sensitivity spectroscopic measurement . . . . . . . . . . . . . . . . . . . . . . . . . . 47

4 Study of local structure of organic thin-films 52

4.1 SubPc thin-film . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4.1.1 As-grown SubPc thin film . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

2
CONTENTS 3

4.1.2 Annealed SubPc thin-film . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

4.2 Cl6 SubPc thin-film . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
4.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58

5 Influence of the molecular orientation on the energy levels 61

5.1 SubPc orientation dependence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
5.2 Cl6 SubPc orientation dependence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

6 Study of local structure in crystalline rubrene 67

6.1 introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
6.2 STXM image . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
6.2.1 Grain boundary analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
6.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

7 Conclusion 78
Chapter 1

Introduction

The Sun produces free and worldwide available energy that can be harvest through photovoltaic devices.
Organic photovoltaic devices are the cleaner and cheaper option among the photovoltaic devices because
the production does not demand much energy and the organic molecules are mainly composed of carbon,
nitrogen, and hydrogen. Up to date, the efficiency of these devices has been measured as 17% [1]. The

Figure 1.1: Organic photovoltaic device with 12% efficiency[2].

rapid evolution in the efficiency of the organic device has been achieved by the development of new
materials and structural engineering. However, the efficiency can be raised more by modification in the
improvement in the molecular arrangement. Since fundamental concepts of organic semiconductors as
charge transport and transfer differ significantly from the inorganic counterpart, many fundamental issues
are still unclear. Thus the improvement of organic devices relies on the understanding of the fundamental
concepts of organic semiconductors.
In this section, the basis of the organic solar cells and organic semiconductor is present.

1.1 Energy demand

The energy demand is increasing as the population on Earth is growing faster and modern lifestyle is
becoming more comfortable. To keep our daily life we need to consume energy. As an example, according

4
CHAPTER 1. INTRODUCTION 5

to U.S Energy Information Administration, in the U.S., the average consumption per capita per month
in 2018 was 867 kWh [3]. However, usually, no one thinks about energy sources.
As observed in Fig. 1.2, for 1973 to 2016 the energy supply more than duplicate, however, the main
sources are still fossil fuels (coal, oil, and natural gas), which the percentage is still higher than 80 % [4].
This is a serious problem because they emit greenhouse gases that contribute to global warming. Also,
the relation between greenhouse gases and the health problems as bronchitis, cancer and heart attack has
been studied for years[5]. Although all these drawbacks, the emission is still increasing[6]. Overall, the
development of sustainable and renewable energy sources are beneficial for all society.

Figure 1.2: World total primary energy supply by fuel from 1871 to 2014. Mtoe: million tonnes of oil
equivalent [4].

In recent years, the use of renewable energy source, although still low, is increasing as displayed in
Fig. 1.3. The most used source is hydropower, which can be harmful to the environment owing to the
large area flooded [7]. Among the other sources, geothermal and ocean-related ones are limited by the
environment, since the first need to be near to hot springs, and the ocean-related ones must be on the
shore areas. Solar photovoltaic energy has the advantage that is not limited by the location and it can be
used on a small scale to produce independent energy, differently from the wind sources that need large
areas. The other advantage of solar cells is that it can be used in diverse applications, from satellites to
wearables.

1.2 Solar cells

The solar cells are one of the most promising alternatives to replace fossil fuels. Although the prices are
still not low enough to replace the other energy sources, with the increase in efficiency and improvement
of production methods, the photovoltaic devices can dethrone the fossil fuels.
The first solar cell was reported in 1839 by Edmond Becquerel. Becquerel discovered that platinum
electrodes coated with silver chloride or silver bromide can generate current and voltage when it is illumi-
nated. Although he couldn’t explain that time, he discovered the photoelectric effect and the operating
system of a solar cell. However, it was in 1954 that the first practical silicon solar cell was demonstrated
CHAPTER 1. INTRODUCTION 6

Figure 1.3: World electrical generation from renewables by source form 1990 to 2016 [4].

and the efficiency was approximately 6% [8]. Since that time photovoltaic research has achieved great
progress [9]. By now the research has gone through three generations and the best deficiencies are dis-
played on the Fig. 1.4. These are the best photovoltaic devices in each category certificated by the
National Renewable Energy Laboratory (NREL).
The first generation uses silicon as the main material and can be divided into two production chains:
single crystal silicon (m-Si) and polycrystalline silicon (p-Si) [11]. These are the most used type of
technology, occupying about 85% of the market (reaching 15-20% of normal performance) due to the
high achievable efficiency, and its superior performance and stability. However, these cell types are rigid
and require a lot of energy during their manufacture. In single-crystal silicon, the entire structure is
composed of the same material, so the molecular structure is uniform. This type of uniformity is ideal
for the efficient transport of electrons through the material. However, in order to the cell achieves good
efficient, it is necessary for silicon to undergo a doping process to produce layers of type p and n doped
silicon reaches a purity level of 99.9999% after undergoing a purification process. While this is expensive,
it is crucial to the cell performance the control of impurities, which is directly connected to cell efficiency
of m-Si devices [12]. On the other hand, p-Si uses simpler and cheaper manufacturing processes than
m-Si silicon manufacturing technology. However, the quality of the material is low due to high levels of
impurities.
The second-generation solar cells, also called thin films, can be divided into three production chains.
Amorphous silicon (a-Si) [13], Copper indium selenide (CIS) [14] and Copper indium gallium selenide
CHAPTER 1. INTRODUCTION 7

Figure 1.4: Evolution of the efficiency for solar cells divided by technology from 1975 to 2019 [10].

(CIGS) [15], and cadmium telluride (CdTe) [16]. This technology occupies about 12% of the market.
Since these materials have good absorption of solar radiation, their structure can be very thin (thickness
of about 1 µm). So, the amount of semiconductor used is less. This type of cell has the potential to
become cheaper than silicon. Besides, the production is done at lower temperature, which contributes to
the low power consumption in the manufacturing process. Another important factor is that the substrate
is flexible, and the applicability of this type of cell in architectural projects is currently targeted. Although
the manufacturing processes contribute to environmental pollution, the low availability of raw materials
makes this type of cell commercially unattractive.
Organic photovoltaic devices are in the third generation of solar cells [17]. They have many advantages
over other technologies owing to lower production costs, flexibility, and lightweight. The highest efficiency
of the organic devices certificated by NREL is 15.6% (Fig. 1.4).

1.3 Organic semiconductors

Organic materials are separated into two groups, the conjugated molecules and unconjugated one. Con-
jugated molecules are the ones that have a double bond with a unique bond between the atoms. The
conjugated molecules are important for organic semiconductor research because they have delocalized
electron, which is the basic requirement of electrical conductivity from an applied electric field. However,
still, the electron transport capacity originated from the delocalization is much lower in organic materials
than in inorganic semiconductor.
A σ bond is formed when two carbon atoms with sp2 orbitals bond to each other using the sp2
orbital. When the hybridization occurs, each carbon remains with one unhybridized 2pz orbital which
CHAPTER 1. INTRODUCTION 8

has a perpendicular direction to the hybridized lobes and has one electron that overlaps forming a π
bonding, as shown in Fig. 1.5. π bonding and a π antibonding orbitals are formed when the two occupied
2pz orbitals (one electron each) overlaps producing the energy level diagram as displayed in Fig. 1.6.
After the π bonding orbital is formed, in the plane region between the nuclei of carbon, it occurred an
increasing of the density of electrons. The region between the molecules is where the π ∗ orbital can be
found and it has a nodal plane in the perpendicular direction of the molecules axis. Since the 2pz orbital
has only one electron each, only one bounding level can be filled and the other antibonding level stays
unoccupied. Thus, the double bond in ethylene is formed from a σ bond and a π bond.

Figure 1.5: (a) σ bond and (b) π bond in ethylene (C2 H4 ).

These two orbital are called highest occupied molecular orbital (HOMO) and lowest unoccupied molec-
ular orbital (LUMO), which can be roughly associated with the valence and conduction band, respectively,
in inorganic semiconductors. The bandgap of organic semiconductors is in the range of 1.5 and 3 eV.
Since the electrons in these orbitals are delocalized, they can contribute to the transport.

1.4 Organic solar cells

Organic photovoltaic devices can be manufactured in three types: single layer, bilayer and bulk devices. A
single layer of anthracene was reported as the primal organic solar cell in 1959 [18]. However, only in 1986
that Tang reported the first device using the concept of donor/acceptor layer.[17] The device achieved the
efficiency of 0.95%. Nowadays, with the search for new materials, device structure and interface studies,
the efficiency of the organic devices have crossed above 10% [19, 20].
Figure 1.7 (a) displays the conventional heterojunction solar cell. The active layer is formed by the
donor and acceptor layers. The materials selected as acceptor are the ones with higher electron affinity
and the ones with lower electron affinity are selected as donors. The hole and electron transport layers
CHAPTER 1. INTRODUCTION 9

Figure 1.6: Diagram displaying the energy levels a bonding(π) and antibonding(π ∗ ) and its respective
molecular orbital for ethylene.

Figure 1.7: (a) Schematic draw of the conventional heterojunction solar cell. (b) Work principle of organic
solar cell.

are used to improve charge separation and collection.

The working principle of the organic solar cell is displayed in Fig. 1.7 (b). When a photon is absorbed
in the organic materials, an electron is excited from the HOMO to LUMO creating an electron-hole pair.
Then, the electron-hole pair relax with a binding energy of between 0.2 and 1.4 eV, which is called exciton.
When the exciton diffuses to the donor/acceptor interface, the electron transfers to the acceptor layer
and the hole stays in the donor layer. Thus, an energy difference of around 0.5 eV in the LUMO levels is
CHAPTER 1. INTRODUCTION 10

required to disassociate the exciton.

The exciton dissociation creates a hole polaron in the HOMO of the donor and electron polaron in
the LUMO of the acceptor material. They are called germinate electron-hole pair and are still weakly
bonded, thus, depending on the distance, they still can recombine due to Coulomb interaction. However,
to generate a photocurrent, the germinate electron-hole pair must completely disassociate to the charges
be collected at the electrodes.
The selection of the preferable donor and acceptor layer is utmost important to enhance the light
absorption, exciton diffusion and separation and charge transport. For small molecules, the materials
used as a donor for photovoltaic applications are usually copper(II) phthalocyanine (CuPc), zinc ph-
thalocyanine (ZnPc), boron subphthalocyanine chloride (SubPc), lead phthalocyanine (PbPc), pentacene
and diindenoperylene (DIP). On the other hand, for the acceptor layer, fullerene (C60 ) is the most used
material owing to the acceptor property and, also it has high electron mobility.

Figure 1.8: (Upper) Absorption spectrum of C60 and C70 . (Down)(a) Structure and (b) energy level of
solar cell comparing C60 acceptor with a generic donor [21].

Although the outstanding transport properties of fullerene, it has some problems that limit the effi-
ciency of photovoltaic devices. First, the main absorption of fullerene is outside the visible spectrum, as
shown in Fig. 1.8. The solar spectrum emission is from around 400 nm to 700 nm and the maximum is
localized approximately around 500 nm. Since the overlap between the absorption of fullerene and emis-
sion of the sun is small, the exciton formation is limited which controls the maximum efficiency. Another
CHAPTER 1. INTRODUCTION 11

problem of fullerene is the deep HOMO (-5.1 eV), which reduces the open-circuit voltage (Voc ). Voc can
be associated with the HOMO/LUMO by the photovoltaic gap (EP V G ), which is the difference between
the HOMO of the donor material and the LUMO of the acceptor. The relation between them is given by
the following equation [22]:

1
Voc = (EP V G − Eloss ), (1.1)
e
where Eloss is the energy loss affected by exciton dissociation, charge collection, recombination. Thus
the replacement of fullerene is a way to improve the power conversion efficiency in organic photovoltaic
devices.
In the past years, many researchers are interested in producing OPV devices with higher efficiency
through the development of free-fullerene OPV devices [23, 24, 25, 26]. Recently, a device that achieves
a huge efficiency (for small molecule) of 8.4 % using α-sexithiophene (6T) in the donor layer and a com-
bination of boron subphthalocyanine chloride (SubPc) and boron subnaphthalocyanine chloride (SubNc)
in the acceptor layer, has been reported [27]. Despite the astonishing performance, the energy loss in the
Voc in SubPc/6T device has been reported as 0.8 eV, which is two times higher than the C60 /6T device.
Since, the non-planar solar cells are extremely affected by the orientation of molecules on the film [28],
the improvement of the morphology in these molecules may improve the efficiency of these devices.

1.5 Topics and goals of this work

To enhance the efficiency of organic solar cells, the replacement of fullerene to non-planar molecules is
necessary. However, for non-planar molecules, the orientation of them in the film affects the efficiency. In
this work, the relation between morphology and electrical properties of high dipole molecules is studied.

1.5.1 Energy-level alignment (ELA)

Organic photovoltaic device usually has many layers between the anode and cathode to improve the
exciton creation, carrier extraction, etc. On the other hand, there are many interfaces that the charge
carrier must cross to generate the photocurrent. Thus, the study of the electronic structure is utmost
important to improve the efficiency of the organic solar cells.
The physics behind the organic-metal and organic-organic interfaces have been extensively studied to
understand the mechanism of the energy-level alignment (ELA) in organic semiconductor[29, 30, 31, 32].
However, complete understanding is missing.
In organic devices, such as solar cells, light-emitting diodes, transistor, the performance highly depends
on the properties of the active layer materials. Moreover, to produce a good device, the properties of
the active layer cannot be isolated, the interface plays an important role. However, researchers are still
focusing on developing new materials to produce better devices, instead of understanding the interfacial
properties. When the interfacial study and electronic properties are taken into account, interesting results
are observed. As an example, one may notice that the open-circuit voltage is limited by the charge
separations due to organic-organic and organic-metal interfaces. The ELA at these interfaces must be
completely understood to obtain the maximum results in organic electronic devices.
CHAPTER 1. INTRODUCTION 12

Organic-organic heterojunction interfaces

There are two well-known models to describe organic/organic and organic/metal interfaces. The first
one mention here will be the integer charge-transfer (ICT) model. The interaction between the substrate
electronic level and molecular orbital is assumed to be negligible in the ICT model. This model has been
successfully used in many organic electronic cases.
The ICT model estimates the energy-level alignment at interfaces using the substrate work function
(ΦSU B ) and the integer charge transfer states (EICT ). The EICT + and EICT − are the energy necessary to
oxidize or reduce the organic molecule at the interface. The determination of the energy-level alignment
occurs due to the equilibrium of the chemical potential by the relative position of the substrate work
function and the ICT states.

Figure 1.9: Schematic illustration of the evolution of the energy-level alignment when a π-conjugated
organic molecule or polymer is physisorbed on a substrate surface when a) ΦSU B > EICT + : Fermi-level
pinning to a positive integer charge-transfer state, b) EICT − < ΦSU B < EICT + : vacuum level alignment,
and c) EICT − < ΦSU B : Fermi-level pinning to a negative integer charge-transfer state. The charge-
transfer-induced shift in vacuum level, ∆, is shown where applicable. Ref. [33].

Figure 1.9 describes the three cases that can occur. If ΦSU B > EICT + , electrons transfer from the
organic layer to the substrate and a Fermi level pinning occurs to EICT + . For EICT − < ΦSU B < EICT + ,
since the Fermi level is in the center, no charge is necessary to be transferred and occurs the vacuum level
CHAPTER 1. INTRODUCTION 13

alignment. The last case is for EICT − < ΦSU B < EICT + , in this case, charge transfers from the substrate
to the organic layer and a Fermi level pinning occurs at EICT + .
This model has been applied successfully to many systems, however, it had some problem in describing
small molecule systems.

Induced density of interface states (IDS) Model

The induced density of interface states (IDIS) model can be used in the system that the chemical interac-
tion is small, but not negligible. This often occurs, as an example, of organic material evaporate in gold.
Thus, the IDS model describes interfaces with stronger interaction than the ones of ICT model scope.
The fundamental concept of the model is that the discrete energy level of isolated molecules are
broadening into a quasi-continuous density of states when the molecules contact the metal or other
organic substrates. Since no covalent bond is formed at the interface, but still there some interactions of
the upper layer with the layer below, occurs a hybridization of the HOMO and LUMO, in especial. If the
underlayer is organic, the interaction must create the hybridization on both layers at the interface. By
filling the induced density of states by the charges of the isolated molecules, the charge neutrality level
(CNL) can be found.

Figure 1.10: Energy level alignment at organic heterojunctions: the initial CNL difference is partially
screened, resulting in the formation of an interface dipole ∆OO and a smaller final CNL offset [34].

For organic-organic interfaces, the charge transfer that occurs inter-molecules at the interface is given
by the relative position of CNL for the film 1 and film 2. If the charge neutrality level for one is higher
than another, the charge will transfer to achieve the equilibrium. The interface dipole ∆OO is given by
the following equation [34]:

(CN L1 − CN L2 )f inal = SOO (CN L1 − CN L2 )initial (1.2)

∆OO = (1 − S00 )(CN L1 − CN L2 )initial (1.3)

where SOO is called the slope parameter. The separation between the CNL levels at the interface is
correlated to the parameter SOO . This parameter describes the ability of the molecular materials to
CHAPTER 1. INTRODUCTION 14

screen the electrostatic-potential difference. Although it has been described that S ≈ (1/1 + 1/2 ), where
 is the dielectric function, SOO cannot be precisely deduced and must be extracted from experimental
data. Although this description works, it is still unknown how this continuous density of states occurs at
organic/organic interfaces, since the interactions are not strong enough and it has never been observed
[35].

