Morphological Studies of Organometal Halide Thin Films for ......solar cells based on organometal halide perovskites have demonstrated outstanding efficiency among photovoltaics. The
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Morphological Studies of Organometal Halide Thin Films for Perovskite Solar Cells
by
Donghan Chen
A thesis submitted in conformity with the requirements for the degree of Master of Applied Science
Department of Materials Science and Engineering University of Toronto
For the last two decades, renewable energy has drawn great attention from research to
investments. As the most easily accessible energy, solar energy can be used at the most place of
the world. Therefore the use of solar energy would be an ideal solution for the increasing energy
demands. Specifically, photovoltaics (PVs), or solar cells, are ideal solutions since they have
many advantages. First of all, solar energy is an abundant energy form that does not rely on the
geographic conditions like other renewable energy. Second, solar cell devices are usually easy to
install and they require relatively less maintaining operation compared to others. Third, solar
cells convert solar energy into electricity so they can directly supply energy for most of our
needs today. Photovoltaic industry has been one of the fastest growing industries in renewable
energy development.
For effective photovoltaic production, it is very important to develop efficient and
affordable techniques. Traditional silicon based photovoltaic techniques cannot decrease
production costs due to the high price of materials and complex manufacturing requirements.
Second-generation photovoltaics, based on thin film technologies, reduce the cost significantly
by simplifying the device manufacturing procedures. Even though this generation’s efficiency is
not as high as silicon based solar cells, the ultra-low manufacturing costs make it a much more
cost efficient technology. Therefore, thin film solar cells have become a rapidly growing, and
increasingly important photovoltaic production type in industry [1, 2].
In terms of categorization, thin film solar cells can be classified into different types
according to their photovoltaic materials, such as cadmium telluride solar cells (CdTe), copper
indium gallium selenide solar cells (CIS or CIGS), and amorphous silicon (a-Si) solar cells.
Several relatively new types of solar cells have emerged in recent years, including quantum dots
solar cells, dye sensitized solar cells (DSSC), organic photovoltaics (OPV, organic molecules or
conjugated polymers), and perovskite solar cells, all of which can also be classified as thin film
solar cells – these are referred to as the third generation of solar cells.
2
Figure 1.1 Comparison of the efficiency of several third-generation photovoltaic technologies [3].
Despite being relatively recent, the popularity of perovskite solar cells has soared since
its efficiency has increased dramatically in a very short time. Figure 1.1 compares the efficiency
development of several new solar cell types (third generation solar cells). The power conversion
efficiency (PCE) of perovskite solar cells reached 15% by the end of 2013, and 19% by 2014 [3],
becoming the highest among third generation solar cells. Perovskites are considered to be great
candidates for solar cell production because of they use inexpensive materials and the same
manufacturing techniques used for other thin film solar cells [4].
3
1.2 Perovskites for High Performance Solar Cells
Today, the term “perovskite” refers to crystal species with structures of ABX3. However,
the term initially only referred to CaTiO3, and was discovered by German mineralogist Gustav
Rose in the Ural Mountains in 1839, and named after Russian mineralogist Count Lev Perovskite
(1792 – 1856). Along with many other compounds found with the same ABX3 crystal structure,
Perovskite became one of most common class of minerals on earth. In 1926, Victor Goldschmidt
described the crystal structure of perovskite for the first time, and illustrated its tolerance factors
[5]. An accurate crystal structure was later published in 1945 using a X-ray diffraction study by
Irish crystallographers [6].
Materials in the perovskites family already presented a wide range of applications, such
as conductors, semiconductors, insulators, and even superconductors [7]. Other physical
properties of perovskite also drew substantial interest, especially in magneto-resistance, ionic
conductivity, and a multitude of dielectric properties [7-9]. One can achieve these properties,
which are of great importance in microelectronics and telecommunications, by either keeping or
modifying the ideal perovskite structure. Due to the flexibility of bond angles inherent in the
perovskite structure, many different types of distortions can occur in the ideal structure,
including tilting of the octahedra, displacements of the cations out of the centers of their
coordination polyhedra, and distortions of the octahedra driven by electronic factors [9, 10].
A group of organic-metal halide perovskites has been found with ideal photovoltaic
properties. Particularly in recent years, methylammonium lead halides (CH3NH3PbI3 and
CH3NH3PbI3-XClX) have shown extremely high power conversion efficiency. Within just two
years, the device efficiency tested in laboratory increased from 8% to 19.3%, which is the
highest among the third generation thin film solar cells [3] . The organometal halide perovskites
have high charge carrier mobility and charge carrier lifetime, which allows light-generated
electrons and holes to move far enough to be extracted as current, instead of creating heat and
losing energy. The effective diffusion lengths of CH3NH3PbI3 are several hundred for both
electrons and holes [3]. For photovoltaic performance of CH3NH3PbI3, open-circuit voltage (VOC)
can approach 1 V, while for CH3NH3PbI3-xClx, VOC > 1.1 V has been reported [11]. Since the
band gaps (Eg) for both materials are 1.55 eV, the ratios of VOC to Eg are higher than what is
4
usually obtained from similar thin film cells. By tuning the band-gap of perovskite, VOC can
reach up to 1.3V, which is the highest performing among thin film photovoltaic devices [3, 12].
Another attention-catching advantage of perovskite solar cells is that the device can be
fabricated in simple planar structure, which drastically simplifies the production process. The
most common spin-coating technique has been shown to be a sufficient method of efficiently
preparing perovskite solar cells. On the other hand, the vapor deposition technique is also
considered as a potential fabrication method that simplifies the production and enhances the
degree of thin film quality control. For example, simple planar heterojunction perovskite solar
cells have already been fabricated without complex nanostructures using the vapor
deposition technique [14]. In addition, compared with other thin film solar cells, perovskite is
free from rare elements requirement, which result in relatively low manufacturing costs.
There remain two areas of concerns for perovskite use in solar cells. The first is the use of
heavy metal (lead) in cell fabrication. Since other substitutions with low toxicity (such as tin
[13]) can potentially be used, this may not be a major concern. Also, compared with the large
amount of lead used in lead-related industries every year, the amount used in perovskite solar
cells is relatively small. A second concern for perovskite solar cell application is the stability of
the organic-inorganic hybrid materials. Since the hybrid materials are very sensitive, the
perovskite films degrade quickly in exposure of ambient environment, and the cell durability is
currently insufficient for commercial use [14]. Since the length of device’s lifetime depends on
the stability of the perovskite thin film, studies of the stability of perovskite are of great value for
understanding the device’s durability, and controlling the production conditions. This research
focuses attention mainly on the investigations of stability in the perovskite materials excluded
from the solar devices and photovoltaic performance. The preparation conditions of pure
perovskites have been optimized and the material tolerances for annealing temperature and air
exposure time have been studied. The discoveries in materials properties will benefit by
controlling the production and post-treatment conditions.
To better understand the fundamentals of perovskite, the following section introduces the
perovskite crystal structure and its distortion features.
5
1.2.1 Crystal Structure of Perovskites
A general perovskite structure in ideal cubic model is shown in Fig. 1.2. The crystalline
architecture consists of three elements A, B and X. Each “A” atom sits at the central position of
the unit cell coordinating to 12 “X” atoms, where “X” atoms are usually O2–, F–, Cl– or other
large ions. “B” cations and “X” ions are coordinated and form [BX6] octahedra. There are 8 [BX6]
octahedra in each unit cell. Therefore, “A” and “X” atoms are close-packed with B is occupying
the centre of octahedra. In the organometal halide perovskites studied in this paper, “A”
represents the methylammonium cation (Ch3NH3+), while “B” stands for the metal cation (Pb2+),
and “X” is the halide ions (I– or Cl–).
In ideal perovskite structure, the unit cell axis, a, can be described by the ionic radii of A,
B and X (rA, rB and rx) by the following equation:
)2()(2a XBXA rrrr . (1)
Figure 1.2 Illustration of ideal ABX3 cubic perovskite crystal structure.
6
1.2.2 Perovskite Distortions
Because of the flexibility of bond angles inherent in the perovskite structure, many
different types of distortions can occur from the ideal model [15, 16]. Only a few compounds,
such as CaRbF3 and SrTiO3, have ideal perovskite structures, and even mineral perovskite
CaTiO3 itself is also distorted [16]. There are three major causes for the distortions in perovskite:
1) the size effect of ions and cations; 2) non-ideal stoichiometric deviation; 3) geometrical
distortion caused by electron configuration, also known as Jahn-Teller effect [15].
First, the perovskite structure easily gets distorted by varying the sizes of A, B or X. To
estimate the degree of distortion in a particular ionic perovskite, the Goldschmidt Tolerance
Factor [5] has been defined as t :
rrrr
XB
XA
2
1t (2)
The ideal cubic perovskite structure has t = 1, due to the high symmetry in cubic system. The
factor t gets smaller with the decrease in the cation A size. When the size of cation A dropped
below a certain value, t will be smaller than 1. In this case the octahedra will tilt to make space
for the large ions, such as the cases in CaTiO3 and GdFeO3 (t = 0.81). But t cannot be smaller
than 0.81, since the structure will then be assigned to an ilmenite structure in that range. For a
system with large A or small B ions than in ideal cubic system, the tolerance factor is larger than
1. The close-packing structure will be stable as a varied hexagonal perovskite, such as BaNiO3.
However, the tolerance factor assumes only ionic bonds existing in the structure so it is limited
to distortion from the ideal perovskite structure.
Second, the non-ideal stoichiometric deviation can cause distortion in perovskite
structures. Taking SrFeOx as an example, the valency of the Fe can be different depending on
either a sample heated in an oxidizing or a reducing environment, whereby the oxygen content
varies from 2.5 to 3 in the perovskite structure. Both +3 and +4 oxidation states can be assigned
to Fe in SrFeO2.875: thus, the FeO5 pyramid structure can be formed instead of octahedra. The
original structure distortion takes place because of the deviations from ideal structure [16].
7
The third distortion type is Jahn-Teller effect, which is a geometrical distortion caused by
a certain electron configuration in order to lower the structure’s overall energy. Particularly,
some perovskite systems hava Jahn-Teller active ions at B position [17]. For example, in the
LnMnO3 in which Ln is La, Pr or Nd structure, Jahn-Teller active Mn3+ ions result in the
elongation of the [MnO6] octahedra [16]. All three types of distortions are illustrated as sketches
in Fig 1.3.
The distortions in perovskite structure have influences on the stability of properties. For
organometal halide perovskites, their hybrid composition enlarges the flexibility of the crystal
structure, which could be one of the causes of their low stability.
Figure 1.3 Illustrative sketches of possible distortions in perovskite structure: (a) ideal cubic structure of perovskite SrTiO3; (b) rotated octahedra structure of perovskite GdFeO3
(Side view); (c) rotated octahedra structure of perovskite BaNiO3; (d) Jahn-Teller distorted octahedra structure of perovskite LaMnO3.
8
1.3 Dye Sensitized Solar Cells (DSSCs) and Perovskite Solar Cells
As a new type of thin film photovoltaics, perovskite solar cells emerged from the
previous generation, DSSCs [18]. It is important to understand the basics of thin film DSSCs and
their connections with perovskite solar cells. Here, we simply demonstrated the typical device
structure of DSSC and its work principle. Also, a historic evolution from the DSSCs to
perovskite solar cells is reviewed in order to illustrate the development of perovskite application
in solar cells, the present device designs, and possible future directions.
1.3.1 Device Structure of DSSC
A typical DSSC is commonly built with a series of layers of materials serving for
specific functionalities for transforming light into electrical energy. Figure 1.4 shows a typical
DSSC structure [19]. The substrate that supports the whole cell structure is usually transparent
glass. A layer of transparent conductive material is coated on glass as anode, which is usually
indium tin oxide (ITO) or fluorine doped tin oxide (FTO). The transparency of the anode allows
the light injecting into the device to provide the energy.
On top of the anode material, n-type semiconductor, TiO2 is typically built for electron
transporting. Commonly, two layers of TiO2 are coated in different morphologies [20, 21]. The
compact TiO2 layer works as electron transport layer in solar cells, and the mesoporous TiO2
layer is usually made into nano-scale structure to provide necessary morphology for dye
sensitizer exhibiting the photovoltaic into its pores. Mixing with the organic dye molecules [22-
26], quantum dots [27-29], or other light sensitizers [30, 31], the porous layer is the essential
layer for absorbing light and generating electron-hole pairs in the device. On top of the TiO2
there is a layer of p-type semiconductor, typically Spiro-OMeTAD, which transports holes from
the photovoltaic material structure. A highly conductive cathode layer finalizes the cell
preparation.
9
Figure 1.4 Illustrative sketch of a typical DSSC structure.
1.3.2 Working Principles
Thin film solar cells are connected to conventional solar cells in their working principle.
For traditional solar cells, both n-type and p-type materials are silicon based layers. The new
generation of thin films solar cells, however, use thin film light absorbing materials.
For traditional solar cells, the performance of photovoltaic cells depends on its core
material, semiconductor, which performs as insulator in their pure form but is able to conduct
electricity in high temperature or combine with other materials. A host semiconductor combined
with electron donor materials develops an excess of free electrons, known as an n-type
semiconductor. A host semiconductor combined with acceptor materials develops excess of
“holes” (or equivalently the removal of electrons), known as a p-type semiconductor. A
photovoltaic cell contains adjacent n-type and p-type materials, and the interface between is
known as a P-N junction. Resulting from the nature of two types of semiconductor, a small
number of electrons always move across the junction from the n-type to the p-type
semiconductor, producing a small voltage output, even in dark. With the presence of light, a
great number of electrons can be activated and flow across the junction thus creating an electric
potential difference at each side. Photovoltaic devices are designed to apply this principle, and
convert solar energy into electrical energy.
10
For the new generation of solar cells, which is the focus of this research, the process of
electricity generation is different from P-N junction-based photovoltaics. After passing through
the transparent glass substrate and anode, the sunlight striking onto the absorbing layer excites
electrons from the materials valence band to the conduction band. Accordingly, the excitation of
the electrons creates a free electron in conduction band and a “hole” in the valence band, which
is referred to as “excitons”. After generating excitons, electrons attract into the TiO2 layer and
conducted into metallic anode. The hole will be replaced by an electron provided by the p-type
material layer, and thus be conducted to the cathode. A potential difference, VOC, for an open
circuit formed from anode to cathode and a current density JSC can be generated by continuous
power supplied form light.
11
1.3.3 Evolution from DSSCs to Perovskite Solar Cells
In thin film solar cells, a variety of organic dyes and inorganic quantum dots were used as
photon sensitizers. Perovskite was also considered as a sensitizer in DSSCs but then discovered
with electron transporting property. This section briefly introduces how perovskites become
photovoltaic material that enables simple planar structured solar cells.