1.5.2 Density of gap states (DOGS)

In the IDIS model, to generate gap states it is necessary that interaction between the substrate and the
organic layer be chemically strong. However, due to the van der Walls interaction between the molecules
to form a material, the HOMO and LUMO energy levels are not more discrete. And the disordered
molecules and structural imperfection lead to very small HOMO-LUMO DOS that causes the broadening
and tailing on the HOMO and LUMO [36, 37].
For inorganic materials, the energy gap is well defined and the valence and conduction band the
density of states (DOS) can be modeled by a parabolic model (Fig. 1.11). However, for organic materials,
the HOMO and LUMO DOS are modeled by a Gaussian. The HOMO DOS broadening is governed by
energetic disorder, since the full width at half maximum (FWHM) for amorphous or film with imperfection
in the ordering, is measured as higher than 0.4 eV. Thus, the tailing of the Gaussian can be represented
as the density of gap states (DOGS).

Figure 1.11: E1/2 -distributed DOS in a model inorganic semiconductor crystal with a clear energy gap
between 1D conduction and valence bands. There is a clear valence band edge (EV ) and conduction band
edge (EC ). (b) Gaussian-distributed HOMO and LUMO DOS in a model organic semiconductor film
with the energy gap between the onsets of HOMO and LUMO DOSs. We define the onset of HOMO (or
LUMO) as the energy position of the HOMO (or LUMO) peak (EH (or EL )) minus (or plus) 2σH (or 2σL ),
which is strictly obtained from the intersection between the baseline and a tangent at the inflection point
of the Gaussian DOS (see panel (b)). The HOMO and LUMO onset values should be used to determine
IEΦ and EAΦ .[38]

The DOGS are originated, primarily in high polar molecules, by defects that appear as tail states from
the HOMO and LUMO inside the region of the bandgap. Ultraviolet photoelectron spectroscopy (UPS)
has been extensively applied to understand the origin of DOGS for the HOMO region.[39]. The tailing
CHAPTER 1. INTRODUCTION 15

that occurs in the density of states (DOS) mainly appears in high polar molecules due to the arrangement
of permanent dipoles in the disordered film, inducing a local electric that originates the tailing states.
Organic films are usually disordered due to the weak interaction of molecules in the solid which affects
the electrical properties [40]. Thus, the study of the physical properties and intermolecular interactions
should be done together.
The structural defects can modify the energy-level alignment at the organic-organic interface. To
exemplify it, We use the example of perfluoropentacene (PFP)/diindenoperylene (DIP) interface reported
by Yoneza et al. (Fig.1.12) [41]. For the disordered film, when the two films are brought together, owing
to the wider DOGS, there are more available states to charge transfer to achieve the thermodynamic
equilibrium. This process causes an accumulation of charges at the interface, which is observed through
the large electrostatic potential. After the reduction of structural defects, by annealing, a reduction of
DOGS has been observed using ultraviolet photoelectron spectroscopy (UPS). Due to the reduction of
DOGS, fewer states are available to charge transfer, reducing the charges that are transferred and the
ones accumulated at the interface. Also, due to morphological modifications, the electrostatic dipole and
the HOMO-LUMO onsets have changed.

1.6 Organization

Organic photovoltaic (OPV) devices have many advantages over other technologies owing to lower produc-
tion costs, flexibility, and lightweight. Many materials have been used as donors, however, concern to the
acceptor layer, the most used materials are fullerene (C60 ) and derivatives (PCBM (Phenyl-C61-butyric
acid methyl ester)). Despite the great properties as high electron mobility and ability to accept electrons,
fullerene has a small absorption in the visible light and small photovoltaic gap with the donor layer that
controls the efficiency. Thus, the efficiency of SubPc devices could be improved by adjusting the density
of gap states at the interfaces through the molecular arrangement.
SubPc has a high permanent dipole, thus it is expected that modification on the morphology of the
film leads to modification on the electrical properties. Non-planar solar cells are extremely affected by
the orientation of molecules on the film [28]. To improve the OPV devices, the effect of modification in
morphology on the electrical properties must be clarified. To do so, this study is divided into 3 parts. The
first part is the effect of misalignment of molecules and the permanent dipole on the energy-level alignment
at a donor-acceptor interface. The electrical properties were analyzed using UPS and X-ray photoelectron
spectroscopy (XPS). The donor-acceptor interface was analyzed comparing different substrates: MoO3 ,
SiO2 and Cs2 CO3 . Next, to connect the electrical properties and the morphology, the same experiment was
conducted on the annealing the organic layers (150°C). Subsequently, the same experiment was carried out,
however, this time, SubPc was replaced for halogenated SubPc (Cl6 SubPc), which has lower permanent
dipole, as the acceptor layer. The second part of the study was the investigation of the morphology using
scanning transmission X-ray microscope (STXM) combined with near-edge X-ray absorption fine structure
(NEXAFS). The morphology of as-grown SubPc and annealed SubPc thin-film could be investigated and
compared. As well the STXM image for Cl6 SubPc was taken. And, in the last section, the relationship
between the molecular alignment of SubPc and Cl6 SubPc and the orbital energies was investigated using
density functional theory (DFT). Besides the main work, the morphology of rubrene was also studied, in
which the grain boundary features could be extracted from the STXM images.
CHAPTER 1. INTRODUCTION 16

Figure 1.12: Energy level diagrams of PFP (3.2 nm)-on-DIP (20 nm) heterostructures with different defect
densities. The thickness of colored boxes represents the width of HOMO and LUMO level with FWHM
(w) measured for HOMO (LUMO width assumed to be similar to that for HOMO). Distribution of gap
states is shown by shading. Onset-to-onset band gap of 2.55 eV for DIP is taken from a previous report.
Density of states (DOS) scheme is shown on right side [41].

This work is divided as follows: Chapter 2, entitle ”Materials and characterization methods”, presents
the methods and measurement techniques, in which contains the physical principles and experimental
setup. Chapter 3, entitle ”Electronic proprieties at donor-acceptor interface” contains the discussion about
the photoemission spectroscopy measurements and results. The morphological analysis using STXM and
image analysis is presented in Chapter 4 entitled ”Study of local structure of organic thin-films”. DFT
calculation was discussed in Chapter 5 entitled ”Influence of the molecular orientation on the energy
levels”. In Chapter 6 entitled ”Study of local structure in crystalline rubrene” the STXM image analysis
at the grain boundary of 2 types of orthorhombic polycrystalline rubrene is done. Chapter 7 is the
conclusion.
Chapter 2

Materials and characterization methods

In this chapter, the physical principals and experimental methods selected to characterize the organic
thin-film are displayed. Also, the organic and inorganic compounds used in this work are introduced
here. The chapter is divided into 3 parts. In the first one, it is addressed the compounds, purification
and growth methods. The second section focuses on photoemission spectroscopy. And, the last section
addresses the principle of STXM and the image analysis process.

2.1 Used materials

The materials used throughout this work are introduced here. To investigate the losses in the OPV
devices, the donor-acceptor interface properties must be clarified. Organic films are usually disordered
due to the van der Walls bonds between the molecules in the solid which affects the electrical properties
[40]. Especially in highly polar molecules that the DOGS are enhanced. To study the interfacial properties
at the active layer of the OPV devices, the selected organic materials are α-sexithiophene (6T), boron
subphthalocyanine chloride (SubPc) and hexachlorinated boron subphthalocyanine chloride (Cl6 Subpc).
For the inorganic materials, MoO3 and Cs2 CO3 were selected. The former has been used as a hole
transport layer [42] and the latter has been used as n-doping material in organic devices [43].

2.1.1 Organic materials

The organic materials presented in this section are 6T, SubPc, and Cl6 Subpc.

• α-sexithiophene

Thiophene is a heterocyclic compound with the formula C4 H4 S, in which this molecule and its deriva-
tives have been used in various applications, such as photovoltaic devices, thin-film transistors, and
light-emitting diodes [44]. One of the most promising oligothiophenes studied is α-sexithiophene (6T)
since it has high absorption in the visible light and charge carrier mobility [45]. 6T consists of six rings
linearly connected, as exhibited in Fig.2.1. The chemical structure of the 4 inner rings is identical, con-
sisting of C, H, and S atoms. The rings at the borders have an extra hydrogen atom each, as observed in
Fig.2.1.

• Boron phthalocyanine chloride

17
CHAPTER 2. MATERIALS AND CHARACTERIZATION METHODS 18

Figure 2.1: Molecular structures of 6T and its optimized geometry.

SubPc is mostly used as a donor material in organic photovoltaic devices combined with fullerene (C60
or C70) acceptor [46, 47], however, recently, SubPc has been used as an acceptor layer [27, 48, 49, 50].
SubPc has relatively cone-shaped molecular structure and in the center of the molecule is observed a
B–Cl atom which is attached to the main structure. This results in the high permanent electric dipole
calculated as 4.59 D. Thus the electronic properties are sensitive to the film conditions.

Figure 2.2: Molecular structures of SubPc and its optimized geometry.

• Hexachlorinated boron subphthalocyanine chloride

Halogenated SubPcs are good candidates to replace fullerene in small molecule organic photovoltaic
devices due to the deep energy levels and high light absorption in the solar spectrum [51]. It has been
reported that organic solar cells using a Cl-substituted SubPc (Cl6 SubPc) acceptor shows an open-circuit
voltage of 1.31 eV, which is higher than 1.10 eV reported for C60 acceptor and SubPc used as donor
[52]. Due to the presence of the 6 Cl atoms attached in the molecule, the permanent electric dipole is
calculated as 1.43 D, which is an alternative to SubPc.
CHAPTER 2. MATERIALS AND CHARACTERIZATION METHODS 19

Figure 2.3: Molecular structures of Cl6 SubPc and its optimized geometry.

2.1.2 Inorganic materials

• MoO3
Molybdenum trioxide (MoO3 ) is usually employed as a hole transport layer in organic photovoltaic devices
[53]. Although MoO3 is an insulator with conductivity measured as 1 × 10−7 S/cm. For thin films
(< 100nm), The oxygen vacancies create gap states closer to the Fermi level that improves the hole
transport on the valence level [54]. Furthermore, the MoO3 improves the hole transport and extraction in
the organic active layers due to the p-doping that induces a band bending at the MoO3 /organic interface
which leads to a built-in field that improves the hole extraction in OLED and reduces recombination in
organic photovoltaic devices [55].
• Cs2 CO3
Cesium carbonate (Cs2 CO3 ) has been used in organic semiconductor field as a dopant due to the low
work function (∼ 3 eV) [56]. This is caused by the decomposition of Cs2 CO3 into Cs and other gases
during the thermal evaporation [57]:
2 Cs2 CO3 −−→ 4 Cs + O2 + 2 CO2 ·

2.1.3 Organic material purification

The purification of organic material is extremely important to obtain high-quality devices. As reported by
Salzman et al., the efficiency is correlated to the number of purification circles that the organic material
has been processed [58]. The impurities modify the electrical and optical proprieties by undesired doping.
Since the impurities in the as-received materials are high, the materials cannot be directly used for organic
devices application, thus the purification of the materials is necessary. As well, it is also mandatory to
avoid contaminants to enter into the high vacuum chamber, which affects the background pressure and
increase the contamination on the films and chamber. There are several methods of purification, however,
since the materials used in this work are solids in the atmospheric pressure and bellow, the gradient
sublimation is preferable.
The purification process is displayed in Fig. 2.4. The source material obtained from a commercial
CHAPTER 2. MATERIALS AND CHARACTERIZATION METHODS 20

Figure 2.4: Schematic representation of the sublimation purification. (a) Before and (b) after the purifi-
cation.

source is loaded onto the end of the glass tube. Then, other sleeves are placed on the glass tube to separate
the material that is sublimed and impurities. To avoid that some material enter inside the pump, a wad
of wool is placed at the end of the glass tube. After finishing the setup, the tube is pumped and the
temperature increase gradually until the sublimation point of the source material. After 24 hours the
purified material is concentrated in the 3rd sleeve, the more volatile impurities move further due to the
gradient in temperature. On the other hand, the residue is composed of nonvolatile impurities. Before
installing the organic materials into the high vacuum evaporation system, they are usually purified in
more than one cycle (multiple cycles) by gradient sublimation. In this work, all the organic materials
were purified at least 4 times before loading into the chamber.

2.2 Spectroscopy techniques

Photoelectron spectroscopy (PES) is one of the most important methods to investigate the electronic
structure of condensed matter. The photoelectric effect, explained by Einstein in 1905, is the principle
behind the photoelectron spectroscopy.
The photoelectric effect proposes that the light is composed of a series of particles, called photons,
and when they interact with the surface of a metal, electrons can be emitted. If the photon energy hν is
CHAPTER 2. MATERIALS AND CHARACTERIZATION METHODS 21

greater than the work function (Φ), an electron can be emitted with kinetic energy (Ekin ) described by:

Ekin = hν − Φ (2.1)

By modifying the energy of the incident photon, different parts of the occupied states can be measured.
For ultraviolet (UV) radiation, the valence region can be observed and for higher energies, X-ray photons,
the core level can be examined.
The photoemission spectroscopy measurements were carried at the Institute for Molecular Science
(IMS) and Photon Factory (PF), High Accelerator Research Organization (KEK). The UPS is used to
observe the valence region, in which we can extract the HOMO information. On the other hand, the
XPS can give information about the core levels and band bending on the films. The study of DOGS was
carried out using the ultrahigh-sensitivity UPS, which has a low background and high sensitivity. The
gap states have been effectively investigated using this technique [59]. The LUMO was investigated using
low-energy inverse photoemission spectroscopy (LEIPS). Since the emitted electrons have low energy, the
degradation in the analyzed sample is reduced [60].

2.2.1 Ultraviolet photoelectron spectroscopy (UPS)

UPS is a powerful technique to investigate valence states. In this method, monochromatic light is used
to irradiate the sample and the distribution of electron can be measured. The radiation source used to
perform UPS is usually He lamp that emits photons with 21.2 eV (He I) and 40.8 (He II). Synchrotron
radiation is often used due to the high resolution that can be achieved. In organic semiconductors, the
topmost layer, which is in contact to vacuum, has lower polarization screening (HOMO shift due to
the low interaction of the molecules) compared to the bulk, that causes an increment in the ionization
potential. A reduction of polarization screening by around 0.3-0.4 eV has been observed [61]. However,
for synchrotron reduces this to around 100 meV [62].
UPS can be used to study the interfacial electronic states as displayed in fig. 2.5. The organic-
metal interface is exemplified in this Fig. 2.5. Figure 2.5 (a) corresponds to the metal substrate. Since
max
for a metal the occupied states go until the Fermi level (EF ), the most energetic electron (Ek(metal) ) is
originated from the EF , where the binding energy is minimum. However, by collecting electron with
lower kinetic energy, there is a moment where no electron is more emitted because they do not have the
energy to overcome the vacuum energy. The electrons with lower energy emitted are the secondary cutoff
electron (SECO), since they are the last to cross the vacuum level barrier, the work function of metal
(Φm ) can be calculated as:
max
Φm = hν − (Ek(metal) − ESECO ) (2.2)

Figure 2.5 (b) displays the spectrum of a thin layer of an organic semiconductor on the metal substrate.
Differently of the metal case, in organic semiconductor the valence electrons are not located on the EF ,
instead of it, they are on the HOMO level. The electrons with maximum energy (Ek(org)max ) are exited
from the HOMO level. So, the HOMO onset (FV ) is described as the relative position of Ek(org)max in
relation to the EF . Also, another physical quantity called ionization potential (IP) can be defined as:

IP = hν − (Ek(org)max − ESECO(org) , (2.3)

which is the minimal energy required to take an electron to free state.

CHAPTER 2. MATERIALS AND CHARACTERIZATION METHODS 22

Figure 2.5: The principle of the UPS applied to an organic/metal interface. (a) and (b) display the
photoemission from the metal and the organic layer on the metal substrate, respectively. On (c) the UPS
spectra of metal and organic layer on metal is shown.hν: photon energy, EF : Fermi energy of metal, Φm :
work function of metal, VL: vacuum level, Ekmax : maximum kinetic energy of photoelectron, FV : energy
of the HOMO relative to the Fermi level of the metal.

As one can see, after the organic layer is deposited, the vacuum level for the organic material shifted
by (∆V L ), as well the valence level. In this case, the work function for the organic material (Φorg ) is
calculated as:
Φorg = IP − Ek(org)max . (2.4)

Usually, the organic film thickness is in order of ten of nanometers thickness. A thicker layer may
compromise the experiment. The thicker organic layer may act as an insulator causing charging on the
sample. Charging is an accumulation of positive charge on the surface of the sample due to the extraction
of the electron during the measurement. This effect needs to be evaluated carefully, especially in the study
of organic-organic interfaces and multilayer systems, since the sample tends to become thicker limiting
the size of the sample.

2.2.2 X-ray photoelectron spectroscopy (XPS)

One of the most important experimental methods used to understand the properties materials is the
X-ray photoelectron spectroscopy (XPS). It is used to study the electronic structure of solids, molecules,
interfaces, and surfaces. Several important phenomena in atomic, molecular and solid state physics were
understood through this method, also, to have practical application in fields such as surface chemistry
and materials science. This technique can be used for the determination of elements and their quantities;
contamination of materials; empirical formula; the energy of connection of electronic states; state of
binding (valence) of elements in samples; thickness of different thin layers. This technique, similar to
UPS, is based on the photoelectric effect, in which electrons are ejected from an atom when it is exposed
to radiation with sufficiently high frequency. Since the photon energy used in the technique is higher than
CHAPTER 2. MATERIALS AND CHARACTERIZATION METHODS 23

the UPS, electrons from the core levels can be exited to the vacuum level. The peaks observed in the
measurement can be associated with a specific atom since each atom has a characteristic binding energy
observed in the core levels. However, it is still highly dependent on the species or molecules that this
atom is bonded with. Thus, charge transfer may shift the internal levels to higher (or lower) binding
energy, if the atoms became positive (or negative) charged due to the attractive (or repulsive) Coulomb
interaction with the nucleus.

Figure 2.6: Escape depth (nm) as a function of photoelectrons energy (eV) [63].

Although X-rays photons have penetration depth in the order of 10 nanometers, this technique is
surface sensitive. In solids, the inelastic mean free path (IMFP) is very small, which results that the
detected electrons are only from the surface up to 10 nm. The universal curve (Fig. 2.6) displays the
mean electron kinetic energy as a function of the mean free path. Electron with energy between 30 eV
to 500 eV, the mean path is around 1 nm, which makes this technique very surface sensitive for chemical
analysis.
The XPS has another advantage, the band bending caused by charge transfer and accumulation at
the interface can be assigned to the XPS peak shift as displayed in Fig. 2.7. From the Poisson equation,
the electrostatic potential Vbb bends the energy levels same amount as a function of distance from the
interface, as described by the following equation:

qVbb (x) = (Ecore (X) − EHOM O (X)) − ((EF − EHOM O (X)) − Ecore (interf.)). (2.5)

The band bending on the valence band can be calculated by the peaks shift using the core level information.
This fact can be very useful when one wants to extract the band bending at the interface of two different
materials.
To assign the band bending at a heterojunction interface using UPS can be highly complex. Since UPS
can just evaluate the valence region, the signal from different layers may overlap affecting the analysis.
However, using XPS, this feature can be overcome. Figure 2.7 exemplifies how to analyze the peak shift
CHAPTER 2. MATERIALS AND CHARACTERIZATION METHODS 24

Figure 2.7: Schematic of the band bending calculated using the core level shift information.

for layer A and B. The HOMO onset and vacuum level, for both layers, are obtained using UPS. The
XPS is used to using the position of Sa , which is an element that A is composed, and the element Sb of
molecule B. Before the first layer of B is deposited, Sa is localized at Sa (A). Upon the deposition of the
1st layer of B, due to the interface formation, the peak from layer A shifts to Sa (B, 1) and a peak from
layer B can now be assigned as Sb (B, 1). By the continuing growth of layer B, the final position of the
peaks is Sa (B) and Sb (B). In this case, the total band bending at layer A (Vb (A)) is defined as:

Vb (A) = Sa (B) − Sa (A). (2.6)

For the layer B, since layer A is still bending upon the increment of the upper layer, the total band
bending at B is calculated as:

Vb (B) = (Sa (B) − Sa (B, 1)) − Vb (A). (2.7)

2.2.3 Near edge X-ray absorption fine structure (NEXAFS)

X-ray absorption spectroscopy (XAS) is an evasive analytical technique, which involves exposing any
material to an X-ray beam, part of which is absorbed. Absorbance is then measured as the ratio of
CHAPTER 2. MATERIALS AND CHARACTERIZATION METHODS 25

Figure 2.8: Schematic of the HOMO/LUMO alignment at organic semiconductor heterointerfaces and
the resultant XPS and UPS photoemission. [64].

incident intensity to transmitted intensity, as in spectroscopy-based radiation absorption [65].

By measuring the amount of absorption by increasing the energy level of X-rays, so-called ”absorption
edges” specific to each element are revealed where the absorption intensity suddenly increases. This
happens when the x-rays have enough energy to emit or excite electrons in the material [65]. The basic
process of x-ray absorption is the excitation of electrons near the atomic nuclei by the absorption of
photons. The typical spectrum of XAS is usually divided into two energy regions (A) high-resolution
absorption edge spectroscopy (XANES) or Near edge X-ray absorption fine structure (NEXAFS) and (b)
extended X-ray absorption fine structure (EXAFS).
NEXAFS is the part of the XAS spectrum that goes from near below to the edge to around 50 eV
above the absorption edge. Figure 2.9 shows the principle of NEXAFS for organic molecules. Above the
Fermi level, there are the unoccupied molecular orbitals. By selecting the appropriate energy a transition
from the core level (e.g., C 1 s) to the unoccupied π ∗ and σ ∗ orbitals. For molecules with a double
and triple bond, the most prominent transition is usually the one to the LUMO π ∗ , which is lower the
ionization potential. Thus, the lifetime is of π ∗ is quite short and the peak is localized. On the other
hand, the σ ∗ orbital are in higher energy than the ionization potential for most of molecules composed
by carbons, nitrogen or oxygen. The resonance of this transition is related to the lifetime of quasibond
state. Due to the overlap of continuum states and the σ ∗ orbitals, the σ ∗ resonance is broader.
NEXAFS can be useful to extract the information about the orientation of the molecule. Considering
quantum mechanical, The Fermi golden rule use the dipole approximation to describe the allowed exci-
CHAPTER 2. MATERIALS AND CHARACTERIZATION METHODS 26

Figure 2.9: Principle of NEXAFS. Absorption resonances in K-shell NEXAFS spectrum arises from the
initial 1 s state to the unoccupied molecular orbital (π ∗ or σ ∗ )

tation process for a electron. The absorption cross section (σx ) for a transition from the initial state (Ψi )
to the final state (Ψf ) is given by:

σx ∝ |hΨf |e · p|Ψi i|2 ρf (E), (2.8)

where e is the unit electric field vector, p is the dipole transition operator and ρf (E) is the density of
final states.
For the linearly polarized light and considering the k-sell excitation (1 s state is spherically symmetric),
the angular dependence of the transition matrix reduces to |hΨf |e · p|Ψi i| = |ehΨf |p|Ψi i|. By applying the
dipole selection rule, the transition from 1s to a final state p orbital is the only allowed. For a 1 s initial
state and a directional final state orbital, the matrix element hΨf |e · p|Ψi i points to the final state orbita
(O). Then, the transition intensity is:

I ∝ |hΨf |e · p|Ψi i|2 ∝ |e · O|2 ∝ cos 2 δ, (2.9)

where δ is the angle between the electrical field e and the direction of the final state orbital O. Thus,
as exemplified in Fig 2.10, if the electric field is perpendicular to the molecular plane, (e.g., benzene) a
transition from C 1s to π ∗ and for parallel, a transition from the C 1s to σ ∗ can be observed.

2.3 Morphology characterization methods

Scanning transmission X-ray microscope (STXM) is a type of X-ray microscopy that uses a zone plate to
focus the radiation in a small region, then the sample is positioned in the focal plane of the zone plate, in
which it can be scanned and the transmitted X-ray intensity is saved as a function of the sample position.
CHAPTER 2. MATERIALS AND CHARACTERIZATION METHODS 27

Figure 2.10: Schematic diagram of polarization dependence of the X-ray absorption.

Figure 2.11: Standard design of scanning transmission X-ray microscope (STXM) optics with a dedicated
beamline for it. FZP: Fresnel zone plate. OSA: order sorting aperture. [66]

The local structure of the organic materials was studied using the compact STXM at PF, KEK [66].
The combination of STXM and NEXAFS is an effective way to evaluate the domain structure, because
the transitions from carbon 1s to the π ∗ and σ ∗ atomic orbitals can be seen by choosing the x-ray energy
near to C 1s ionization energy. The orientation of the molecule can be extracted using the following
equation:

OD = M P cos2 (β + θ) + (1 − P ) sin2 (β + θ) + C


(2.10)

where OD is the optical density and it is defined as -ln(I/I0 ). I and I0 are the transmitted intensity
and the background. respectively. By modifying the polarization from vertical (β =90°) to horizontal
(β =0°), the in-plane orientation θ, in relation to β, can be calculated for all points. To image the raw
CHAPTER 2. MATERIALS AND CHARACTERIZATION METHODS 28

data and proceed to further analysis, an algorithm using Python 3 was written.
Chapter 3

Electronic proprieties at donor-acceptor

interface

This chapter has been partly published in [67, 68] and it is divided into three parts. In the first part, to
have a broader view about the electronic properties at the organic-organic interface, it was studied the
SubPc/6T interface by growing the organic materials on different inorganic materials used as substrates:
MoO3 , SiO2 and Cs2 CO3 . Next, the same experiment was carried out, however, to investigate the
difference at the energy levels by controlling the morphology, the organic layers were annealed to produce
a more ordered film. And, in the last part, SubPc was replaced by Cl6 Subpc to evaluate the influence of
the permanent dipole in the energy-level alignment.
The donor-acceptor heterojunction was formed by depositing layer-by-layer of SubPc (Cl6 Subpc) on
6T/substrate. The compounds used as substrates were selected due to the wide range of the work
function values, from 3.0 to 6.7 eV, as displayed in Fig. 3.1. UV-ozone treatment was used to oxidize
the silicon wafer to obtain a thin layer of Si/SiO2 . The others substrate films were deposited by growing
approximately 5 nm of MoO3 and Cs2 CO3 on Si. Afterward, 5 nm of 6T and the acceptor layer were
evaporated step-by-step inside an ultra-high vacuum deposition chamber. In the analysis chamber, the
structures were annealed in-situ for 60 s at 150°C.

3.1 SubPc/6T/substrate

3.1.1 SubPc/6T/MoO3

In this section, the organic-organic interface of SubPc evaporated on 6T/substrate is discussed. The work
function of the substrate layer varies from 3.0 to 6.7 eV. The wide range of the work function value allows
us to obtain a broader view of the electrical proprieties at this interface.

As-grown

To get a complete overview of the interface properties between the organic molecules at the active layer,
the interface between the organic and inorganic layer (6T/MoO3) is firstly investigated. Fig. 3.2 displays
the UPS spectra for 6T evaporated in MoO3 as a function of 6T overlayer. The first spectrum refers to
5-nm-thick MoO3 film, in which the valence and gap states bond energy are measured as 2.71 and 0.28

29
CHAPTER 3. ELECTRONIC PROPRIETIES AT DONOR-ACCEPTOR INTERFACE 30

Figure 3.1: Optimized molecular structure obtained using Gaussian 9 for the SubPc and 6T is presented
on the left and on the right side is displayed the energy levels of the studied system before the contact.
The values for work function and ionization potential were measured using UPS. The bandgap energies
were obtained from the literature [69].

eV, respectively, in relation to the Fermi Level (EF ). Literature values for valence and gap states have
been reported as 2.74 and 0.36 eV [70], respectively, which agrees with our measurements. The HOMO
is found at the binding energy of 0.24 eV for the 0.4 nm-thick 6T layer. The onset of HOMO is pinned
around the EF until 2.5 nm-thick layer. When the final thickness of 5 nm is achieved, the onset of HOMO
is measured as 0.59 eV. Although the HOMO is pinned at 0.24 eV until the fourth deposition of 6T,
one can notice that the HOMO-1 and also the HOMO-2 levels have changed to higher binding energy,
suggesting that band bending occurs at the interface due to charge transfer.
XPS measurements are preferred to calculate the band bending at interfaces, since the UPS signal at
the very interface overlaps with the spectrum from the layer below, as discussed in Chapter 3. Figure
3.3 displays the XPS spectrum as a function of 6T overlayer, in which the S 2p peak can be observed.
After 0.4-nm-thick 6T is evaporated on MoO3 , the S 2p3/2 is found at 163.35 eV. Upon next evaporation,
the peak position moves by 0.45 eV to higher binding energy, which is consistent with HOMO+1 and
HOMO+2 shift observed in the UPS measurements. Once the 6T film is completed (5 nm), a total
displacement of 0.53 eV is observed. Since shift from the 1st to the 2nd evaporation is greater than the
shift from the 2nd to the 3rd one, we can consider,close to the interface, band bending can be observed .
In the next step, SubPc is evaporated on the 6T/MoO3 . Figure 3.4 displays the UPS spectra as a
function of SubPc thickness. As the increment of SubPc thickness, the HOMO onset shifts to higher
binding energy. The HOMO onset of 0.4 nm-thick SubPc is measured as 0.73 eV. After the film is
completed, the 5 nm-thick SubPc HOMO onset is localized at 0.86 eV. Since the SubPc spectrum covers
the 6T one for the 1.6 nm-thick SubPc, which suggests that the SubPc film completely covered the 6T
film. Also, 0.11 eV shift of HOMO-1 towards higher binding energy has been observed from the first to
the last thickness of SubPc, which suggests band bending.
As discussed in Chapter 2, to study the charge transfer process and band bending at the SubPc/6T
CHAPTER 3. ELECTRONIC PROPRIETIES AT DONOR-ACCEPTOR INTERFACE 31

Figure 3.2: UPS spectra displaying the valence region for MoO3 and 0.4, 0.8, 1.6, 2.5, and 5.0nm of 6T
deposited on MoO3 .

interface, the band bending on the underlayer can be separated from the upper layer using XPS mea-
surements. Figure 3.5 shows the S 2p (related to 6T) and Cl 2p (related to SubPc) core-level spectra for
SubPc evaporated on 6T/MoO3 . The S 2p3/2 peak is measured at 163.88 eV for the 5-nm thick 6T film.
When 0.4 nm-thick SubPc is evaporated, the S 2p3/2 peak shifts by 0.10 eV to higher binding energy and
when 5nm of SubPc is evaporated, a total peak shifted of 0.33 eV to higher binding energy is observed.
The Cl 2p3/2 peak is measured at 198.97 eV for 0.4 nm-thick, for the second thickness, a shifted to BE
higher of 0.12 eV was observed. And, in total, a displacement of 0.30 eV was noticed for 5 nm-thick
SubPc film.
The relative peaks shift of S 2p3/2 and Cl 2p3/2 as a function of SubPc increment are displayed in Fig.
3.6 (a). The band bending for 6T at the organic-organic heterojunction interface is directly extracted
from the shift of S 2p3/2 . On the other hand, the band bending of SubPc can be calculated using
the relation Cl 2p3/2 −S 2p3/2 (Fig. 3.6 (b)). When the interface is formed, due to the donor-acceptor
interface formation, charges are reorganized at the interface. Since SubPc has been reported with acceptor
property [71], charges transfer from 6T to SubPc. The positive charges accumulate at the underlayer and
the negative in the upper one. When the film gets thicker, the equilibrium can be obtained with lesser
charges resulting in the band bending. Figure 3.7 summarizes the result.
CHAPTER 3. ELECTRONIC PROPRIETIES AT DONOR-ACCEPTOR INTERFACE 32

Figure 3.3: Sulfur 2p peak for 6T deposited on MoO3 .

Annealed

The interface of the annealed film is examined to investigate the relationship between the electronic
properties and the morphology. Firstly, the 6T(5 nm)/MoO3 (5 nm) stack layer is grown on Si substrate.
The conditions are kept the same as the previous experiment however, this time, the stack layer is annealed
for 60 s 150°. Next, SubPc is grown on the annealed stack layer and after each evaporation of SubPc, the
structure is annealed. Figure 3.8 shows the UPS spectra for the annealed stack layers as a function as
SubPc thickness. The 6T HOMO onset shifts by 0.24 eV to higher binding energy, located at 0.79 eV.
Subsequent, SubPc is evaporated and annealed on 6T/MoO3 . The HOMO onset is located 0.73 eV, for
the first layer, and 1.20 eV, for the 5 nm-thick SubPc film. The evolution of HOMO position behaves
quite differently from the as-grown sample. At the final thickness, the annealed SubPc HOMO onset is
localized in higher binding energy of 0.34 eV compared to the former case.
Figure 3.9 displays the spectra for the sulfur 2p and chlorine 2p peaks of the annealed sample. Although
some fluctuation is observed, the peak shift is smaller than 0.10 eV for both analyzed peaks. The reduction
of peak shift means that the band bending has been reduced due to a decrease of the charges accumulated
at the interface.
To further explore the effect of annealing on SubPc, the comparison of the HOMO width for both
CHAPTER 3. ELECTRONIC PROPRIETIES AT DONOR-ACCEPTOR INTERFACE 33

Figure 3.4: UPS spectra displaying the valence region for 6T(5 nm) evaporated on MoO3 and for SubPc
evaporated on 6T/MoO3 .