Mitzi and co-workers discovered organic-inorganic halide perovskite as a possible
candidate for thin film transistor and light-emitting diode (LEDs) early 1990s [4, 32]. Since then,
properties of photovoltaic are anticipated but have not sufficiently studied due to concerns of
lead toxicity and low stability [4]. Miyasaka [33] was the first to report the photovoltaic
performance of perovskite in the nanoporous TiO2 layer of dye-sensitized cells. The method used
for their study was to spin coat the sensitized layer of perovskite with solution of CH3NH3I and
PbI2 on top of the TiO2 film. The initial efficiency of their CH3NH3PbI3 cell was 2.2%, and they
were then able to increase efficiency to 3.8% by replacing bromine with iodine. This result was
not a groundbreaking achievement in terms of efficiency; however, it demonstrated the potential
for a new light absorbing material in photovoltaic cells. With the further investigation, perovskite
was proven to be an ideal candidate of light sensitizer, which could achieve appreciable
efficiency.
Subsequently, Park [34] implemented perovskite in similar structure by depositing
sparsely spaced hemispherical nanoparticles that were approximately 2.5 nm in diameter. Along
with surface treatment on TiO2, they achieved an efficiency of 6.5% in 2011 [34]. The
performance of perovskite became comparable to that of organic dyes at this time, but degraded
rapidly since perovskite material can easily dissolve in its electrolyte cell. This lead Park, Gratzal
and his co-works to consider replacing the electrolyte with a solid-state hole transport material,
35]. The perovskite material penetrates the nanoporous structure of the TiO2 layers in all their
structures, and is thus considered a sensitizing layer rather than a fully functioned photovoltaic
layer. It not only improved the stability, as expected, but also it improved efficiency to 9.7% at
2012 [35].
12
Around the same time, Snaith and his workers [19] also reported success in using spiro-
MeOTAD in perovskite structure, and developed photovoltaic cells in several ways. First, they
employed a mixed-halide material CH3NH3PbI3-xClx and improved both the stability and
performance-efficiency of the cell, compared to the pure iodine equivalent. Furthermore, they
capped the TiO2 porous structure with a thin perovskite layer and the device was still functional.
By replacing the TiO2 layer with a similar but non-conducting Al2O3 network, VOC of the device
increased and efficiency boosted to 10.9%. This research demonstrated that perovskite has the
potential to transport both electrons and holes between cell electrodes. The single-step perovskite
deposition process created large morphological inconsistency, and low stability in device
performance. With the method using both perovskite in TiO2 scaffold structure and a capping
layer of pure perovskite overlaying the scaffold, the efficiency jumped to 12.0% [36]. Same
efficiency was observed in a similar structure with Br content in perovskite, and this structure
was proven with high humidity stability [37]. There exists a structural transition from tetragonal
to pseudo-cubic mainly due to a higher t factor, which is caused by the smaller ionic radius of Br.
In 2013, Gratzel’s group used TiO2 scaffold along with two-step iodide deposition, and
improved the efficiency up to 14%. Snaith’s group creatively used a two-source deposition
method, but avoided the previous scaffold structure, and achieved an increased efficiency of
15.4%. This method greatly enhanced the morphology by physical vapor deposition (PVD) and
encouraged our work in this research. A similar structured solar cell with an efficiency of 19.3%
was reported in May 2014 [38].
In summary of the evolution path of perovskite solar cells, perovskite was initially used
as an organic-inorganic hybrid dye in DSSC and also as a solid state DSSC (ssDSSC). It was
substituted for the organic dye molecules, and acted as active material extremely thin absorbers
(ETAs) in DSSC [39]. The discovery of perovskite’s ability to transport charge carriers enabled
it to act as the photovoltaic materials, without depending on the TiO2 electron transporting
materials. Perovskite meso-superstructured solar cells (MSSC) exhibited a high photovoltaic
efficiency on insulation AlO3[40]. Therefore, the possible future directions of perovskite solar
cells could be MSSC, simple planar “p-i-n” thin film solar cell, or p-n heterojunction cells [4, 41,
42]. Figure 1.5 shows the perovskite solar cell with a simplified planar structure, which is a “p-i-
n” thin film solar cell.
13
Figure 1.5 Sketch of a perovskite solar cell with simplified planar structure.
14
1.4 Achievements in Morphological Studies of Perovskite Thin Film
The most progressive morphological study in surface uniformity and consistency is the
co-vapour deposited perovskite thin film, conducted by Snaith’s group in 2013 [43]. This study
discovered significant differences regarding film morphology between solution-processed and
vapour deposited perovskite films. In this work, Snaith’s research team achieved a 5% increase
in efficiency by removing scaffold structure, and simply building the solar cell structure with a
planar heterojunction thin film. The vapour deposition method produced extremely
homogeneous perovskite film with highly uniform thickness[44], which provided contrasts to the
inconsistent thickness and uneven morphology of the solution-processed films (shown in Fig. 1.6
[43]).
Figure 1.6 shows SEM images comparing perovskite solar cells prepared by vapour
deposition and solution processed techniques. The top row photos show that vapour deposition
technique achieved full surface coverage of the film, and the solution processed material
unevenly distributed. The vapour deposited sample is fully covered with perovskite materials and
no vacancies or voids could be observed, whereas the solution processed sample showed a
significant portion of area with substrate exposure. The cross-sectional figure also indicates that
some part of the solution processed film may suffer from short circuits by the varying thickness
from 0 nm to 410 nm, which hinders the performance of the device because of the pinhole
formation. But the vapour deposited film with evenly covered layer of perovskite material is of
no concern in this problem.
This morphological study stressed the advantage of the vapour deposition, in that this
method allows precise control of the thickness of the perovskite film. It is important to optimize
the film thickness since a thicker film is needed to absorb enough light for generating excitons.
However, one needs to maintain the thickness at a thin enough level to allow for the transport of
electrons and holes. Last, optimized perovskite with film thickness of 330 nm indicates that
electron-hole diffusion length exceeds such a length in perovskite material.
The investigation of stability of the perovskite films is also discussed in this report
regarding solution processed perovskite thin film [44]. Using a scanning electron microscope
15
(SEM) is the proper technique for examining the morphology change of the perovskites surface.
Particularly, the surface coverage rate is an important parameter that can be used to describe the
film’s formation, or the surface loss in post-treatment. SEM has thus been used in this research,
and changes of surface morphology and coverage are used to evaluate the degree of degradation.
Figure 1.6 Thin-film topology characterization: (a-d) SEM images of top-view and cross sectional view of perovskite thin film and perovskite solar cells made by solution processed technique and vapour deposition technique. (e-f) SEM images of large cross sectional view
images of perovskite make by solution processed technique and vapour deposition technique [11].
16
1.5 Motivations and Thesis Overview
Thin film solar cell technologies are important in making photovoltaics more affordable.
One goal of studies in thin film solar cells is to achieve of high efficiency devices with easy
processing methods and accessible element resources. Organometal halide perovskites based
solar cells have been demonstrated to have significantly high efficiency. The morphology
conditions of solution-processed perovskite, such as uniformity of thickness and surface
coverage, are key factors to control film quality and device performance. Using the vapour
deposition method in high vacuum system, we prepared perovskite thin films with uniform
thickness and full surface coverage.
In this study, we aimed to optimize the vapor deposition conditions for preparing
perovskite thin films that have been used in a number of solar cells. Perovskite samples were
prepared and annealed under a series of temperatures. By checking the crystal structures with X-
ray diffraction (XRD), and surface morphologies with scanning electron microscope (SEM), the
effect of different annealing temperatures was investigated. The stability of the perovskite thin
film was studied by examining the changes of the material exposing in air. AFM was used to
check high-resolution surface information that cannot be captured by SEM.
In sum, this chapter introduced the fundamentals of perovskite structures, and provided a
brief recount of the evolution from DSSC to perovskite solar cells. This chapter also reviewed a
study of morphological control of the thin film perovskite. Next, chapter 2 illustrates the
materials used in preparing the perovskite thin film, as well as the experimental details of our
study. Chapter 3 shows the experimental results and the conclusions of our investigation. Last,
Chapter 4 is a brief summary of the entire thesis and several suggestions on future research.
17
Chapter 2 – Materials and Experimental Methods
2.1 Materials
Our organometal halide perovskite is made from organic and inorganic sources through
the vapour deposition technique. The organic source is methylammonium iodide (CH3NH3I), and
the inorganic source is lead chloride (PbCl2). CH3NH3I was synthesized by reacting methylamine
with hydroiodic acid at 0 °C for two hours, while stirring. PbCl2 was purchased from sigma-
Aldrich. This chapter will describe materials and experimental methods, including both solution
processing and vapour deposition. Principles of characterization method used in this study will
also be introduced.
2.2 Methods for Perovskite Thin Film Preparation
There are two main techniques for preparing the perovskite thin films, solution processed
technique and vapour deposition. The solution processed method is more commonly used
because of the simplicity of equipment set-up. Vapour-deposition requires complex experimental
conditions, but it pays off by providing the thin film with highly uniformed thickness. In this
study, we mainly focused on perovskite thin films that were made by vapour deposition in high
vacuum system.
2.2.1 Solution Processed Technique
For perovskite thin film preparation, the most commonly used technique is spin coating
the precursor solution on substrate, followed with an annealing treatment. In early studies of thin
film perovskite solar cells, perovskite precursor solution was prepared and spin coated directly
onto substrate to form thin films [19]. The precursor solution is usually the methylammonium
lead halide (CH3NH3PbI2Cl or CH3NH3PbI3),) in N, N-dimethylformaide. Spin coating is usually
conducted at ambient conditions, and the annealing procedure is carried out on sample
immediately after spin coating. Before annealing, the thin film has lower crystallinity (smaller
crystal size) and contains an excess of PbI2 component mixed in the thin film. With the annealing
process, the perovskite film forms larger crystal grains, and pure perovskite crystal structures [3].
Some other methods, based on the solution processed technique, were developed by
varying part of the original procedures in order to obtain perovskite thin films more easily, or
18
with higher performance. A sequential solution deposition method was carried out with high
efficiency in device performance, based on the previous study [14]. Instead of spin coating
perovskite solution directly onto the substrate, the inorganic component PbI2 solution was first
spin coated, and then the substrate was dipped into a CH3NH3I solution. Further modification
was made to introduce the organic component by vapour treatment on the substrate, spin coated
with an inorganic source [11]. CH3NH3I powder was then spread out around the PbI2 coated
substrates in a covered petri dish. By heating the petri dish for desired time, perovskite formed
from the CH3NH3I vapor-treated PbI2 substrate. Although preparation methods seem to differ
from each other in details, all have been proven to be effective ways for perovskite film
preparation. High performance devices with efficiency higher than 12% can be obtained by all
the methods discussed above.
2.2.2 Vapour Deposition Technique
A novel method is to use the vapour deposition technique to produce highly pure
perovskite thin films with high device efficiency. For the vapour deposition technique, thin films
are prepared onto substrates through condensation, or the reaction of vaporized materials. The
preparation procedure is conducted in a multi-technique vacuum system, which has a central-
distribution chamber (CDC) with several sub-chambers located around it. All chambers are
connected and kept in a vacuum at a pressure between 1×10-10 to 5×10-9 Torr. Perovskite thin
films were deposited in an organic deposition chamber, which allows for relatively low
temperature deposition compared to the oxides deposition chamber and the metal deposition
chamber. In the organic deposition chamber, chemical sources are evaporated from the
permanently mounted Knudsen cells (K-cells) whose heating temperature can be precisely
controlled. The two source materials for perovskite preparation are vaporized in the vacuum
system as the gas phase. They interact on the surface of substrate located on the top of K-cells,
and where the chemical reaction takes place forming the thin film.
The illustrative diagrams of organic deposition chamber are shown in Fig. 2.1. Each part
of the chamber is labeled with a letter. Part (a) is the chamber body where the evaporation
deposition takes place. Part (b) is a transfer-arm evaporator cell (TAE-cell) that allows substrate
moving into and out of the chamber body. A sample holder, part (c), is connected at the end of
the transferring arm and it can carry four pieces of substrate for each deposition. Substrates
19
prepared for perovskite deposition are placed at the sample holder so that the substrate can be
delivered into the chamber. Part (d) is a quartz crystal microbalance (QCM) used for monitoring
the films’ thickness. Part (e) is the shutter that controls the exposure of substrate to the chemical
vapour. Part (f) is the connector to the central chamber, while part (g) is the K-cell. These K-
cells are isolated from one another and the temperature of K-calls can be individually controlled.
Each K-cell has its own shutter that controls the evaporation for coming out from the K-cell. Part
(h) is the ion pump used to maintain the required vacuum. The lower panel of Fig. 2.1b illustrates
the top view of the organic chamber.
Figure 2.2 describes the vapour deposition process for perovskite thin film preparation.
During the deposition process, organic source CH3NH3I and inorganic source PbCl2 are placed in
different K-cells. Once the K-cells are heated to the preset temperature, the chemicals start
evaporating and moving upward to the substrate. When both PbCl2 and CH3NH3I vapor reaches
onto the substrate, a chemical reaction is initiated, and begins to form perovskite films on the
substrate’s surface. At the same height level of the substrate holder, there is a QCM device serving
as the monitor for film thickness. QCM measures the mass per unit area by measuring the change
in frequency of a quartz crystal resonator. The resonance changes with increases of a small mass
due to the film deposition at the surface of the acoustic resonator. Film thickness can be
monitored based on the mass addition on unit area. Usually, there is some difference from the
real deposited thickness to the QCM readings. The ratio of the real deposition rate over the QCM
reading is defined as the tooling factor of the QCM. Each QCM has a specific tooling factor in a
certain deposition system. During the process, the system is maintained at pressure of 10-6 – 10-7
Torr. As such, only chemical sources evaporated upon the film participate in the chemical
reaction, which ensures the purity of the reagents and uniformity of the film thickness.
For solution processed perovskite, it is known that the annealing procedure assists film
formation and material crystallization[14, 34]. Most of the films after spin coating need an
annealing process for better crystallization and film formation. To prepare an ideal perovskite
thin film with solution processed method, it is very important to control the annealing
temperature and treatment time. For vapour deposited perovskite, however, there is little research
that addresses the annealing effect from previous studies. This dearth of information provided
motivation for us to conduct an investigation on the annealing effects for perovskite over vapour
deposited thin films.
20
Figure 2.1 Illustrative sketches of organic chamber used for vapor deposition: (a) Side view of the organic chamber (b) top view of the organic chamber. The label of each part a – h is
described in page 18.
21
Figure 2.2 Illustrative sketch of vapour deposition process.