Figure 3.5: (a) sulfur 2p and (b) chlorine 2p core level region for 5 nm of 6T grown on MoO3 and for
SubPc evaporated on 6T/MoO3 .
CHAPTER 3. ELECTRONIC PROPRIETIES AT DONOR-ACCEPTOR INTERFACE 34

Figure 3.6: (a) S 2p3/2 and Cl 2p3/2 peak shift as a function of SubPc deposition thickness for the MoO3
substrate. (b) Cl 2p3/2 - S 2p3/2 as a function of the thickness of SubPc layer.

films is displayed in Fig. 3.10(a). The HOMO levels were fitted using a Gaussian function:
E2
 
N0
g(E) = √ exp − 2 , (3.1)
σ 2π 2σ
where g(E) is the HOMO distribution, N0 is half of the number density of HOMO electron, E is the
√
binding energy, and FWHM is given by 2 2 ln 2σ. The FWHM is calculated as 1.11 eV and 0.86 eV for
the as-grown and annealed films, respectively. Although the background noise of the UPS measurement is
high to perform this type of analysis, the reduction of the width of DOS after annealing can be observed.
Owing to annealing, SubPc film became more ordered, consequently, the misalignment of the permanent
dipole inside the film is reduced that minimizes the local electric field, controlling the width of HOMO
DOS and tailing states.[72, 40] A complete discussion regarding the annealing effect in the morphology of
SubPc is done in the Chapter 4.1, in which addresses that, after annealing the film became more ordered
compared to the amorphous as-grown case. To observe the tail states the logarithm of the intensity is
preferable [37]. Thus, the HOMO intensity in logarithm scale as a function of the relative binding energy
HOMO peak is displayed in Fig.3.10 (b). Although the tailing is observed for the annealed and as-grown
films, the width of DOS of the annealed one has been demonstrated to be narrower.

The Voc loss

The open-circuit voltage (Voc ) is related to the photovoltaic gap by the Eq. 1.1 [73]. Thus, a wider
photovoltaic gap is essential to achieve higher Voc . Another way to improve it is by reducing energy loss.
For SubPc/6T device, it has been reported an energy loss of around 0.8 eV [69]. However, the origin of
this huge energy loss is still unclear.
Figure 3.11 displays the diagram of the interfacial energy-level alignment, which shows the difference
between annealed and as-grown structure. Due to the band bending, the photovoltaic gap is reduced,
CHAPTER 3. ELECTRONIC PROPRIETIES AT DONOR-ACCEPTOR INTERFACE 35

Figure 3.7: (a) Diagram of the energy-level alignment of the organic layers grown on MoO3 . (b) Charge
neutrality level diagram for SubPc evaporated on 6T/MoO3 . The SOO is 0.08 eV in this system.

that might cause a huge energy loss. After annealing, the charge transfer and band bending are reduced
that can improve the efficiency due to the enhancement of the photovoltaic gap.

3.1.2 SubPc/6T/Cs2 CO3

Here Cs2 CO3 is used as substrate. During the evaporation, the material disassociates in cesium which is
used as n-dopant in organic semiconductor field [74]. Before the evaporation of the organic layers, the
work function of Cs2 CO3 was measured as 3.0 eV. The experiment is carried out following specifications
discussed in the previous section.
CHAPTER 3. ELECTRONIC PROPRIETIES AT DONOR-ACCEPTOR INTERFACE 36

Figure 3.8: UPS spectra displaying the valence region for 6T (5 nm) evaporated on MoO3 and for SubPc
evaporated on 6T/MoO3 (Annealed).

Figure 3.9: (a) sulfur 2p and (b) chlorine 2p core level region for 5 nm of 6T grown on MoO3 and for
SubPc evaporated on 6T/MoO3 . (Annealed).

As-grown

Figure 3.12 displays the S 2p XPS spectra as a function of 6T increment thickness evaporated on Cs2 CO3 .
The S 2p3/2 peak has been displaced by 0.23 eV to lower binding energy from the 0.4 nm-thick 6T film
CHAPTER 3. ELECTRONIC PROPRIETIES AT DONOR-ACCEPTOR INTERFACE 37

Figure 3.10: HOMO region of annealed and as-grown 5 nm-thick SubPc grown 6T/MoO3 as a function
of the binding energy measured from the center of the HOMO. (a) As-grown and annealed SubPc film
HOMO displayed in linear intensity. The FWHM was calculated as 0.86 and 1.11 eV, respectively, by
fitting using a Gaussian function. (b) The same plot as in (a) using a logarithmic scale to enhance the
tailing states.

to the last thickness (5 nm). This means that the charge transfer occurs from 6T to the substrate to
obtain the equilibrium. One must note that the charge transfer and band bending are inverses compared
to the MoO3 case. It has been discussed in the literature the bipolar characteristic of 6T that enhances
the ability to donate and accept electrons [75].
Next, SubPc is evaporated on the 6T/Cs2 CO3 stack layer and the UPS spectra are obtained. Figure
CHAPTER 3. ELECTRONIC PROPRIETIES AT DONOR-ACCEPTOR INTERFACE 38

Figure 3.11: Comparison of the band bending in the organic-organic interface calculated using UPS and
XPS.

3.13 displays the UPS spectra for 6T film (5 nm) on Cs2 CO3 and 0.4 to 5 nm of SubPc grown on the
stack layer. We notice that the HOMO onset of 6T is located much further away from the Fermi level
(1.16 eV), which is in agreement with the n-doping caused by the Cs2 CO3 doping effect. Also, compared
to the MoO3 case, the HOMO onset of SubPc has shifted by 0.51 eV, located at 1.37eV. One can notice
that with the increment of SubPc thin film, the HOMO level shifts closer to the Fermi level, which can
be associated with band bending. However, due to the overlap of the UPS signal, it is difficult to confirm
it. Thus, we move to the investigation of the core levels.
Figure 3.14 (a) and (b) show the sulfur 2p and chlorine 2p peaks of SubPc deposited on 6T/Cs2 CO3 .
It can be seen in both images a shift to lower binding energy, which is inversed tp the direction observed
in the organic layers deposited on MoO3 . Figure 3.15 (a) and (b) display the relative shift of sulfur 2p
and chlorine 2p peaks shift and the difference of them for the SubPc evaporated on 6T/Cs2 CO3 . It is
interesting to notice that for a very thin layer of SubPc, a huge shift is observed due to the influence of
the substrate, however, for thicker layers, an upward band bending is observed due to the SubPc acceptor
property. The schematic diagram of the organic-organic interface is displayed in Fig. 3.16.
CHAPTER 3. ELECTRONIC PROPRIETIES AT DONOR-ACCEPTOR INTERFACE 39

Figure 3.12: Sulfur 2p peak for 6T deposited on Cs2 CO3 .

Annealed

Next, the experimental conditions were kept as before and the sample was annealed after each deposition
of SubPc. From the XPS spectra (Fig. 3.17), the peak shift observed at S 2p and Cl 2p was reduced, which
corroborates to our discussion about the reduction of tailing states due to the morphology control that
leads to a decrease in the available states to charges be transferred. Consequently, the thermodynamic
equilibrium can be achieved with lesser charge transfer than the as-grown case.

3.1.3 SubPc/6T/SiO2

The last substrate studied is the SiO2 one. Although SiO2 is an insulator, for a very thin layer of the
substrate grown on Si it can act as semiconductor [76]. The Si is oxidized (≤ 2 nm) using plasma cleaner
CHAPTER 3. ELECTRONIC PROPRIETIES AT DONOR-ACCEPTOR INTERFACE 40

Figure 3.13: UPS spectra displaying the valence region for 6T(5 nm) evaporated on Cs2 CO3 and for
SubPc deposited on 6T/Cs2 CO3 .

in ambient with O2 for 10 min. Next, 6T is evaporated on SiO2 /Si and the XPS spectra for S 2p is taken
(Fig. 3.18). In this case, the peak shift cannot be noticed, thus no detectable charge transfer is observed
to achieve the equilibrium.
Next, SubPc is evaporated on the stack layer and the UPS spectra are displayed in Fig. 3.19. It can
be noticed that the HOMO onset of 6T is located in an intermediate position (0.73 eV) compared to 6T
thin film grown on MoO3 and on Cs2 CO3 . After evaporation of 5 nm-thick SubPc thin-film on the stack
layer, the HOMO onset became located at 1.24 eV. It also can be observed that the HOMO and HOMO-1
peak do not appear to shift with the increment of SubPc thickness.
The XPS spectra are displayed in Fig. 3.20 (a) and (b), for the S 2p and Cl 2p, respectively, as a
function of SubPc overlayer thickness. With the increment of SubPc thickness, one can notice that neither
S 2p nor Cl 2p shifted. This fact corroborates the hypothesis that the tailing states enhance the charge
CHAPTER 3. ELECTRONIC PROPRIETIES AT DONOR-ACCEPTOR INTERFACE 41

Figure 3.14: XPS spectra of the S 2p peak and (b) Cl 2p peak for 5 nm of 6T deposited on Cs2 CO3 and
for SubPc evaporated on 6T/Cs2 CO3 .

Figure 3.15: ((a) S 2p3/2 and Cl 2p3/2 peak shift as a function of SubPc deposition thickness for the
Cs2 CO3 substrate. (b) Cl 2p3/2 - S 2p3/2 as a function of the thickness of SubPc layer.

transfer. Since the work function of the SiO2 /Si was measured as 4.1 eV, the Fermi level is between the
gap states of SubPc. Thus, the charge transfer is undetectable by the XPS measurement.
CHAPTER 3. ELECTRONIC PROPRIETIES AT DONOR-ACCEPTOR INTERFACE 42

Figure 3.16: Diagram of the energy-level alignment of the organic layers grown on Cs2 CO3 .

Figure 3.17: (a) sulfur 2p and (b) chlorine 2p core level region for 5 nm of 6T grown on Cs2 CO3 and for
SubPc evaporated on 6T/Cs2 CO3 (Annealed).
CHAPTER 3. ELECTRONIC PROPRIETIES AT DONOR-ACCEPTOR INTERFACE 43

Figure 3.18: Sulfur 2p peak for 6T deposited on SiO2 /Si.

3.2 SubPc tailing states control summary

Based on the peak shift summarized in Figure 3.21, it is possible to obtain the band bending on the
interface. The S 2p peak shift corresponds to a band bending of 6T film and Cl 2p - S 2p peak shift
corresponds to the band bending at SubPc. We could observe that due to the tail states on the SubPc film
after the interface is formed, charges are transferred to the tail states to achieve the equilibrium which
causes the charge accumulated at interface and band bending. The band bending at the donor-acceptor
interface can limit the Voc which could be the origin of the huge loss at the SubPc/6T device. Thus, if
the band bending could be suppressed, the device might achieve higher efficiency.
Here, it is discussed how the morphological modification, due to annealing, affects the electronic
properties at the organic-organic interface. Figure 3.22 summarizes the XPS result for annealed SubPc
CHAPTER 3. ELECTRONIC PROPRIETIES AT DONOR-ACCEPTOR INTERFACE 44

Figure 3.19: UPS spectra displaying the valence region for 6T (5 nm) evaporated on SiO2 /Si and for
SubPc deposited on 6T/SiO2 /Si.

on 6T/substrate.
Upon annealing, it was noticed a substantial control of the core-level peak displacement at the
SubPc/6T interfaces, which can be associated with the reduction of charge transfer and band bend-
ing at the organic-organic interface. This effect can be explained as a contraction of the DOGS of SubPc
by the better molecular stacking, which decreases the number of available states that the charges can
transfer to achieve the thermodynamic equilibrium.
Owing to annealing, SubPc film became more ordered, consequently, the misalignment of the per-
manent dipole inside the film is reduced that minimizes the local electric field, controlling the width of
HOMO DOS and tailing states. Due to the decrease of available tail states that charge can be transferred
into, the charge transfer is reduced, which reduces the charge accumulated at interface and band bending.
The fluctuation on the peak shift is caused by the non-uniformity of the annealed film. By annealing,
the film becomes more uniform, however, the process is not effective enough to align all molecules. The
CHAPTER 3. ELECTRONIC PROPRIETIES AT DONOR-ACCEPTOR INTERFACE 45

Figure 3.20: (a) sulfur 2p and (b) chlorine 2p core level region for 5 nm of 6T grown on SiO2 /Si and for
SubPc evaporated on 6T/SiO2 /Si.

Figure 3.21: a) S 2p3/2 and Cl 2p3/2 peak shift as a function of as-grown SubPc deposition thickness for
the three substrates. (b) Cl 2p3/2 - S 2p3/2 as a function of the thickness of SubPc layer.

misaligned molecules due to the non-complete ordering could affect the XPS peak location and the DOGS.
Since the organic layers in the OPV devices are very thin (< 100 nm), the annealed film non-uniformity
could affect more the efficiency than the improvement due to the reduction of DOGS. Thus, in the next
section, SubPc is replaced to Cl6 SubPc.
CHAPTER 3. ELECTRONIC PROPRIETIES AT DONOR-ACCEPTOR INTERFACE 46

Figure 3.22: S 2p3/2 and Cl 2p3/2 peak shift as a function of the thickness of annealed SubPc evaporated
on 6T/MoO3 and on 6T/Cs2 CO3 .

Figure 3.23: (a) S 2p3/2 and Cl 2p3/2 peak shift as a function of the thickness of Cl6 SubPc thickness on
MoO3 and Cs2 CO3 substrates. (b) Cl 2p3/2 - S 2p3/2 as a function of thickness of Cl6 SubPc film.

3.3 Cl6 SubPc/6T/substrate

Due to the presence of additional Cl atoms, the permanent dipole of the molecule is reduced from 4.59
D to 1.43 D. The reduction of the dipole improves the molecular stacking that reduces the local electric
CHAPTER 3. ELECTRONIC PROPRIETIES AT DONOR-ACCEPTOR INTERFACE 47

field and fluctuation of polarization energy that leads to a contraction of the DOS and reduction of tail
states. Figure 3.23 displays the S 2p3/2 and Cl 2p3/2 peak position as a function of the Cl6 SubPc film
thickness. The control of the peak shifts at the organic-organic interface grown on MoO3 can be observed.
We hypothesize that due to the lower permanent dipole, the molecules have a better stack which causes a
reduction of the width of the DOS, similar to the annealed SubPc film grown on 6T/MoO3 . Consequently,
the band bending is reduced. However, for Cs2 CO3 , due to the high electronegativity of Cl, the charge
transfer is enhanced.
To support our hypothesis about Cl6 SubPc and annealed SubPc film, we performed a high-sensitivity
spectroscopic measurement.

3.4 High-sensitivity spectroscopic measurement

To obtain a trustful result about the tail states, we use low background UPS to study the HOMO level
and low energy inverse photoemission spectroscopy (LEIPS) to obtain information about LUMO. The
LEIPS is more precise than the ordinary one because due to the improvement in the photon detector, it
can detect a small density of photons excited by electrons near-ultraviolet (5 eV) range [60]. IPS has been
used extensively in the literature to obtain information about the LUMO level [77, 78, 79, 27], however,
the film can be easily damaged due to the high electron energy.
Figure 3.24 displays the HOMO-LUMO region for all materials here studied. The HOMO was in-
vestigated using He I (21.2 eV) and the LUMO using LEIPS, thus the result is surface sensitive. The
energy gap for 6T (2.43 eV), as-grown SubPc (2.20 eV), annealed SubPc (2.42 eV) and Cl6 SubPc (2.38)
were obtained. It can be observed that the bandgap for the as-grown and annealed SubPc have increased
from 2.20 eV to 2.38 eV. The increment of the bandgap is owing to annealing that it may imply that this
process changed the crystallization of the SubPc thin film.
The UPS spectra taken using Xe I source (8.437 eV) are shown in Fig. 3.25 (a) and (b) for SubPc
and Cl6 SubPc on 6T/MoO3 . Due to the low energy photon, the penetration depth is high and the
measurement is bulk sensitive. SubPc and Cl6 SubPc were deposited on 6T (5 nm)/MoO3 . The HOMO
onset is measured as 0.24 eV for 5 nm-thick 6T film. After the evaporation of SubPc and Cl6 SubPc,
the HOMO onset of the acceptor layer are measured as 0.71 and 1.44 eV. The deeper level of Cl6 SubPc
has already been observed in the literature [80]. The broadening of the width of DOS was investigated
using high-sensitive UPS and the result is displayed in Fig.3.26. Due to the replacement of SubPc to
Cl6 SubPc, it was observed a reduction of the HOMO DOS full width at half maximum (FWHM). This
fact corroborates our assumption that the reduction of permanent dipole improves the molecular stalking
and controls the broadening of the width of DOS. Cl6 SubPc can be used as an acceptor in the OPV
devices and might improve the Voc due to the reduction of tail states.
Figure 3.27 displays the diagram of energy-level alignment for SubPc and Cl6 SubPc. Although the
photovoltaic gap is reduced from 1.73 eV to 1.18 eV by replacing SubPc to Cl6 SubPc, the photovoltaic
device may have better efficiency. Due to the reduction of tailing states, the band bending is reduced
which can improve the Voc .
CHAPTER 3. ELECTRONIC PROPRIETIES AT DONOR-ACCEPTOR INTERFACE 48

Figure 3.24: HOMO and LUMO spectra for 6T, SubPc, annealed SubPc and Cl6 SubPc obtained using
He I (21.2 eV) and LEIPS.
CHAPTER 3. ELECTRONIC PROPRIETIES AT DONOR-ACCEPTOR INTERFACE 49

Figure 3.25: High-sensitivy UPS spectra of valence level region for 6T(5 nm) evaporated on MoO3 and
for (a) SubPc and (b) Cl6 SubPc deposited on MoO3 .
CHAPTER 3. ELECTRONIC PROPRIETIES AT DONOR-ACCEPTOR INTERFACE 50

Figure 3.26: HOMO of 6T/MoO3 , as-grown SubPc, annealed SubPc and Cl6 SubPc on 6T/MoO3 . The
FWHM values were calculated by a Gaussian fit.
CHAPTER 3. ELECTRONIC PROPRIETIES AT DONOR-ACCEPTOR INTERFACE 51

Figure 3.27: Energy-level alignment obtained using high-sensitive UPS and LEIPS for SubPc and
Cl6 SubPc used as acceptor material.
Chapter 4

Study of local structure of organic

thin-films

Organic molecules, differently from the inorganic counterpart, are weakly bonded due to the van der Walls
interactions. Thus, organic thin-film used in devices are usually quite disordered, negatively impacting
the electrical and transport properties [32]. The energy-level alignment at the organic-organic interface is
susceptible to morphological modifications [40, 41]. It has been reported an increment of the interfacial
dipole at the organic-organic interface due to induction of defect using a non-reactive gas [59].
In Chapter 3, it is observed a reduction of the charge transfer after annealing SubPc layer. In our
hypothesis, the annealed film has a better molecular alignment that controls the gap states reducing the
charge transfer and band bending. To support our results, we carry out grazing incidence X-ray diffraction
(GIXRD) and atomic force microscopy (AFM). Although the XRD profile for 6T displays peaks associated
with molecules in a stand position (Figure 4.1), the profile for SubPc displays a glass amorphous peak.
As well, the AFM images, Figure 4.2, exhibit difference after annealing the SubPc film, however, it is not
clear enough to differentiate the morphological condition. In this case, a direct measurement is preferable
to clarify the modifications due to annealing.
In this Chapter, the morphology of the organic films is analyzed using scanning transmission x-ray
microscope (STXM). This technique has been used to analyze the orientation on many systems and it
has many applications for the study from porcelain to Kevlar [81, 82, 83, 84]. STXM combined with
NEXAFS has been successfully used to study the domain structure in polycrystalline films of organic
semiconductors.
Here, the morphology of SubPc and Cl6 SubPc were studied using STXM and image analysis. The
raw data was analyzed using an algorithm developed in Phyton 3 and the orientation was extracted using
Eq. 2.10. Part of the work in this chapter has already been published [68]. This chapter is divided into 3
sections, the first section addresses the comparison between the as-grown and annealed SubPc thin film.
In Section 4.2, the morphology of Cl6 SubPc is inspected. The section 4.3 summarizes the results found
here.