22
2.3 Characterization
2.3.1 Powder X-Ray Diffraction (XRD)
XRD, a traditional technique for crystal characterization, examines long-range ordering
and phase purity. The periodicity of the electron density in a crystal structure generates X-ray
scattering from electrons, and causes coherent “diffraction pattern”. The intensity of the
diffracted X-ray can be plotted as the function of angle 2θ (θ is the angle of electron and the
crystal plane) and thus depict the powder diffraction pattern. Peaks appear in the powder
diffraction pattern at angles with maximum interference that satisfy Bragg’s Law [45]:
sin2dn (5)
Each crystal structure has a unique diffraction pattern with characteristic peaks and
relative intensity. The XRD patterns of known structures has been tested and archived in
databases. Therefore, one can compare the experimental XRD patterns with the standard
patterns in order to identify the sample structure. In this paper, all X-ray diffraction patterns of
the perovskite thin films were obtained on an X-ray diffractometer (Panalytical X’Pert Pro), with
Cu-Kα radiation (λ=1.54056Å). Figure 2.3 shows an illustrative diagram of Bragg’s Law, as
well as sketches of the working principle of a XRD measurement.
To verify the crystal structure of our perovskite thin film samples, we referred to the
database and labeled the theoretical peak positions of perovskite structure. We can thus compare
our XRD patterns with standard XRD patterns, and check if the thin film prepared by vapour
deposition also has perovskite crystal structure. To investigate the effect of other impurity crystal
structure, XRD patterns of possible lead halide (PbI2 and PbCl2) are also obtained. As noted,
XRD is an ideal tool for checking the crystal structure evolution in annealing process; results of
XRD tests provide strong support for the investigation of our thin film and thin film changes.
23
Figure 2.3 (a) Illustrative diagram of Bragg’s Law; (b) Principle of the XRD measurement [46].
24
2.3.2 Scanning Electron Microscopy (SEM)
SEM is an imaging tool that uses a beam of focused electrons, scanning on sample
surfaces, to produce images. The scanning electrons interact with atoms on sample surface
producing signals of surface topography. There are several types of signals produced by SEM,
such as secondary electrons (SE), back-scattered electrons (BSE), characteristic X-rays, cathode-
luminescence (CL), specimen current and transmitted electrons. Secondary electrons are the
most commonly used signal type. During the measurement, the electron beam scans over the
surface of the sample dot-by-dot and line-by-line, producing many types of signals. Secondary
electrons that are dislodged from the surface atoms have unique patterns at each dot. A detector
that counts the secondary electrons scattered from the sample surface receives information of
secondary electrons of the sample. Other sensors also detect BSE, X-ray, and other signals. By
detecting the information along with the scanning beam by different sensors, a great deal of
information is processed by the computer, and displayed by different levels of brightness on a
monitor.
SEM images were collected at the Centre for Nanostructure Imaging at Chemistry
Department of University of Toronto using the Quanta FEG 250 environment SEM with both a
bright field and dark field detector. The crossbeam combines a high resolution SEM for imaging,
with a focused ion beam (FIB) for micromachining by sputter milling with a sub-100 nm lateral
resolution. The spatial resolution of the images can focus down to 1 nm, depending on the
material’s conductivity.
SEM is a powerful imaging tool that can be used in conductive and semi-conductive
materials. For our perovskite sample, it is the proper choice with which to image surface
topography. By detecting the surface of perovskite sample, we are able to understand the surface
uniformity, thickness, and even chemical components in a rough scale. In addition, the change of
the surface morphology observed from SEM benefits to explain the possible material affect.
25
2.3.3 Atomic Force Microscopy (AFM)
Distinguished from its predecessor – Scanning Tunneling Microscopy (STM), which relies
on tunneling current between scanning tip and sample surface atoms – AFM mechanically
contacts with materials directly. Therefore, AFM affords the possibility of observing on almost
any type of materials.
Working principle of AFM is also relatively simple among advanced characterization
techniques. Typically, AFM contains three main parts, as shown in Figure 2.4: the scanning
system, controller, and computer. Samples are located on a piezoelectric scanner that provides
precise movement on x, y and z direction. A very sharp tip attached to a soft cantilever scans on
the sample surface line-by-line with the motion of the piezoelectric scanner. During scanning,
the controller adjusts the z canner to maintain a constant tip-sample force value to make sure the
tip-sample distance has a constant value. At the backside of the tip, a beam comes out of the
laser source and reflects to a photodiode detector. With the monitoring of the laser beam signal,
height change and deflection information of tips can be precisely captured by the controller.
Thus, morphology images can be constructed in three dimensions by recording height change
during lateral dimensional scanning.
Figure 2.4 Illustrative sketch of a typical AFM system.
26
AFM achieves extremely high resolution on sample surface morphology, depending on
sharpness of the probe tip, precision of the scanner, and the optical detection system. The probe
tip usually has a radius of a few nanometers (~10nm), which only has a small amount of atoms at
the end. High sensitivity piezoelectric ceramic provides accurate three-dimensional
displacement. The long path of the laser beam amplifies slight changes of the tip, and the laser
signal is precisely captured by photodiode. Tip-sample interaction also helps AFM achieve
atomic level resolution. Tip-sample interaction may be described by the Lennard-Jones potential
[ω(r)] [47], which considers the ideal interaction between two atoms:
rr
126
BAω(r) (6)
Where r is the distance between two bodies and A, B are constants for fixed atoms. The
force interaction is then:
rr
127
B12A6
dr
dωF (7)
Figure 2.5 The dependence of force on the probe distance.
27
As shown in the above equation, as well as in Figure 2.5, the Lennard-Jones force depends
directly on the probe distance. The curve increases steeply when separation distance is less than a
certain benchmark, which provides the sensitivity to detect slight height changes from repulsive
force.
Different scanning modes are available, depending on several sample-tip relative motion
types. In contact mode, the tip scans under an applied force, and a feedback system keeps the
tip-sample interaction force in a constant value. In dynamic mode, oscillating cantilever beats on
the sample surface to minimize the tip-sample contact effect, and improves the ability to detect
soft materials. The oscillation amplitude and phase change are taken as parameters to report
surface information. Therefore, AFM not only depicts morphology information, but also
presents material surface property by phase shift information.
Although AFM is a powerful tool for characterizing a variety of materials in different
environments, it still has some limitations. As a result of the sharpness of the tip, the AFM used
in this research can only scan within 100μm2 in image size, and the scanning speed is much
slower than other characterization methods, such as electron microscopy. Unsuitable tip shape
introduces image artifacts in the morphology image. Artifacts are mainly due to changes in tip
shape, such as sticking impurities from sample surface, or a small piece knocked off from the tip
end. As a detector originally designed for flat materials, AFM works poorly on morphology with
steep walls or overhangs.
28
Chapter 3 – Results: Morphological Studies of Perovskite Thin Films
3.1 Introduction
Organo-metal halide perovskites have garnered great attention over the last three
years[48, 49] because they provide the highest efficiency among third-generation photovoltaic
materials. The vapour deposition technique allows for the making of perovskite thin film, which
can potentially be assembled into photovoltaic devices through easy fabrication. Compared with
solution-processed methods that were previously common, vapour deposition possesses several
advantages over controlling the film thickness and morphology. First, by controlling the
deposition rate of two chemical sources, vapour deposition precisely controls thickness at the
nanometer level [43]. Second, vapour deposited perovskite thin films have uniform morphology
suitable for further device assembly [36, 44]. For an ideal perovskite thin film, uniform
morphology effectively avoids the possible short circuit in the device, and provides
homogeneous performance across the entire device. As such, the uniformity of morphology
contributes to enhanced efficiency of such devices. Lastly, vapour deposition techniques are
easier for large-scale production, which is a key factor when one considers the potential for
future manufacturing.
The primary goal of this research is to make perovskite structures with organic and
inorganic source chemicals in a PVD system, which has been described in the experimental
section. There are several parameters that need to be optimized in order to obtain the perovskite
thin film. Two major factors for making highly crystalized perovskite are: deposition rate, and
ratio of two source chemicals. By varying the temperature of the K-cells, we controlled the
deposition rate of the film as well as the organic-inorganic ratio. Optimum temperature of each
K-cell was decided by checking the crystal structure of the thin film with XRD.
Besides preparing perovskite thin films, another important issue is perovskite stability.
The stability of perovskite is the main factor determining device performance and lifetime. It is
also decisive in choosing fabrication conditions (e.g. vacuum requirement). Therefore,
investigations on stability of perovskite thin film will benefit from understanding the thin film
properties, and predicting device performance and lifetime. Instead of checking the perovskite
29
performance and stability in solar devices, two studies were conducted directly on the perovskite
thin film. The first study focused on the effects of annealing temperature on film formation and
degradation. Annealing is one of the key procedures aiming to improve perovskite performance.
By heating the film above a certain temperature, materials become more homogeneous and
exhibit better properties. In this study, we observed the effect of annealing temperatures, and
examined the temperature tolerance of perovskite thin film. To further investigate the stability of
the perovskite film, we studied the degradation behavior of perovskite thin film exposed to air.
Morphological information such as film thickness, surface coverage, and crystal grain
sizes can assist in evaluating the rate of film formation, consistency of the preparation
conditions, and changes upon treatments. Solution processed thin film with rough surface
morphology is not suitable for monitoring the change of film surface [43]. However, vapour
deposited film can achieve very uniform surfaces as a baseline for morphology studies, and
enables capturing morphology changes. With a uniform surface, we clearly observed thin film
morphology changes with varying annealing temperature, air exposure time, and other conditions
that may cause the perovskite degradation.
It is important to note that the vapour deposition method confronts several obstacles
during experimental procedures. First, methylammonium iodide (CH3NH3I) does not stick on the
substrate alone, so it is very hard to monitor its deposition rate, and precisely control the organic
component independently. Second, the CH3NH3I powder easily gets clustered upon heating. The
ideal solid material should disperse freely as powder; thus, the clustering makes it difficult to
maintain a stable evaporation rate. Last, the deposition system needs to be calibrated often so as
to keep the same experimental condition from batch to batch, especially the tooling factor of the
QCM. In all, preparation of perovskite sample is time consuming, and a great number of samples
are needed to obtain significant data points.
In this chapter, we described conditions for preparing perovskite thin film with uniform
thickness by the vapour deposition method. By controlling the heating temperature of K-cells,
we successfully optimized the deposition rate of the CH3NH3I and PbCl2 in order to obtain
perovskite thin film. The relationship of the deposition rate and the temperature was also studied.
The crystal structure was examined using XRD, while the surface morphology was studied using
SEM and AFM. Surface coverage changes were also investigated.
30
3.2 Experimental Details
Materials for vapour deposition of perovskite thin films are prepared separately.
CH3NH3I was usually synthesized with following method: 24 mL of CH3NH2 reacted with 10
mL of HI in a 250 mL round-bottom flask at 0 °C for two hours, with constant stirring. The
precipitate of the reaction was then collected using a rotary evaporator, by carefully removing
the solvents at 50 °C. The obtained product was re-dissolved in 80 mL absolute ethanol, and
precipitated with the addition of 300 mL diethyl ether; this procedure was repeated twice for
purification. The final product of CH3NH3I was collected and dried at 60 °C in a vacuum oven
for 24 hours. The Ted Sargent group from the Department of Electrical and Computer
Engineering provided the CH3NH3I, and PbCl2 was purchased from Sigma-Aldrich.
Substrate cleaning is a key procedure in guaranteeing final thin film uniformity in
thickness. A substrate with insufficient cleaning will cause uneven film patterns, which lack
uniformity. As we aimed to image the morphology of the perovskite film with SEM and AFM,
very flat substrates were desired. As such, we chose a silicon wafer as the substrate that would
provide flat surface, and would also allow the film to form on the surface.
The cleaning procedure of the silicon wafer consists of three steps. The first step is to
thoroughly wash the substrates with water, acetone, and methanol, in that order; extra wiping or
sonication may be required depending on the surface condition. The use of a different solvent is
to ensure that the substrate was cleaned with solvents of different polarity. The next step is to dry
the substrate with a flow of nitrogen gas right after the solvent wash. Finally, the substrate is
treated with UV-Ozone radiation. The UV-Ozone treatment not only cleans the surface with
high-energy radiation, but also it oxidizes the surface in order to gain higher work function and
surface potential, which helps the deposition process and enhances the adhesive property. The
whole cleaning procedure was conducted immediately before the deposition, ensuring the least
air exposure after cleaning. In addition, the color of the film is also a key indication of film
quality. So glass slides were used as substrate to prepare sample for colour check.
31
Vapour deposition process is carried out on substrates following sufficient clean. Varying
the heating temperature of the K-cells that contain source materials – CH3NH3I and PbCl2 –
controlled the evaporation rate. The common temperature used for heating PbCl2 is around 330
ºC, and for CH3NH3I it is 125 ºC in the chamber, with the pressure of 10-6 – 10-7 Torr. These
optimal temperatures were decided by varying the temperatures of K-cells and recording the
deposition rate of each source. The device performance and the thin film quality depend greatly
on the component of the CH3NH3I and PbCl2. To optimize the component ratio of CH3NH3I to
PbCl2, the ratio of 1:1 to 10:1 were conducted, and the optimum ratio was found to be around 8:1
for making perovskite thin film in this system.
Since it is difficult to deposit CH3NH3I on a silicon substrate when deposited alone, we
set the QCM as the set of PbCl2, and the deposition rate of CH3NH3I was calculated accordingly.
The optimum perovskite film thickness is around 100 nm, and each deposition takes around two
hours. This thickness is proper for subsequent characterizations [43].
After the deposition process, the sample was taken out from the chamber for annealing.
As discussed, the perovskite is very sensitive especially to moisture; as such, we kept the sample
in a glove box except when annealing or characterizing.
The samples are annealed in a vacuum oven to separate the sample from air and humidity
during heating. Some reported studies did not keep the sample in vacuum during annealing [44,
50]. Instead, the device was annealed in air with the relative humidity at around 30%. Our study
kept the same from exposure to air, since moisture in air was considered to be a major factor in
degradation of perovskite film [50]. All annealed samples were heated in the determined
temperature for 20 minutes. To anneal samples at different temperatures, certain number of
choice sample were taken out of the glove box. During the annealing process of the first sample,
other samples were exposed to air. As such, the second sample actually stayed in air for 20
minutes more than the first sample.
X-ray diffraction (XRD) is a powerful technique with which to study the long range
ordering of perovskite crystal. It is used to examine the crystallinity of perovskite material,
which indicates the purity of the thin film and the grain size of individual crystal. The tolerance
of annealing temperature was detected by the XRD, and showed how the structure changes along
32
with increases to the annealing temperature. Taking advantage of the morphology uniformity
provided by vapour deposition, it is convenient to study perovskite thin film morphology with a
series of characterization methods. Using SEM, we can observe morphology of the perovskite
thin film. AFM is a great tool with which to present morphological changes in very high
resolution.