4.1 SubPc thin-film

First, two samples were prepared, one as-grown and the other one annealed. To do so, 100 nm-thick

52
CHAPTER 4. STUDY OF LOCAL STRUCTURE OF ORGANIC THIN-FILMS 53

Figure 4.1: XRD profile for 6T (100 nm)/MoO3 and for SubPc (100 nm)/6T/MoO3 .

SubPc thin-film was evaporated on silicon nitride membrane in a deposition chamber (< 5.0x10−9 Torr).
The growth rate was fixed at 0.02 nm/s for both films. The annealing was performed in a hot plate,
ex-situ, for 60 s at 150°C.
Before the measurement, the condition of the samples was verified using an optical microscope. Since
the silicon nitride substrate is very thin and small (1 mm2 ), it can be easily damaged, or some artifact
might be introduced. Figure 4.3 displays the optical microscope image for the as-grown and annealed
SubPc film. It can be observed that the as-grown film is uniform, however, some small spherulites can
be observed. These patterns are caused by the degradation of the film. Since the measurement was
not prepared on the same day as the measurement, degradation might occur. In the annealed film,
however, the spherulites are not observed, moreover, all the film became more orientated. To confirm this
assumption, we need to proceed to the STXM images.
4.1.1 As-grown SubPc thin film

We investigated the as-grown SubPc thin film using STXM, as displayed in Fig. 4.4 (a). We select three
regions in 4.4 (a) to extract the NEXAFS spectra for the in-plane horizontal (Fig. 4.4 (b)) and vertical
(Fig. 4.4 (c)) polarization. As observed by the resonance intensity for the two polarization, peaks at
284.8, 286.05, 287.9 and 289.4 eV, in the π ∗ region, can be addressed. The peaks are in agreement with
C 1s absorption spectrum for metal phthalocyanine (m-Pc) molecules [85].
The absorption spectra for horizontal and vertical polarization for the as-grown SubPc (Fig. 4.4 (b)
CHAPTER 4. STUDY OF LOCAL STRUCTURE OF ORGANIC THIN-FILMS 54

Figure 4.2: AFM images of (a) MoO3 (30 nm) and as-grown (b) 6T (15 nm) grown on MoO3 and (c)
SubPc (15 nm) grown on 6T/MoO3 ; AFM images of annealed (d) 6T (15 nm) grown on MoO3 and (e)
SubPc (15 nm) grown on 6T/MoO3 .

and (c)) are not affected by the position and/nor polarization, which means that the as-grown SubPc film
is amorphous.

4.1.2 Annealed SubPc thin-film

The STXM image for the annealed SubPc is displayed in Fig. 4.5 (a). In this case, color contrast can
be observed in the optical density (OD) image taken at 284.8 eV. Also, the spectra for horizontal (Fig.
4.5 (b)) and vertical (Fig. 4.5 (b)) resonance intensity display a complementary intensity for different
polarization, as well, the OD changes according to the position on the film. This means that the annealed
film is more ordered than the as-grown one.
The difference due to annealing is noticeable. The as-grown film is amorphous, however, the annealed
one presents two distinct orientations. To extract further information about the crystallization of SubPc,
additional analysis of the images is required.
CHAPTER 4. STUDY OF LOCAL STRUCTURE OF ORGANIC THIN-FILMS 55

Figure 4.3: Optical microscope image for as-grown SubPc and annealed SubPc film evaporated on silicon
nitride membrane.

The analysis was focused on the images obtained for the C 1s to π transition resonance peak (284.8
eV). We observed some cluster regions by subtracting the vertical mode from the horizontal images.
This can be associated with a non-perfect crystallization caused by non-sufficient annealing time and
temperature or thickness effect.
The orientation map can be extracted using Eq.2.10. To do so, we select the OD map taken at 284.8
eV. By dividing the vertically polarized image by the horizontally polarized one, the molecular angular
orientation as a function of the position can be obtained (Fig. 4.6). By extracting the line profile, two
plateaus can be observed. It also can be noticed that the plateaus are not completed ordered since the
line profile is not uniform.
To proceed, the histogram for Figure 4.6 is calculated and displayed in Fig. 4.7. The histogram has
two distinct peaks, thus the film became more ordered. However, due to the broadening of the peaks,
we can confirm that the film did not become completely crystalline, but more ordered than the as-grown
one.
The average molecular orientation in these two distinct regions can be extracted, which are calculated
as 36±1°and 64±1°for the lower and higher angular regions, respectively. The mapping for the molecular
orientation corroborates that annealing modifies the morphology of SubPc thin-film (amorphous) to a
more ordered film.

4.2 Cl6 SubPc thin-film

To continue the image evaluation of the acceptor layer, we replace SubPc to Cl6 SubPc. 50 nm-thick
Cl6 SubPc thin film was grown on silicon nitride membrane in a deposition chamber (< 5.0x10−9 Torr).
The growth rate was fixed as 0.02 nm/s. To avoid degradation, the sample was kept in vacuum condition
until the measurement.
First, the film condition is evaluated using an optical microscope, as displayed in Fig. 4.8. Since we
take care of the degradation, no spherulite structure is observed. And the film appears to be homogeneous
in all the sample which implies that the film could be amorphous.
The STXM image and the NEXAFS spectrum are displayed in Fig. 4.9. The absorption peaks localized
CHAPTER 4. STUDY OF LOCAL STRUCTURE OF ORGANIC THIN-FILMS 56

Figure 4.4: (a) Untreated image obtained at 284.8 eV for as-grown SubPc thin-film. NEXAFS spectra
for the delimited areas in (a) using (b) vertically and (c) horizontally polarized radiation.
CHAPTER 4. STUDY OF LOCAL STRUCTURE OF ORGANIC THIN-FILMS 57

Figure 4.5: (a) Untreated image obtained at 284.8 eV for the annealed SubPc thin-film. NEXAFS spectra
for the delimited areas in (a) using (b) vertically and (c) horizontally polarized radiation.
CHAPTER 4. STUDY OF LOCAL STRUCTURE OF ORGANIC THIN-FILMS 58

Figure 4.6: Orientation map for the annealed sample and the line profile in the orientation map.

at 285.1, 286.2, 287.2 and 288.1 eV correspond to transitions from carbon 1s to π ∗ . By comparing the
in-plane vertical and horizontal absorption, we confirm that the Cl6 SubPc film is amorphous.

4.3 Summary

In this chapter, we compared the as-grown and annealing SubPc. We could notice that the annealing
film presents a better stacking due to the highlighting of domains structures, however it is not perfectly
crystallized. This observation contributes to our assumption that the better molecular stacking after
annealing contributes to the reduction of the width of the density of states, as observed in Chapter 3.
On the other hand, Cl6 SubPc thin-film appears to be amorphous as SubPc thin-film. However, since the
permanent dipole is quite lower for the first, the intern electrical field is reduced that also control the
width of the density of states.
CHAPTER 4. STUDY OF LOCAL STRUCTURE OF ORGANIC THIN-FILMS 59

Figure 4.7: Histogram showing the distribution of angles. To obtain the histogram, it was applied a
median filter to reduce the noise.

Figure 4.8: Optical microscope image for Cl6 SubPc film evaporated on silicon nitride membrane.
CHAPTER 4. STUDY OF LOCAL STRUCTURE OF ORGANIC THIN-FILMS 60

Figure 4.9: Untreated STXM image taken at 284.9 eV and (b) NEXAFS spectrum for Cl6 SubPc film
using horizontal and vertical polarization.
Chapter 5

Influence of the molecular orientation on

the energy levels

In Chapter 3 we observe a reduction of the width of the density of states (DOS), in which we could
associate with better alignment of the molecules in the acceptor layer. In Chapter 4 the images of these
films are used to support our hypothesis. However, one point remains unclear: How does the permanent
dipole affect the energy levels? To clarify this point, here, the dependence of the orbital energies and the
permanent dipole is analyzed using density functional theory (DFT).
DFT is one of the most famous and popular methods in computation chemistry and physics and
solid state physics. DFT is a computational modeling method that can approximate the solution of the
Schrodinger equation for many-body systems, thus it is used in physics, chemistry and materials science
to study the electronic structure of molecules, and condensed phases. The main principle of this theory is
the use of functionals (a function of another function) of the electron density to determine the properties
of many-body systems. Since it uses the function of the electron density, the name density functional
theory was given to it.
Many physical properties of various metal phthalocyanine (m-Pc) molecules have been investigated
using DFT methods [86, 87, 88, 89, 90, 91, 92, 93]. The time-dependent DFT (TD-DFT) calculations
have been used to analyze the UV-vis spectra of them. The DFT model has been applied to study the
electronic structure orientation dependence of lead phthalocyanine in dimers [94]. The orbital energy of
SubPc considering different molecular orientation in dimers and trimers has already been calculated [95],
however they did not focus on the permanent dipole.
In this work, the molecular orbital energies for a single molecule and dimer of SubPc and Cl6 SubPc
were calculated using the hybrid functional b3lyp/6-31G(g,p). The calculation was done using Gaussian
9 and the molecular orbitals. The optimized geometry was drawn using Avogrado software.

5.1 SubPc orientation dependence

First, the orbital energy is calculated for SubPc in a gas phase. Figure 5.1 displays the HOMO and LUMO
orbital molecular and their respectively orbital energy levels for SubPc. We can see that the HOMO and
LUMO are located at -5.58 eV and -2.89 eV.

61
CHAPTER 5. INFLUENCE OF THE MOLECULAR ORIENTATION ON THE ENERGY LEVELS62

Figure 5.1: HOMO and LUMO calculated for the optimized SubPc.

LUMO+3 LUMO+2 LUMO+1 LUMO HOMO HOMO+1 HOMO+2 HOMO+3

−1.288 −1.293 −2.893 −2.893 −5.585 −7.220 −7.220 −7.365

Table 5.1: SubPc orbital energies calculated using Gaussian 9

Table 5.1 displays the orbital energies for SubPc from LUMO+3 to HOMO+3. Note that the LUMO
and LUMO+1 have the same energy, which means that they are degenerated. The energy difference
from HOMO+1 to HOMO is higher compared to HOMO+2 and HOMO+1. In normal porphyrinoids is
reported degeneracy of HOMO and HOMO+1, however, in the case of phthalocyanine, these levels became
non-degenerated due to the higher electronegativity of nitrogen atoms at the meso-positions compared to
the of carbon atoms around [96].
For the dimers calculations, we set three different positions for the SubPc molecules on the dimer. In
position 1 the molecules are facing each other, which reduce the permanent dipole to 0.038 D. The HOMO
is calculated as -5.60 eV and LUMO as -2.19 eV. On the other hand, by aligning the molecules, position 2
displays the higher permanent dipole measured as 9.51 D and the HOMO and LUMO are found at -5.48
eV and -2.99 eV, respectively. And in the last position, the molecules are not aligned to each other, thus
the permanent dipole is calculated as 8.91 D and the HOMO and LUMO are located at -5.54 eV and
-3.06 eV, respectively.
In the dimer configuration, the LUMO has became destabilized, although near degeneracy is still
observed. The ∆LUMO, which is the difference between LUMO+1 and LUMO, is calculated as 0.04952,
0.00054 and 0.01596 for position 1, 2 and 3, respectively. The total energy of the system can infer which
dimer is more stable. The energy for position 1 to 3 is calculated as −3470.817, −3470.838 and −3470.793
a.u., respectively. This suggests that position 3 is more stable than others. This result corroborates our
CHAPTER 5. INFLUENCE OF THE MOLECULAR ORIENTATION ON THE ENERGY LEVELS63

assumption that the annealed subPc has a better molecular stacking because from Fig. 5.2, lower the
permanent dipole, deeper is the orbital energy levels, which is in accordance with Chapter 3. Based on
the results of this section, we can support that the annealing originates a more ordered film with lower
local electrostatic potential that reduces the width of the density of states and tailing states. Which is
responsible for the controlling of the change transfer and band bending in the annealed film.

Figure 5.2: Orbital energy for SubPc dimers calculated for different relative molecular orientations.

5.2 Cl6 SubPc orientation dependence

In this section, the same analysis that was done to SubPc is repeated for Cl6 SubPc. Due to the presence
of the chlorine atoms replacing hydrogen atoms at the exterior benzene rings, the DFT calculation became
heavier and the calculation time longer.
The HOMO and LUMO orbital molecular and the orbital energy levels in the gas phase for Cl6 SubPc
are displayed in Figure 5.3. In this case, the energy position of HOMO and LUMO are located at -6.36
CHAPTER 5. INFLUENCE OF THE MOLECULAR ORIENTATION ON THE ENERGY LEVELS64

eV and -3.60 eV. We can observe that the Cl atoms contribute to the HOMO and LUMO, which is the
reason for the deep energy level observed in Chapter 3 and as well, reported in the literature [80].

Figure 5.3: HOMO and LUMO calculated for the optimized Cl6 SubPc.

Table 5.2 displays the orbital energies for Cl6 SubPc from LUMO+3 to HOMO+3. As observed for the
SubPc case, the LUMO and LUMO+1 are degenerated. Also, LUMO+2 and LUMO+3 have the same
energy. The HOMO is deeper than the one for SubPc as mention previously.

LUMO+3 LUMO+2 LUMO+1 LUMO HOMO HOMO+1 HOMO+2 HOMO+3

−1.953 −1.953 −3.604 −3.604 −6.357 −7.580 −7.581 −7.646

Table 5.2: Cl6 SubPc orbital energies calculated using Gaussian 9

For the Cl6 SubPc dimers calculations, we set three different positions as displayed in Fig.5.4. However,
in this case, position 1 refers to the lower dipole, position 2 has higher dipole. And position 3 has a low
total permanent dipole as position 1, but in this case, the Cl atoms are not aligned. In position 1, the
permanent dipole is calculated as 0.10 D and the HOMO (-6.57 eV) and LUMO (-3.88) are also calculated.
For position 2 the dipole increase to 3.41 D and the HOMO and LUMO are measured as -6.41 and -3.98
eV. The last one, position 3, the dipole is very similar to the first one (0.3 D), however, the HOMO (-5.87
eV) and LUMO (-4.02 eV) have changed from the former case, since the electrostatic repulsion is reduced.
Due to the chlorine atoms on the Cl6 SubPc, the coulomb interact between the molecules has increased
compared to SubPc. Owing to the Coulomb repulsion, the orbital energy has split for position 2. This
suggests that this configuration is unstable. By observing the total energy, for position 1 to 3, −8985.219,
−8986.152 and −8986.146 a.u., respectively, we can confirm that position 1 and 2 are more unstable. And
the more stable one is position 3, due to the reduction of the repulsion of the Cl atoms.
It already has been reported the fluorination of copper phthalocyanine can change the intrinsic prop-
erties to donate and accept electrons [97]. Jiang, et al. observed that with the increment of Fluorine
CHAPTER 5. INFLUENCE OF THE MOLECULAR ORIENTATION ON THE ENERGY LEVELS65

Figure 5.4: Orbital energy for Cl6 SubPc dimers calculated for different relative molecular orientations.

atoms, the electron injection is modified, which can change the transport properties from hole to electron
transport. Here, we could notice the same process, Cl6 SubPc has more acceptor properties than SubPc,
which could result in a better photovoltaic device.