Characterization: X-ray diffraction pattern (2θ scans) of the perovskite thin film were
obtained on an X-ray diffractometer (Panalytical X’Pert Pro), with Cu-Kα radiation
(λ=1.54056Å). SEM analysis was performed on a Quant FEG 250 environmental SEM. The
AFM used is a JPK NanoWizard II AFM. A cantilever with nominal spring constants between 40
and 50 N/m (NCH probes, Nanoworld Innovative Technologies) was operated under the
dynamic force mode. In this mode, the cantilever is vibrated at around the resonant frequency,
and its amplitude reduces when the tip is in proximity to the sample surface, which is caused by
tip – sample interaction. Reduced amplitude is set as the feedback parameter (set point) so that
the AFM system scans the surface contour of the sample with minimized error signals (the
difference between the set point and the amplitude measured) by adjusting the distance between
the tip and the sample surface. Mapping of this distance constructs a topographic image of the
surface morphology. Mapping the error signal resulted in an image removing the height
contribution, and stressing only the shape of surface features. When the height range is large,
surface features with small height differences are obscured. In this case, it is advantageous to use
the error signal image in order to show the shapes of surface features, while using the
topographic image to estimate the height distribution. The scan rate for obtaining images is 1 Hz.
The experiment was conducted in air with a relative humidity of ~40%.
33
3.3 Results and Discussion
For each deposition, four pieces of substrates were aligned in the sample holder so that
four identical thin films could be obtained. Usually, one of the four substrates is a glass slide that
allows for a quick check of the film colour. The samples of 100 nm-thick perovskite film made
by vapour deposition technique are shown in Fig. 3.1. Only samples deposited on glass are
shown here because of the significant contrast between film and substrate. Samples deposited on
silicon wafers are not well depicted on a photo with clear contrast. The left sample in Fig. 3.1
with light yellow colour is a typical failure deposition due to insufficient perovskite formation on
the substrate surface. The yellow colour is mainly due to PbI2 mixed in the film, indicating that
PbI2 is the intermediate in the reaction between CH3NH3I and PbCl2.
A successfully-deposited perovskite film deposition should have dark color, indicating it
absorbed most of the visible light (Fig. 3.1 sample on the right). The color of the film also gets
darker with the increase of the thickness. From the colour of the sample, we can quickly check if
the deposition constitutes a desired perovskite film.
Figure 3.1 Photos of 100-nm thin film perovskite samples made by the vapour deposition technique: left photo, a typical failure-deposition on glass; right photo, a typical successful
deposition of the organometal halide perovskite.
34
The ideal temperatures for making a good perovskite CH3NH3PbI3 are around 330 ºC for
PbCl2 and 125 º C for CH3NH3I. To set the deposition condition, we fixed the deposition rate of
PbCl2, and adjust the deposition rate of CH3NH3I. Controlling the temperature at these conditions
provides the ratio of deposition rate of CH3NH3I:PbCl2 is 8:1. Under this ratio the deposited thin
film easily shows good perovskite structure as examined with XRD. For another previous study
using vapour deposition, the temperature of PbCl2 and CH3NH3I are 320 ºC and 116 ºC [43]. The
possible reasons for the different deposition temperatures could be the result of many factors.
Different chamber geometry, especially the distance from the K-cells to the substrate, has an
influence on the deposition rate. Also, if the vacuum conditions are not exactly the same the
deposition temperature may change form one system to another.
3.3.1 Characterization of Perovskite and Related Materials
According to previous studies of perovskites, the main XRD peaks assigned to the (110),
(220) and (330) crystal planes are at 14.1º, 28.4º, and 43.2º, respectively [19, 36]. Based on
theoretical calculations (100), (200), (300) peaks are also closely aligned at these three respective
positions with lower intensity. At 2θ = 14.1º, (100) the peak is usually covered by the (110) peak,
so a small peak should would be expected. And similar situations happen at 28.4º for (200) and
(220) peaks, and 43.2º for (300) and (330) peaks. Fig. 3.2a shows an XRD pattern of a typical
pure CH3NH3PbI3 perovskite thin film. In this pattern, three main diffraction peaks at 14.1º,
28.4º, and 43.2º presents with strong intensity. But each peak’s width is broader than a typical
single peak, indicating that the signal is a doublet of two peaks. Solution processed perovskite
have shown identical peaks in previous studies, which demonstrate that both techniques are able
to produce perovskite with same structure [19].
In order to interpret the XRD results of perovskite made by vapour deposition, several
standard patterns of related materials are also examined, including the inorganic source PbCl2
and the possible side product PbI2. PbI2 has a very high intensity of (110) at 12.65º, as shown in
Fig. 3.2b; with two other notable peaks at 39º and 52.5º. PbCl2 has a more complex pattern (Fig.
3.2c), two broad peaks at around 14º and 27.8º, five sharp peaks from 22.5º to 30º, and some
other minor peaks around 40º - 50º. With the XRD patterns of PbI2 and PbCl2, the deposited
perovskite structure can be explained if there are mixed components of PbI2 and PbCl2 in
resultant thin films.
35
.
Figure 3.2 XRD patterns: (a) Pure perovskite thin film of CH3NH3PbI3; (b) PbI2 powder purchased from Sigma-Aldrich; (c) PbCl2 powder used for film preparation purchased
from Sigma-Aldrich.
36
3.3.2 Annealing Effects for Perovskite Thin Films
The annealing process is one of the key procedures during solar cell fabrication. In
materials science, annealing is a heat treatment that alters physical and sometimes chemical
properties in order to improve the working properties. This heat treatment usually consists of
heating a material to above a certain temperature, maintaining that for certain time, and then
cooling to room temperature. Commonly, annealing is applied to soften material, relieve internal
stresses, and refine material structure for making it homogeneous, which eventually makes the
material workable with certain ideal properties. This study investigated the annealing effects on
perovskite thin films. As described in the experimental section (chapter 2), the annealing
procedure is conducted in vacuum oven under 260 ºC, and on a hot plate over the same
temperature – this is done because of the heating limitations of the vacuum oven.
Perovskite preparation requires precise control of the evaporation system, including its K-
cells temperature and QCM readings. Only several batches of samples are presenting signature
peaks of perovskite, so the study is limited to discuss only a few samples with significant
perovskite structures.
Powder X-ray diffraction was used to study the effect of annealing, and selected patterns
are shown in Fig. 3.3. Pattern (a) is the as-made sample (having been just removed from the
deposition chamber), and (b) is a sample annealed at 140 ºC in a vacuum oven. The pattern of
as-made sample shows (110), (220), and (330) peaks of perovskite at 14.1º, 28.5º, and 43.2º,
respectively, demonstrating that perovskite material has been synthesized by the vapour
deposition method. These sharp peaks are dominant signals in the pattern, which means that
CH3NH3PbI3 perovskite has been prepared as the main composition of the thin film. And as
discussed in last section, the widths of all three characteristic peaks are broadening by some
shoulder peaks. These peaks are (100), (200), and (300) peaks, covered under (110), (220), and
(330) peaks. However, this pattern also presents a PbI2 (110) peak at 12.6º, 27º, and 39º, which
reveals that the thin films are actually a mixed composite of CH3NH3PbI3 perovskite and PbI2.
Previous XRD studies have observed peaks of CH3NH3PbCl3, indicating the existence of a
mixed halide perovskite of CH3NH3Pb3-xClx. In this study, however, a peak of CH3NH3PbCl3
was not noticeable. This is mainly because the “x” is usually much smaller than 0.1, so XRD is
not necessarily able to detect its existence.
37
According to the observation of PbI2 peaks in perovskite thin films, the chemical process
of the film formation reaction of organic and inorganic source chemicals can be explained.