5.3 Summary

By modifying the relative position of the molecules it was possible to calculate the orbital energy con-
sidering a total high dipole and a low dipole for both molecules. It was observed that variation in the
permanent dipole affects the orbital energies, which corroborates the UPS results. We also observed that,
for Cl6 Subpc, the position of the Cl, due to the high electronegativity, has a more noticeable effect on the
orbital energies than the permanent dipole.
To support our experimental results we performance DFT calculation for the molecules in gas phase
and in dimer position. We notice that for SubPc and Cl6 Subpc the most energetic favorable position to
stay is the ones with lower total permanent dipole. One assumption is that on the annealing the molecules
CHAPTER 5. INFLUENCE OF THE MOLECULAR ORIENTATION ON THE ENERGY LEVELS66

are more energetic to move, thus they can minimize the total energy (lower permanent dipole). Since the
internal electrical repulsion is reduced, the width of DOS is controlled.
Chapter 6

Study of local structure in crystalline

rubrene

6.1 introduction

In chapter 4 we investigate the morphology of SubPc and Cl6 Subpc thin-films. We started developing an
algorithm to analyze the STXM images, however, the analysis is limited by the disorder in the organic
films. To improve the analysis method, need to study a polycrystalline material with well-defined domains.
Thus, in this section polycrystalline rubrene is studied using scanning transmission and X-ray microscope
(STXM) and image analysis. Rubrene is a type organic semiconductor, with hole mobility measured
up to 20 cm2 V−1 s−1 in a single crystal [98, 99]. Although rubrene thin films are usually amorphous
when as-evaporated, the film can crystallize in three distinct phases: monoclinic, orthorhombic and
triclinic [100, 101]. The orthorhombic phase is the one of mainly interest because it has demonstrated
the highest charge carrier mobility [102]. To produce crystalline orthorhombic rubrene, many techniques
have been used in the literature [101, 103, 104]. The technique here used to obtain the orthorhombic film
has been reported by Fusella et al. [105]. By adjusting the annealing temperature of rubrene thermally
evaporated on a suitable underlayer, orthorhombic single-crystal domains with dimensions on the order
of 100 µm can be achieved. However, by annealing at higher temperatures, orthorhombic spherulite
domains can be formed. Although both films are orthorhombic, the rubrene with platelet domains has
hole mobility 4 times higher than the spherulite film [105].
The rubrene film was prepared by thermal evaporation on silicon nitride membrane by our collabo-
rators at Princeton University. 40 nm of rubrene was evaporated on 6 nm-thick film of TPTPA (tris(4-
(5-phenylthiophen-2-yl)phenyl)amine). TPTPA was used as an underlayer to assist the crystallization
as described previously [105]. Crystallization was achieved by annealing the samples on a preheated hot
plate in a nitrogen glovebox. The platelet samples were annealed at 140 ℃ for 5 min, and the spherulite
samples were annealed at 170 ℃ for 30 - 45 s. The STXM measurement was performed at KEK.

6.2 STXM image

Before the measurement using synchrotron radiation, the sample condition is examined using an optical
microscope. The optical microscope image is displayed in Fig. 6.1. By observing the image, there is

67
CHAPTER 6. STUDY OF LOCAL STRUCTURE IN CRYSTALLINE RUBRENE 68

almost no difference between the two films.

The near-edge X-ray absorption fine structure (NEXAFS) spectra of the platelet rubrene thin film
using in-plane polarized light are displayed in Figure 6.2. Figure 6.2 shows NEXAFS spectra for two
different grains of platelet film, which can be noticed that the resonance intensities for these two regions are
different. This difference means that they have different orientations. The peaks associated to transitions
from C 1s to the π ∗ molecular orbital at 284.2, 285.1 and 285.9 eV have already been assigned as α, β
and γ, respectively [106]. The resonance peak of α and γ peaks are sensitive to the backbone structure,
while the β intensity is sensitive to the wing structure of rubrene [107]. And the transitions from C 1s
to the σ ∗ molecular orbital are assigned as two broader peaks at 289 and 294 eV. Since the samples are
fixed normal to the incident radiation, we can obtain the information of the backbone of the molecule and
the wing project in-plane. In this experiment set-up, the out-of-plane information cannot be extracted
with STXM due to the incident radiation, however, X-ray and EBSD analysis showed that the rubrene
molecules are uniformly aligned out-of-plane [105].

Figure 6.1: Optical microscope images of the platelet and spherulite films. The regions selected to analyze
using STXM are delimited by the red rectangle.

The STXM images for the rubrene spherulite film taken at various photon energy are displayed in
Figure 6.3. The STXM images can be associated with the intensity of the NEXAFS spectra (Fig. 6.2).
The images taken at 284.2 (Fig. 6.3 (a) and (d)) and 285.9 eV (Fig. 6.3 (c) and (f)), as mentioned before,
the similarities in the color contrast can be clearly observed since they both are associated with the same
region (backbone). The uniformity of the color in the images suggests three grains domains. However,
in the image taken at 285.1 eV (Fig. 6.3 (b) and (e)), the image contrast cannot be observed, since the
resonance peak at 285.1 eV overlaps with the other two peaks as already reported in the literature [107].
One can also notice that this is the same reason that the images taken 285.9 eV are darker than the one
taken at 284.2 eV. To reduce the computational error and to get more precise information, we focus the
analysis on the α peak (284.2 eV). To conclude, Fig. 6.3(g) displays the pre-edge image, taken at 283.8
CHAPTER 6. STUDY OF LOCAL STRUCTURE IN CRYSTALLINE RUBRENE 69

Figure 6.2: NEXAFS spectra for two distinct regions of the rubrene platelet film using in-plane vertical
and horizontal polarization. The peaks at 284.2, 285.1 and 285.9 eV have been assigned in the literature
as α, β and γ, which correspond to C 1s to π ∗ molecular orbital transitions. The complementary pattern
observed in the three peaks for the same polarization, may suggest that the molecules in these regions
have different orientations.

eV. Since no pattern is observed here, our discussion is supported. The same is done for the one platelet
film, to comparison, and the result is shown in Fig. 6.4(a)-(f).
In the Fig. 6.5 (a) and (b), the optical density (OD) map for the platelet and spherulite samples are,
respectively, displayed. While two uniform grains can be identified in the platelet, the spherulite sample
exhibits micro-domains within a larger domain and a sinuous and unclear domain boundaries. Since the
annealing temperature was below TPTPA crystallization temperature, we can suppose that the difference
in the OD images come from the rubrene.
The in-plane molecular orientation map can be obtained from the Eq. 2.10 [108]. Figure 6.6 (a) and
(b) display the orientation map for the platelet and spherulite film, respectively, in which can be seen that
molecules in the platelet film are more homogeneous inside each domain in relation with the molecules in
the spherulite domain. Two distinct grains can be clearly distinguished in the platelet rubrene thin film
(Fig. 6.6(a)). Although the homogeneity in platelet film, defects in the film can be observed in the upper
corner of the grain boundary (Fig. 6.6(a)) inset, however, we do not have enough resolution to resolve
the defect. Figure 6.6(b) displays that the grain boundary in the spherulite film is less abrupt due to the
intradomain grain boundaries that are characteristic of the spherulite morphology. We define intradomain
grain boundary as the boundary between the rays (microdomains) created by the non-perfect alignment
of molecules in the crystal due to rapid crystallization, which displays some periodicity, however, we do
not have the necessary resolution to classify them. It is observed that the rays (microcrystalline domain)
are not exactly parallel one to another. Which causes the unclear boundary in the 1-2 region (Fig. 6.6(b)).
On the other hand, since the 1-3 is formed due to the encounter of rays originated in different nucleation
center, the grain boundary is much clearer defined.
Next, the distribution of molecular orientation is investigated by obtaining the histogram of the
orientation map for the platelet (Fig. 6.6 (c)) and spherulite (Fig. 6.6 (d)) films. Two distinct peaks are
observed in the histogram after applying a median filter to reduce the noise. However, for the spherulite
film, although the same filter was applied in the image, the distribution is much more spread, which means
CHAPTER 6. STUDY OF LOCAL STRUCTURE IN CRYSTALLINE RUBRENE 70

Figure 6.3: Images for a spherulite rubrene film taken atat 284.2, 285.1 and 285.9 eV, using (a-c) in-plane
horizontally and (d-f) vertically polarized radia. (g) The image using horizontally polarization taken at
283.8 eV is used as pre-edge (10 µm × 10 µm untreated STXM).

a more disordered film. By using a Gaussian function to fit the histogram, the full width at half maximum
(FWHM) could be extracted. The histogram for the spherulite film has been fitted by two Gaussians
function center located at 34.77° and 46.83°with the FWHM calculated, for both grains, as 1.12°. The
histogram for the spherulite film is fitted with three Gaussian centered at 40.45°, 42.88° and 45.26°. The
FWHM for the higher and lower angular distribution has been calculated as 2.38° and 2.03° , respectively.
The gain disordered can be quantitatively studied by the broadening of the angular distribution. As well,
the tailing observed in the border of the distribution, that cannot be fitted by using a Gaussian, indicates
that there is a concentration of discorded molecules.
The line profile for bolt films are displayed in Fig. 6.6 (e) and (f) (blue curves). While the platelet film
has an abrupt grain boundary and a smother molecular orientation, the spherulite film, on the other hand,
has a not well-defined boundary and the molecules are more disordered. Moreover, in the line profile for
the spherulite sample can be observed at the boundary, inside the same grain, a molecular mismatch.
The interaction between near molecules inside the crystal is low since the molecules are bonded via van
der Walls interactions, which is very weak. Thus the molecules from the adjacent positions can affect the
CHAPTER 6. STUDY OF LOCAL STRUCTURE IN CRYSTALLINE RUBRENE 71

orientation of them, causing this misaligned region. To compare, the average of the line profile along is
calculated and plotted as a function of position displayed in Fig. 6.6 (e) and (f) (green curves). One can
notice that the misalignment occurs throughout the gain boundary, which supports that the tails of the
histogram (Fig. 6.6(d)) come from the disorder at the grain boundary.

6.2.1 Grain boundary analysis

We can observe that the center of the grain boundary of the platelet sample, based on the line profile
and the average along the grain (Fig. 6.6(e)), stays around the same position compared to a much more
sinuous one for spherulite sample (Fig. 6.6(f)).To quantify and investigate the grain boundary sinuosity,
we use the root mean square (RMS) calculated for the center of the boundary. Firstly, the center of the
boundary is defined as x0 . So, we use a sigmoid function to fit the line profile:
b
f (x) = a + , (6.1)
1+ eλ(x−x0 )
where x is position. The center of the boundary x0 is calculated by optimizing the parameters a, b and λ
as displayed in Fig. 6.7. Next, the fitting is done all over the grain boundary. To reduce error due to noise
in the measurement, we fit the average of 3 pixels using the sigmoid function (Eq. 6.1). The mismatched
region near the grain boundary can be integrated to suggest numerically the molecular disorder and the
result is displayed in Fig. 6.7. After obtaining the result, the value for x0 can be plotted over the grain
boundary for both samples as displayed in Fig. 6.8 (a) and (b). The images were rotated by 48.80° and
33.26°in Fig. 6.8 (a) and (b), respectively, to align with the horizontal axis. Now, we can compute the
RMS for the platelet and spherulite border. The values are 0.04 and 0.14 µm for the former and latter,
respectively. The result suggests that the molecules are more misaligned and do not align at a straight
line. The RMS value for the spherulite boundary is higher due to the independent growth of the rays
from the nucleation center. In contrast, the platelet film does not have the microdomains allowing the
formation of a straight border.
Since the grain boundary for the platelet film is almost completely straight, we focus this analysis
on the spherulite sample. We investigate the region of 1 µm on either side of the grain boundary. The
histogram for the spherulite grain border region is displayed in Fig. 6.8(c), which looks exactly like the
full case (Fig. 6.6(d)). With two peaks that are centered at 39.41° and 45.72°, they have the FWHM
calculated as 3.80° and 2.20°, respectively. It can be noticed that the misaligned molecules concentrate
more in the grain boundary than in the interior of the grain. As well, the tailing observed in Fig. 6.8(c),
is more protuberant in the grain boundary region, which is connected to the misaligned molecules at the
very boundary.

6.3 Summary

The platelet and spherulite morphology can be achieved by thermal evaporation of rubrene on an appro-
priate organic underlayer and selecting the correct annealing temperature [105]. The underlayer improves
the crystallization of rubrene by assisting the molecules to rotate to find the lowest energy position and
orientation. However, annealing at high temperature may inhibit long-range order owing to more energy
and more rapid crystallization, generating the spherulite morphology. Although they have the same crys-
tal phase, the charge mobility of the platelet form has already been reported as 3-4 times higher than
CHAPTER 6. STUDY OF LOCAL STRUCTURE IN CRYSTALLINE RUBRENE 72

the spherulite one. As observed in the STXM image analysis, it is not only the presence of more grain
boundaries, but the grain boundary of the spherulite crystal has a rougher boundary and more misaligned
molecules which may also have point, perhaps dominant, in limiting the charge carrier mobility.
In this chapter, we consistently improved the STXM image analysis method used in Chapter 4 to study
the morphology of organic thin-films. Since we analyzed a polycrystalline film with well-defined domains,
we could refine the analysis method and extract parameters to quantify the morphological properties of
the film. As well, this method can be extended and applied to other materials.
CHAPTER 6. STUDY OF LOCAL STRUCTURE IN CRYSTALLINE RUBRENE 73

Figure 6.4: Images for a platelet rubrene film taken at 284.2, 285.1 and 285.9 eV, using (a-c) in-plane
horizontally and (d-f) vertically polarized radiation (10 µm × 10 µm untreated STXM).
CHAPTER 6. STUDY OF LOCAL STRUCTURE IN CRYSTALLINE RUBRENE 74

Figure 6.5: OD transmission image taken using vertically polarized light at 284.2 eV for the (a) platelet
and (b) spherulite polycrystalline film.
CHAPTER 6. STUDY OF LOCAL STRUCTURE IN CRYSTALLINE RUBRENE 75

Figure 6.6: Maps of the angular orientation for (a) platelet and (b) spherulite samples. The black dots
observed in the high contrast image of the grain boundary at the inset (a) are film defects in the film.
To easily identification, the spherulite grains in (b) are divided and numbered from 1 to 3. The angular
distribution for (c) platelet and (d) spherulite films are displayed in red solid and the Gaussian fit in
dashed line. The for the lines in (a) and (b) are shown in (e) and (f), respectively (blue). Average of the
angle as a function of position for the along the grain boundary (green).
CHAPTER 6. STUDY OF LOCAL STRUCTURE IN CRYSTALLINE RUBRENE 76

Figure 6.7: Line profile (blue solid line) for (a) platelet and (b) spherulite film. Step function used to fit
the line profile (yellow solid line). The mismatch of the values from the fitted curve in the region 1 µm
around the center of the grain boundary (in circle area) are shown.
CHAPTER 6. STUDY OF LOCAL STRUCTURE IN CRYSTALLINE RUBRENE 77

Figure 6.8: Enlarged region at grain boundary for (a) platelet and (b) spherulite rubrene film. The center
of grain boundary x0 was found using using equation 6.1 (black solid line). The blue dashed lines limits
the 1 µm around the center of the grain boundary for the spherulite film. (c) The histogram for the region
between the dashed line in (b) is painted in red. The blue dashed display the Gaussian fit, in which the
peaks were found at 39.41° and 45.72°, and FWHM is calculated as 3.80° and 2.20°, respectively.
Chapter 7

Conclusion

• Main results

This work focuses on the study of the relation between the morphology and the electrical properties in
organic semiconductors to improve organic photovoltaic devices. To do so, we first studied the electronic
properties using UPS and XPS and the morphology was mainly investigated by STXM. Our results also
were supported by DFT calculation.
It was observed that the loss in the Voc in SubPc/6T device may come from the band bending at
the interface. It was also shown that annealing process modifies the film structure, which can reduce
the charge accumulated and charge transferred at the organic-organic interfaces, due to the reduction of
the DOGS. By the analysis of STXM images, it was noticed that the annealing process can modify the
morphology of SubPc film, which corroborates our assumption that the control of defect reduction can
reduce the charge transfer. However, the annealed film presents some corrugation that may affect the
power conversion efficiency. In this scenario, the replacement for a molecule with lower permanent dipole
(Cl6 Subpc) may be more effective due to the reduction of tail states.
Using high-sensitive UPS, the annealed SubPc film and Cl6 SubPc film showed a narrower width of
the HOMO density of states, which suggest a reduction of defects and gap states that reduces the states
available to the charge be transferred. Thus, the thermodynamic equilibrium is achieved with fewer
charges to be transferred.
To achieve the main result, it has been developed an algorithm to analyze the STXM data. By
applying the algorithm to a crystallized film, which is rubrene, we could extract information about the
grain boundary as tracking the position of the boundary and the sinuosity of it. We could conclude that
the misaligned molecules near the grain boundary in the spherulite film may act as scattering centers to
the charges, which can explain the difference in the mobility between the platelet and spherulite films.

• Future tasks

One main future task is to study the annealed SuPc layer in the photovoltaic device to confirm the
hypothesis of the control of the gap states can improve the efficiency of the devices. Also, the same process
must be done using Cl6 SubPc. The conductivity of these films needs to be measured to clarify how the
morphology and the permanent dipole affect the transport properties.
Regard to the STXM data acquisition and analysis, due to the high intensity of the transmitted X-ray,
the possibility of damage and charging effect is very high that can add artifacts to the final result. The

78
CHAPTER 7. CONCLUSION 79

improvement of the algorithms to analyze the data could reduce the exposition time and increase the
details of the images.
Bibliography

[1] Lingxian Meng, Yamin Zhang, Xiangjian Wan, Chenxi Li, Xin Zhang, Yanbo Wang, Xin Ke, Zuo
Xiao, Liming Ding, Ruoxi Xia, et al. Organic and solution-processed tandem solar cells with 17.3%
efficiency. Science, 361(6407):1094–1098, 2018.

[2] Heliatek. Heliatek consolidates its technology leadership by establishing a new world record for
organic solar technology with a cell efficiency of 12%. 2013.

[3] https://www.eia.gov/tools/faqs/faq.php?id=97&t=3. Accessed: 2019-05-31.

[4] https://www.iea.org/statistics. Accessed: 2019-05-31.

[5] https://www.who.int/air-pollution/news-and-events/how-air-pollution-is-destroying-our-health.
Accessed: 2019-05-31.

[6] J Kurokawa, T Ohara, T Morikawa, S Hanayama, G Janssens-Maenhout, T Fukui, K Kawashima,

and H Akimoto. Emissions of air pollutants and greenhouse gases over asian regions during 2000–
2008: Regional emission inventory in asia (reas) version 2. Atmospheric Chemistry and Physics,
13(21):11019–11058, 2013.

[7] Emilio F Moran, Maria Claudia Lopez, Nathan Moore, Norbert Müller, and David W Hyndman.
Sustainable hydropower in the 21st century. Proceedings of the National Academy of Sciences,
115(47):11891–11898, 2018.

[8] WG Pfann and W Van Roosbroeck. Radioactive and photoelectric p-n junction power sources.
Journal of Applied Physics, 25(11):1422–1434, 1954.

[9] Keiichiro Masuko, Masato Shigematsu, Taiki Hashiguchi, Daisuke Fujishima, Motohide Kai, Naoki
Yoshimura, Tsutomu Yamaguchi, Yoshinari Ichihashi, Takahiro Mishima, Naoteru Matsubara, et al.
Achievement of more than 25% conversion efficiency with crystalline silicon heterojunction solar cell.
IEEE Journal of Photovoltaics, 4(6):1433–1435, 2014.

[10] https://www.nrel.gov/pv/cell-efficiency.html. Accessed: 2019-05-31.

[11] Jianhua Zhao, Aihua Wang, Martin A Green, and Francesca Ferrazza. 19.8% efficient “honeycomb”
textured multicrystalline and 24.4% monocrystalline silicon solar cells. Applied Physics Letters,
73(14):1991–1993, 1998.

[12] Sergio Pizzini. Solar grade silicon as a potential candidate material for low-cost terrestrial solar
cells. Solar energy materials, 6(3):253–297, 1982.

80
BIBLIOGRAPHY 81

[13] David E Carlson and Cristopher R Wronski. Amorphous silicon solar cell. Applied Physics Letters,
28(11):671–673, 1976.

[14] I Kaiser, K Ernst, Ch-H Fischer, R Könenkamp, C Rost, I Sieber, and M Ch Lux-Steiner. The
eta-solar cell with cuins2: A photovoltaic cell concept using an extremely thin absorber (eta). Solar
Energy Materials and Solar Cells, 67(1-4):89–96, 2001.

[15] Kannan Ramanathan, Miguel A Contreras, Craig L Perkins, Sally Asher, Falah S Hasoon, James
Keane, David Young, Manuel Romero, Wyatt Metzger, Rommel Noufi, et al. Properties of 19.2% ef-
ficiency zno/cds/cuingase2 thin-film solar cells. Progress in Photovoltaics: research and applications,
11(4):225–230, 2003.

[16] J Britt and C Ferekides. Thin-film cds/cdte solar cell with 15.8% efficiency. Applied physics letters,
62(22):2851–2852, 1993.

[17] Ching W Tang. Two-layer organic photovoltaic cell. Applied Physics Letters, 48(2):183–185, 1986.

[18] H Kallmann and M Pope. Photovoltaic effect in organic crystals. The Journal of Chemical Physics,
30(2):585–586, 1959.

[19] Bin Kan, Miaomiao Li, Qian Zhang, Feng Liu, Xiangjian Wan, Yunchuang Wang, Wang Ni, Guankui
Long, Xuan Yang, Huanran Feng, et al. A series of simple oligomer-like small molecules based on
oligothiophenes for solution-processed solar cells with high efficiency. Journal of the American
Chemical Society, 137(11):3886–3893, 2015.

[20] Miaomiao Li, Ke Gao, Xiangjian Wan, Qian Zhang, Bin Kan, Ruoxi Xia, Feng Liu, Xuan Yang,
Huanran Feng, Wang Ni, et al. Solution-processed organic tandem solar cells with power conversion
efficiencies¿ 12%. Nature Photonics, 11(2):85, 2017.

[21] Margaret A Stevens and Alexi C Arango. Open-circuit voltage exceeding the outermost homo-lumo
offset in cascade organic solar cells. Organic Electronics, 37:80–84, 2016.

[22] Barry P Rand, Diana P Burk, and Stephen R Forrest. Offset energies at organic semiconductor
heterojunctions and their influence on the open-circuit voltage of thin-film solar cells. Physical
Review B, 75(11):115327, 2007.

[23] Haijun Bin, Liang Gao, Zhi-Guo Zhang, Yankang Yang, Yindong Zhang, Chunfeng Zhang, Shanshan
Chen, Lingwei Xue, Changduk Yang, Min Xiao, et al. 11.4% efficiency non-fullerene polymer solar
cells with trialkylsilyl substituted 2d-conjugated polymer as donor. Nature communications, 7:13651,
2016.

[24] Haijun Bin, Yankang Yang, Zhi-Guo Zhang, Long Ye, Masoud Ghasemi, Shanshan Chen, Yindong
Zhang, Chunfeng Zhang, Chenkai Sun, Lingwei Xue, et al. 9.73% efficiency nonfullerene all or-
ganic small molecule solar cells with absorption-complementary donor and acceptor. Journal of the
American Chemical Society, 139(14):5085–5094, 2017.

[25] Qing-Ya Li, Jingyang Xiao, Lu-Ming Tang, Hua-Chun Wang, Ziming Chen, Zhiyong Yang, Hin-Lap
Yip, and Yun-Xiang Xu. Thermally stable high performance non-fullerene polymer solar cells with
BIBLIOGRAPHY 82

low energy loss by using ladder-type small molecule acceptors. Organic Electronics, 44:217–224,
2017.

[26] Guoxiao JIA, Shaoqing ZHANG, Liyan YANG, Chang HE, Huili FAN, and Jianhui HOU. Develop-
ment of benzodithiophene-based ada small molecules with different acceptor end groups for efficient
organic solar cells. Acta Physico-Chimica Sinica, 35(1):76–83, 2017.

[27] Kjell Cnops, Barry P Rand, David Cheyns, Bregt Verreet, Max A Empl, and Paul Heremans.
8.4% efficient fullerene-free organic solar cells exploiting long-range exciton energy transfer. Nature
Communications, 5:3406, 2014.

[28] Takeaki Sakurai, Susumu Toyoshima, Hikaru Kitazume, Shigeru Masuda, Hiroo Kato, and Kat-
suhiro Akimoto. Influence of gap states on electrical properties at interface between bathocuproine
and various types of metals. Journal of Applied Physics, 107(4):043707, 2010.

[29] Martin Oehzelt, Kouki Akaike, Norbert Koch, and Georg Heimel. Energy-level alignment at organic
heterointerfaces. Science Advances, 1(10):e1501127, 2015.

[30] Slawomir Braun, William R Salaneck, and Mats Fahlman. Energy-level alignment at organic/metal
and organic/organic interfaces. Advanced materials, 21(14-15):1450–1472, 2009.

[31] Martin Oehzelt, Norbert Koch, and Georg Heimel. Organic semiconductor density of states controls
the energy level alignment at electrode interfaces. Nature communications, 5:4174, 2014.

[32] Jin-Peng Yang, Fabio Bussolotti, Satoshi Kera, and Nobuo Ueno. Origin and role of gap states
in organic semiconductor studied by UPS: as the nature of organic molecular crystals. Journal of
Physics D: Applied Physics, 50(42):423002, 2017.

[33] Slawomir Braun, William R Salaneck, and Mats Fahlman. Energy-level alignment at organic/metal
and organic/organic interfaces. Advanced materials, 21(14-15):1450–1472, 2009.

[34] H Vazquez, F Flores, and Antoine Kahn. Induced density of states model for weakly-interacting
organic semiconductor interfaces. Organic Electronics, 8(2-3):241–248, 2007.

[35] Georges Hadziioannou and Paul Van Hutten. Semiconducting polymers. Wiley Online Library, 2000.

[36] Tomoki Sueyoshi, Haruya Kakuta, Masaki Ono, Kazuyuki Sakamoto, Satoshi Kera, and Nobuo
Ueno. Band gap states of copper phthalocyanine thin films induced by nitrogen exposure. Applied
Physics Letters, 96(9):39, 2010.

[37] Tomoki Sueyoshi, Hirohiko Fukagawa, Masaki Ono, Satoshi Kera, and Nobuo Ueno. Low-density
band-gap states in pentacene thin films probed with ultrahigh-sensitivity ultraviolet photoelectron
spectroscopy. Applied Physics Letters, 95(18):290, 2009.

[38] Jin-Peng Yang, Lin-Tai Shang, Fabio Bussolotti, Li-Wen Cheng, Wen-Qing Wang, Xiang-Hua Zeng,
Satoshi Kera, Yan-Qing Li, Jian-Xin Tang, and Nobuo Ueno. Fermi-level pinning appears upon
weak electrode-organic contact without gap states: a universal phenomenon. Organic Electronics,
48:172–178, 2017.
BIBLIOGRAPHY 83

[39] Jin-Peng Yang, Fabio Bussolotti, Satoshi Kera, and Nobuo Ueno. Origin and role of gap states
in organic semiconductor studied by UPS: as the nature of organic molecular crystals. Journal of
Physics D: Applied Physics, 50(42):423002, 2017.

[40] T Hosokai, H Machida, A Gerlach, S Kera, F Schreiber, and N Ueno. Impact of structural imper-
fections on the energy-level alignment in organic films. Physical Review B, 83(19):195310, 2011.

[41] Keiichirou Yonezawa, Alexander Hinderhofer, Takuya Hosokai, Kengo Kato, Rintaro Makino, Frank
Schreiber, Nobuo Ueno, and Satoshi Kera. Structural defects control the energy level alignment at
organic/organic interfaces. Advanced Materials Interfaces, 1(5):1400004, 2014.

[42] Do Young Kim, Jegadesan Subbiah, Galileo Sarasqueta, Franky So, Huanjun Ding, Irfan, and
Yongli Gao. The effect of molybdenum oxide interlayer on organic photovoltaic cells. Applied
Physics Letters, 95(9):224, 2009.

[43] Junwei Ma, XY Jiang, Zhang Liang, Jin Cao, Xiaowen Zhang, and ZL Zhang. Highly power efficient
organic light-emitting diodes based on cs2co3 n-doped and moo3 p-doped carrier transport layers.
Semiconductor science and technology, 24(3):035009, 2009.

[44] Denis Fichou. Structural order in conjugated oligothiophenes and its implications on opto-electronic
devices. Journal of Materials Chemistry, 10(3):571–588, 2000.

[45] Bernhard Klett, Caterina Cocchi, Linus Pithan, Stefan Kowarik, and Claudia Draxl. Polymorphism
in α-sexithiophene crystals: relative stability and transition path. Physical Chemistry Chemical
Physics, 18(21):14603–14609, 2016.

[46] I Hancox, P Sullivan, KV Chauhan, N Beaumont, LA Rochford, RA Hatton, and TS Jones. The
effect of a moox hole-extracting layer on the performance of organic photovoltaic cells based on
small molecule planar heterojunctions. Organic Electronics, 11(12):2019–2025, 2010.

[47] Yunlong Zou and Russell J Holmes. Influence of a moox interlayer on the open-circuit voltage in
organic photovoltaic cells. Applied Physics Letters, 103(5):138 1, 2013.

[48] Nicola Beaumont, Sang Wan Cho, Paul Sullivan, David Newby, Kevin E Smith, and Tim S Jones.
Boron subphthalocyanine chloride as an electron acceptor for high-voltage fullerene-free organic
photovoltaics. Advanced functional materials, 22(3):561–566, 2012.

[49] Chunhui Duan, Germán Zango, Miguel Garcı́a Iglesias, Fallon JM Colberts, Martijn M Wienk,
M Victoria Martı́nez-Dı́az, René AJ Janssen, and Tomás Torres. The role of the axial substituent in
subphthalocyanine acceptors for bulk-heterojunction solar cells. Angewandte Chemie International
Edition, 56(1):148–152, 2017.

[50] Bernd Ebenhoch, Nor BA Prasetya, Vincent M Rotello, Graeme Cooke, and Ifor DW Samuel.
Solution-processed boron subphthalocyanine derivatives as acceptors for organic bulk-heterojunction
solar cells. Journal of Materials Chemistry A, 3(14):7345–7352, 2015.
BIBLIOGRAPHY 84

[51] Hans Gommans, Tom Aernouts, Bregt Verreet, Paul Heremans, Anaı̈s Medina, Christian G
Claessens, and Tomas Torres. Perfluorinated subphthalocyanine as a new acceptor material in
a small-molecule bilayer organic solar cell. Advanced functional materials, 19(21):3435–3439, 2009.

[52] Paul Sullivan, Amelie Duraud, lan Hancox, Nicola Beaumont, Giorgio Mirri, James HR Tucker,
Ross A Hatton, Michael Shipman, and Tim S Jones. Halogenated boron subphthalocyanines as
light harvesting electron acceptors in organic photovoltaics. Advanced Energy Materials, 1(3):352–
355, 2011.

[53] Vishal Shrotriya, Gang Li, Yan Yao, Chih-Wei Chu, and Yang Yang. Transition metal oxides as
the buffer layer for polymer photovoltaic cells. Applied Physics Letters, 88(7):073508, 2006.

[54] Yuzheng Guo and John Robertson. Origin of the high work function and high conductivity of moo3.
Applied Physics Letters, 105(22):222110, 2014.

[55] Jens Meyer, Rebecca Khalandovsky, Patrick Görrn, and Antoine Kahn. Moo3 films spin-coated
from a nanoparticle suspension for efficient hole-injection in organic electronics. Advanced Materials,
23(1):70–73, 2011.

[56] Jinsong Huang, Zheng Xu, and Yang Yang. Low-work-function surface formed by solution-processed
and thermally deposited nanoscale layers of cesium carbonate. Advanced Functional Materials,
17(12):1966–1973, 2007.

[57] G Li, C-W Chu, V Shrotriya, Jinsong Huang, and Y Yang. Efficient inverted polymer solar cells.
Applied Physics Letters, 88(25):253503, 2006.

[58] Rhonda F Salzman, Jiangeng Xue, Barry P Rand, Alex Alexander, Mark E Thompson, and
Stephen R Forrest. The effects of copper phthalocyanine purity on organic solar cell performance.
Organic Electronics, 6(5-6):242–246, 2005.

[59] Fabio Bussolotti, Jinpeng Yang, Alexander Hinderhofer, Yuli Huang, Wei Chen, Satoshi Kera, An-
drew TS Wee, and Nobuo Ueno. Origin of the energy level alignment at organic/organic interfaces:
The role of structural defects. Physical Review B, 89(11):115319, 2014.

[60] Hiroyuki Yoshida. Near-ultraviolet inverse photoemission spectroscopy using ultra-low energy elec-
trons. Chemical Physics Letters, 539:180–185, 2012.

[61] WR Salaneck. Intermolecular relaxation energies in anthracene. Physical Review Letters, 40(1):60,
1978.

[62] MB Casu, Y Zou, S Kera, D Batchelor, Th Schmidt, and E Umbach. Investigation of polarization
effects in organic thin films by surface core-level shifts. Physical Review B, 76(19):193311, 2007.

[63] M Pl Seah and WA Dench. Quantitative electron spectroscopy of surfaces: A standard data base
for electron inelastic mean free paths in solids. Surface and interface analysis, 1(1):2–11, 1979.

[64] R Schlaf, BA Parkinson, PA Lee, KW Nebesny, and Neal R Armstrong. Homo/lumo alignment at
ptcda/znpc and ptcda/clinpc heterointerfaces determined by combined ups and xps measurements.
The Journal of Physical Chemistry B, 103(15):2984–2992, 1999.
BIBLIOGRAPHY 85

[65] Joachim Stöhr. NEXAFS spectroscopy, volume 25. Springer Science & Business Media, 2013.

[66] Y Takeichi, N Inami, H Suga, C Miyamoto, T Ueno, K Mase, Y Takahashi, and K Ono. Design and
performance of a compact scanning transmission x-ray microscope at the photon factory. Review of
Scientific Instruments, 87(1):013704, 2016.

[67] Alexandre L. Foggiatto and Takeaki Sakurai. Charge transfer induced by MoO3 at boron sub-
phthalocyanine chloride/α-sexithiophene heterojunction interface. Japanese Journal of Applied
Physics, 57(3S2):03EE01, jan 2018.

[68] Alexandre L. Foggiatto, Hiroki Suga, Yasuo Takeichi, Kanta Ono, Yoshio Takahashi, Kentaro Kut-
sukake, Takahiro Ueba, Satoshi Kera, and Takeaki Sakurai. Dependence of substrate work function
on the energy-level alignment at organic–organic heterojunction interface. Japanese Journal of
Applied Physics, 58(SB):SBBG06, feb 2019.

[69] James Endres, István Pelczer, Barry P Rand, and Antoine Kahn. Determination of energy level
alignment within an energy cascade organic solar cell. Chemistry of Materials, 28(3):794–801, 2016.

[70] Kaname Kanai, Kenji Koizumi, Satoru Ouchi, Yoshiaki Tsukamoto, Kei Sakanoue, Yukio Ouchi,
and Kazuhiko Seki. Electronic structure of anode interface with molybdenum oxide buffer layer.
Organic Electronics, 11(2):188–194, 2010.

[71] Kjell Cnops, German Zango, Jan Genoe, Paul Heremans, M Victoria Martinez-Diaz, Tomas Torres,
and David Cheyns. Energy level tuning of non-fullerene acceptors in organic solar cells. Journal of
the American Chemical Society, 137(28):8991–8997, 2015.

[72] A Dieckmann, H Bässler, and PM Borsenberger. An assessment of the role of dipoles on the density-
of-states function of disordered molecular solids. The Journal of chemical physics, 99(10):8136–8141,
1993.

[73] Barry P Rand, Diana P Burk, and Stephen R Forrest. Offset energies at organic semiconductor
heterojunctions and their influence on the open-circuit voltage of thin-film solar cells. Physical
Review B, 75(11):115327, 2007.

[74] Y Cai, HX Wei, J Li, QY Bao, X Zhao, ST Lee, YQ Li, and JX Tang. Mechanism of cs 2 co 3 as
an n-type dopant in organic electron-transport film. Applied Physics Letters, 98(11):60, 2011.

[75] Tomo Sakanoue, Masayuki Yahiro, Chihaya Adachi, Kazuo Takimiya, and Akio Toshimitsu. Elec-
trical characteristics of single-component ambipolar organic field-effect transistors and effects of air
exposure on them. Journal of Applied Physics, 103(9):094509, 2008.

[76] Taketoshi Matsumoto, Hiroki Nakajima, Daichi Irishika, Takaaki Nonaka, Kentaro Imamura, and
Hikaru Kobayashi. Ultrathin sio2 layer formed by the nitric acid oxidation of si (naos) method to
improve the thermal-sio2/si interface for crystalline si solar cells. Applied Surface Science, 395:56–60,
2017.

[77] IG Hill, Antoine Kahn, ZG Soos, and RA Pascal Jr. Charge-separation energy in films of π-
conjugated organic molecules. Chemical Physics Letters, 327(3-4):181–188, 2000.
BIBLIOGRAPHY 86

[78] Weiying Gao and Antoine Kahn. Controlled p-doping of zinc phthalocyanine by coevaporation with
tetrafluorotetracyanoquinodimethane: A direct and inverse photoemission study. Applied Physics
Letters, 79(24):4040–4042, 2001.

[79] Peter I Djurovich, Elizabeth I Mayo, Stephen R Forrest, and Mark E Thompson. Measurement
of the lowest unoccupied molecular orbital energies of molecular organic semiconductors. Organic
Electronics, 10(3):515–520, 2009.

[80] Hyunbok Lee, Sun Woo Ahn, Sim Hee Ryu, Bo Kyung Ryu, Myeung Hee Lee, Sang Wan Cho,
Kevin E Smith, and Tim S Jones. Interfacial electronic structure of cl 6 subpc non-fullerene acceptors
in organic photovoltaics using soft x-ray spectroscopies. Physical Chemistry Chemical Physics,
19(47):31628–31633, 2017.

[81] AP Smith and H Ade. Quantitative orientational analysis of a polymeric material (Kevlar® fibers)
with x-ray microspectroscopy. Applied Physics Letters, 69(25):3833–3835, 1996.

[82] Ebrahim Najafi, Daniel Hernández Cruz, Martin Obst, Adam P Hitchcock, Bastien Douhard, Jean-
Jacques Pireaux, and Alexandre Felten. Polarization dependence of the c 1s x-ray absorption spectra
of individual multi-walled carbon nanotubes. small, 4(12):2279–2285, 2008.

[83] Björn Bräuer, Ajay Virkar, Stefan CB Mannsfeld, David P Bernstein, Roopali Kukreja, Kang Wei
Chou, Tolek Tyliszczak, Zhenan Bao, and Yves Acremann. X-ray microscopy imaging of the grain
orientation in a pentacene field-effect transistor. Chemistry of Materials, 22(12):3693–3697, 2010.

[84] Benjamin Watts, Torben Schuettfort, and Christopher R McNeill. Mapping of domain orientation
and molecular order in polycrystalline semiconducting polymer films with soft x-ray microscopy.
Advanced Functional Materials, 21(6):1122–1131, 2011.

[85] Marco Vittorio Nardi, Francesca Detto, Lucrezia Aversa, Roberto Verucchi, Giancarlo Salviati,
Salvatore Iannotta, and Maurizio Casarin. Electronic properties of cupc and h 2 pc: an experimental
and theoretical study. Physical Chemistry Chemical Physics, 15(31):12864–12881, 2013.

[86] Ahmad Irfan, Abdullah G Al-Sehemi, Abdullah M Asiri, Muhammad Nadeem, and Khalid A
Alamry. A study on the electronic and charge transfer properties in tin phthalocyanine (snpc)
derivatives by density functional theory. Computational and Theoretical Chemistry, 977(1-3):9–12,
2011.

[87] Yuexing Zhang, Xianxi Zhang, Zhongqiang Liu, Yongzhong Bian, and Jianzhuang Jiang. Structures
and properties of 1, 8, 15, 22-tetrasubstituted phthalocyaninato-lead complexes: the substitutional
effect study based on density functional theory calculations. The Journal of Physical Chemistry A,
109(28):6363–6370, 2005.

[88] N Papageorgiou, Y Ferro, E Salomon, A Allouche, JM Layet, L Giovanelli, and G Le Lay. Geometry
and electronic structure of lead phthalocyanine: Quantum calculations via density-functional theory
and photoemission measurements. Physical Review B, 68(23):235105, 2003.
BIBLIOGRAPHY 87

[89] Michinori Sumimoto, Teruyuki Honda, Yukio Kawashima, Kenji Hori, and Hitoshi Fujimoto. The-
oretical and experimental investigation on the electronic properties of the shuttlecock shaped and
the double-decker structured metal phthalocyanines, mpc and m (pc) 2 (m= sn and pb). Dalton
Transactions, 41(23):7141–7150, 2012.

[90] Jakub D Baran and J Andreas Larsson. Inversion of the shuttlecock shaped metal phthalocya-
nines mpc (m= ge, sn, pb)—a density functional study. Physical Chemistry Chemical Physics,
12(23):6179–6186, 2010.

[91] Anton S Nizovtsev and Svetlana G Kozlova. Electronic rearrangements during the inversion of lead
phthalocyanine. The Journal of Physical Chemistry A, 117(2):481–488, 2013.

[92] Yuexing Zhang, Xue Cai, Xianxi Zhang, Hui Xu, Zhongqiang Liu, and Jianzhuang Jiang. Time-
dependent density functional theory studies of the electronic absorption spectra of metallophthalo-
cyanines of group iva. International Journal of Quantum Chemistry, 107(4):952–961, 2007.

[93] Yuexing Zhang, Xianxi Zhang, Zhongqiang Liu, Hui Xu, and Jianzhuang Jiang. Comparative
density functional theory study of the structures and properties of metallophthalocyanines of group
iv b. Vibrational spectroscopy, 40(2):289–298, 2006.

[94] Nobutsugu Hamamoto, Hiromitsu Sonoda, Michinori Sumimoto, Kenji Hori, and Hitoshi Fujimoto.
Theoretical study on crystal polymorphism and electronic structure of lead (ii) phthalocyanine using
model dimers. RSC Advances, 7(14):8646–8653, 2017.

[95] Xue Chen and Shaohui Zheng. Inferring the molecular arrangements of boron subphthalocyanine
chloride in thin film from a dft/tddft study of molecular clusters and experimental electronic ab-
sorption spectra. Organic Electronics, 62:667–675, 2018.

[96] Soji Shimizu and Nagao Kobayashi. Structurally-modified subphthalocyanines: Molecular design
towards realization of expected properties from the electronic structure and structural features of
subphthalocyanine. Chemical Communications, 50(53):6949–6966, 2014.

[97] Hui Jiang, Peng Hu, Jun Ye, Yongxin Li, Henan Li, Xiaotao Zhang, Rongjin Li, Huanli Dong,
Wenping Hu, and Christian Kloc. Molecular crystal engineering: Tuning organic semiconductor from
p-type to n-type by adjusting their substitutional symmetry. Advanced Materials, 29(10):1605053,
2017.

[98] V Podzorov, E Menard, A Borissov, V Kiryukhin, JA Rogers, and ME Gershenson. Intrinsic charge
transport on the surface of organic semiconductors. Physical Review Letters, 93(8):086602, 2004.

[99] Y Okada, K Sakai, T Uemura, Y Nakazawa, and J Takeya. Charge transport and hall effect in
rubrene single-crystal transistors under high pressure. Physical Review B, 84(24):245308, 2011.

[100] Oana D Jurchescu, Auke Meetsma, and Thomas TM Palstra. Low-temperature structure of rubrene
single crystals grown by vapor transport. Acta Crystallographica Section B: Structural Science,
62(2):330–334, 2006.
BIBLIOGRAPHY 88

[101] Liwei Huang, Qing Liao, Qiang Shi, Hongbing Fu, Jinshi Ma, and Jiannian Yao. Rubrene micro-
crystals from solution routes: their crystallography, morphology and optical properties. Journal of
Materials Chemistry, 20(1):159–166, 2010.

[102] Xiaopeng Wang, Taylor Garcia, Stephen Monaco, Bohdan Schatschneider, and Noa Marom. Effect
of crystal packing on the excitonic properties of rubrene polymorphs. CrystEngComm, 18(38):7353–
7362, 2016.

[103] Takeshi Matsukawa, Masashi Yoshimura, Kenichi Sasai, Masahito Uchiyama, Masakazu Yamagishi,
Yukihiro Tominari, Yoshinori Takahashi, Junichi Takeya, Yasuo Kitaoka, Yusuke Mori, et al. Growth
of thin rubrene single crystals from 1-propanol solvent. Journal of Crystal Growth, 312(2):310–313,
2010.

[104] Hyeok Moo Lee, Hanul Moon, Hyo-Sik Kim, Yong Nam Kim, Sung-Min Choi, Seunghyup Yoo,
and Sung Oh Cho. Abrupt heating-induced high-quality crystalline rubrene thin films for organic
thin-film transistors. Organic Electronics, 12(8):1446–1453, 2011.

[105] Michael A Fusella, Siyu Yang, Kevin Abbasi, Hyun Ho Choi, Zhuozhi Yao, Vitaly Podzorov, Amir
Avishai, and Barry P Rand. Use of an underlayer for large area crystallization of rubrene thin films.
Chemistry of Materials, 29(16):6666–6673, 2017.

[106] D Käfer, L Ruppel, G Witte, and Ch Wöll. Role of molecular conformations in rubrene thin film
growth. Physical Review Letters, 95(16):166602, 2005.

[107] Tapas Mukherjee, Sumona Sinha, and Manabendra Mukherjee. Electronic structure of twisted
and planar rubrene molecule: A density functional study. Physical Chemistry Chemical Physics,
20:18623–18629, 2018.

[108] Benjamin Watts, Torben Schuettfort, and Christopher R McNeill. Mapping of domain orientation
and molecular order in polycrystalline semiconducting polymer films with soft x-ray microscopy.
Advanced Functional Materials, 21(6):1122–1131, 2011.
Acknowledgements

I would like to thank Prof. Takeaki Sakurai, my academic advisor, for the patient guidance, encouragement
and advice he has provided throughout my time as his student. He spent endless hours proofreading
my research papers and giving me excellent suggestions, which always resulted in improved versions of
documents. Evaluating new topics with him was always interesting for me to investigate and provided me
with alternative ways to improve my dissertation. He taught me how to be a researcher and encouraged
me to determine appropriate solution methodologies to organize research findings.
Besides my advisor, I would like to thank the rest of my thesis committee: Prof. Takahashi Suemasu,
Prof. Masahiro Sasaki, and Prof. Yutaka Wakayama, for their encouragement and insightful comments. I
would also like to acknowledge professor Akimoto Katsuhiro; I am gratefully indebted to his very valuable
comments on this thesis.
I would like to thank the experts who were involved in the validation survey for this research project,
Prof. Satoshi Kera, Prof. Kazuhiko Mase, Prof. Yasuo Takeichi and Prof. Kentaro Kutsukake for
their excellent advises and insightful discussions. Without their passionate participation and input, the
validation survey could not have been successfully conducted.
I also appreciate the financial support of the Japanese government, Ministry of Education, Culture,
Sports, Science, and Technology (MEXT), during my Ph.D. study. Thanks to all my close friends and
Semiconductor Materials Design (Sakurai-Monirul Lab) community of Master and Ph.D. student and
their esteemed families for the joyful environment and all their supports.
My acknowledgement would be incomplete without thanking the biggest source of my strength, my
family. The blessings of my parents Mrs. Aurea Lira Foggiatto & Mr. Vilmar Miguel Foggiatto and the
love and care of my sister Janaina Lira Foggiatto, have all made a tremendous contribution in helping
me reach this stage in my life. I thank them for putting up with me in difficult moments where I felt
stumped and for goading me on to follow my dream of getting this degree. This would always not have
been possible without their unwavering and unselfish love and support given to me.

89
List of achievements

Publications

1. A. L. Foggiatto and T. Sakurai. Charge transfer induced by MoO3 at boronsubphthalocyanine

chloride/α-sexithiophene heterojunction interface.Japanese Journal of Applied Physics, 57 03EE01
(2018);

2. A. L. Foggiatto , H. Suga, Y. Takeichi, K. Ono, Y. Takahashi, K. Kutsukake, T. Ueba, S. Kera, and

T. Sakurai. Dependence of substrate workfunction on the energy-level alignment at organic–organic
heterojunction interface.Japanese Journal of Applied Physics, 58 SBBG06 (2019);

3. A. L. Foggiatto, Y. Takeichi, K. Ono, H. Suga, Y. Takahashi, M. A. Fusella, J. T. Dull, B. P. Rand,

K. Kutsukake, and T. Sakurai. Study of local structure at crystalline rubrene grain boundaries via
scanning transmissionx-ray microscopy. Organic Electronics, 74 315 (2019).

International conference presentation

1. A. L. Foggiatto, Matthias Meissner, Takuma Yamaguchi, Satoshi Kera and Takeaki Sakurai.
Investigation of the permanent dipole on the energy-level alignment at organic-organic interfaces.
The Tenth International Conference on Molecular Electronics and Bioelectronics (M &BE10), Nara,
Japan, June 25-27, 2019.

2. A. L. Foggiatto, Hiroaki Suga, Yasuo Takeichi, Kanta Ono, Yoshio Takahashi, Takahiro Ueba,
Satoshi Kera and Takeaki Sakurai. Dependence of substrate work function on the energy-level
alignment at organic-organic heterojunction interface. The 50th International Conference on Solid
State Devices and Materials (SSDM) 2018, Tokyo, Japan, September 9-13, 2018;

3. A. L. Foggiatto , Takahiro Ueba, Satoshi Kera, and Takeaki Sakurai. Influence of the Work Func-
tion of the Substrate in the Energy-Level Alignment at SubPc/6T Interface. 2018 Joint Symposium
on Energy Materials Science and Technology, Tsukuba, Japan, March 8-9, 2018;

4. A. L. Foggiatto and Takeaki Sakurai. Charge transfer induced by MoO3 at organic-organic

heterojunction interface. Tsukuba Global Science Week, Tsukuba, Japan, September 25-27, 2017.

5. A. L. Foggiatto and Takeaki Sakurai. Charge transfer induced by MoO3 at boron subphthalo-
cyanine chloride/α-sexithiophene heterojunction interface. The 9th International Conference on
Molecular Electronics and Bioelectronics (M &BE9), Kanazawa, Japan, June 26-28, 2017.

90
BIBLIOGRAPHY 91

Domestic conferences presentation

1. A. L. Foggiatto, Yasuo Takeichi, Kanta Ono, Hiroki Suga, Yoshio Takahashi, Michael A. Fusella,
Jordan T. Dull, Barry P. Rand, Kentaro Kutsukake and Takeaki Sakurai. Study of local structure
in rubrene thin films by Scanning Transmission X-ray Microscopy. The 66th JSAP Spring Meeting.

2. A. L. Foggiatto , Yasuo Takeichi, Kanta Ono, Hiroki Suga, Yoshio Takahashi, Michael A. Fusella,
Jordan T. Dull, Barry P. Rand, Kentaro Kutsukake and Takeaki Sakurai. Study of local structure
in rubrene thin films by STXM;

3. A. L. Foggiatto , Takahiro Ueba, Satoshi Kera, and Takeaki Sakurai. Influence of the work
function of the substrate in the energy-level alignment of organic-organic heterojunction interface.
The 65th JSAP Spring Meeting. The 4th TIA Light and Quantum Measurement Symposium.

4. A. L. Foggiatto and Takeaki Sakurai. Charge transfer induced by MoO3 at SubPc/6T hetero-
junction interface. The 78th JSAP Autumn Meeting.

5. A. L. Foggiatto and Takeaki Sakurai. Band Bending induced by MoO3 at organic-organic inter-
face.The 64th JSAP Spring Meeting.