During the vapour deposition, the reactions taking place on the substrate can be described
through the equation below:
2CH3NH3I + PbCl2 = 2CH3NH3Cl + PbI2 (1)
PbI2 + CH3NH3I =CH3NH3PbI3 (2)
Pattern (b) is the sample annealed at 140 ºC in a vacuum oven. The pattern has the same
peak positions with a slight enhancement in the signal intensity. This enhancement attributes to
two changes in perovskite thin film. First, the annealing process helps the crystallinity. Second,
the perovskite crystal grains tend to gain size during the annealing process, because it provides a
condition for recrystallization of small grains into bigger crystals. In all, the study supports the
finding that annealing assists the crystallinity of perovskite thin film. This conclusion was
confirmed by further study on the annealing process, and will be explained in detail in the next
section.
38
Figure 3.3 XRD patterns of: (a) perovskite sample before annealing treatment and (b) perovskite ample after annealing treatment in 140 ºC for 20 minutes.
39
3.3.3 The Effects of Annealing Temperature on Perovskite Thin Films
3.3.3.1 XRD Results
To further examine the influence of annealing temperatures, we used an XRD to examine
the crystal structure of the perovskite film after annealing at various temperatures. In order to
guarantee that the thin films are identical before annealing, four pieces of film samples were
deposited in one batch so that each sample could be regarded as one part of the same film. After
each deposition, one of the film samples was used to check the purity of the perovskite with
XRD. To make deposition conditions the same, all parameters were maintained between each
deposition, and depositions in this section were conducted continuously without pauses or
sources evaporating.
The selected patterns of perovskite thin films after annealing at different temperatures are
shown in Fig. 3.4. The 100 ºC annealed sample has identical peak alignments with the 140 ºC
annealed sample, as well as the non-annealed sample shown in Fig. 3.3. Moreover, its peak
intensities are also comparable with the 140 ºC annealed sample, indicating improvement of
crystallinity after annealing at 100 ºC. This study found that for temperature ranges from 90 ºC to
160 ºC, 140 ºC is the optimal temperature for annealing treatment because the crystal structure
detected by the XRD has the greatest-intensity peaks, whereas the surface coverage still stays at
ideal level for devices.
The sample annealed at 180 ºC exhibits the disappearance of the main (110), (220), and
(330) peaks of perovskite at 14.1º, 28.5º, and 43.2°. Comparably, the peaks of PbI2 become
dominant at 12.8º, indicating the PbI2 is the major remaining crystal structure at this stage. This
is indicative that perovskite has completely disappeared after annealing at 180 ºC
After annealing at 220 ºC, the XRD pattern only shows the peak of PbI2 at 12.5º.
However, samples heated at temperature beyond 260 ºC led to the emergence of a very broad
peak at 27.0º, and a noticeable sharp peak at around 33.5º. These two peaks are identical to those
of PbCl2. The sample annealed at 400 ºC only shows weak signal of PbCl2.
40
Figure 3.4 XRD patterns of perovskite thin films annealed at temperature ranging from 100 ºC to 400 ºC. Peaks labeled with (*) are characteristic peaks of perovskite; peaks
labeled with (↓) are PbI2 peaks; peaks label with (▼) are PbCl2 peaks.
The XRD results support the perovskite thermal dynamic property observed in our daily
work. As mentioned, prepared thin films with darker color show identical perovskite peaks in
their XRD patterns. Along with the increase of annealing temperatures, the sample thin films
tend to become a lighter yellowish colour, and almost transparent after high temperature
annealing (300-400 ºC). This suggests that the dark color of the sample film is mainly
attributable to the perovskite material, rather than PbI2.
In addition, the study clearly indicates that perovskite material CH3NH3PbI3 is eliminated
from the film sample as the temperature increases. The threshold temperature is between 140 –
180 ºC, and the disappearing happens in a small temperature range without involving any new
structural evolution in the 180 ºC sample. The main concern of CH3NH3PbI3 dissolving concerns
their reaction with moisture in air, as noted in previous studies [3, 51]. However, the samples are
41
annealed in a vacuum oven, which is a condition lacking of moisture. The possible dissolving
process can be described as:
CH3NH3PbI3 = CH3NH3I + PbI2 (3)
Since the dissolving process happens between 140 -180 ºC, the mechanism is that
CH3NH3I is actively reacting to other form, and thus accelerating the dissolving of perovskite.
Limited by a lack of information on how CH3NH3I further reacts, or behaves in high
temperatures, here we cannot exhibit the whole dissolving process. However, it can be concluded
that the CH3NH3PbI3 perovskite is dissolving at a specific temperature range.
As shown in the XRD results, there are characteristic peaks of PbCl2 appearing in the
XRD of samples annealed 260 ºC, 300 ºC and 400 ºC. However, there is no Cl source contacting
the samples after the deposition, and no Cl related structure has been observed in previous XRD
studies. As mentioned, these samples are actually annealed on hot plates in ambient conditions,
instead of in a vacuum oven. As such, one possible sources of the Cl is the dirt chemicals on the
surface of the hot plate.
The as-made samples were stored in a glove box that maintained a low humidity and
oxygen environment, to ensure that samples remained free from degradation. For measuring the
XRD patterns, a certain number of samples were taken out of the glove box and transferred into
the XRD test room. The samples were exposed to air during the transferring process, and all
pieces not annealed at once. During the annealing of the first sample, the rest of samples are also
exposed in air. Therefore, the annealed sample in the second order stayed in air for around 20
minutes. The influence of this time gap may affect the crystal structure since the perovskite is
very sensitive to moisture in the air. But this possible influence is not avoidable due to the
limitations of the experimental procedure.
This hypothesis would not be valid, however, if all of the PbI2 comes from the original
deposition, and not the dissolving process. In order to confirm whether the dissolving of the
perovskite produces more PbI2, quantitative examination needs to be carried out. A very easy and
straightforward method to understand the material change is to monitor the morphology of the
sample surface. Morphological quantitative information, such as coverage and material volume,
can thus be used as tools with which to evaluate quantitative information. We therefore used
42
SEM to further examine the morphological changes of the perovskite thin films, along with the
temperature increases.
43
3.3.3.2 SEM Results
XRD results show the influence of annealing temperatures on perovskite thin films. It can
be determined that the highest tolerant temperature of perovskite is about 140-180 ºC. A series of
crystal structural changes became clear from the XRD results. This would be helpful to predict
tolerant conditions of perovskite during the fabrication of electronic devices. Based on the crystal
structural changes observed in the XRD study, thin film morphology is very likely to shed some
light on the annealing temperature change. Combining knowledge of thermal dynamic behavior
with the morphological information, a solid prediction of the perovskite thin film thermal
dynamic property could be made. In the SEM study on the influence of annealing temperatures,
we were still using the same samples examined under XRD for section 3.3.2. We observed
perovskite thin films closely under SEM with different resolutions. In Fig. 3.5 we only show
selected images with present significant features on sample surfaces.
Figure 3.5a is the SEM image of the perovskite thin film without annealing treatment.
The thin film shows a full coverage on top of the substrate surface. Unlike other thin films made
by vapour deposition, our perovskite thin film does not show extremely uniform and flat film
surface; this is possibly because this thin film is a mixture of aimed perovskite and its
intermediate PbI2. For samples without annealing, we noticed that the film degrades rapidly in a
moist atmosphere, possibly due to the hygroscopicity of CH3NH3+ cation. We discovered 140 ºC
to be the ideal temperature for annealing, but the sample still suffered from reduction of surface
coverage at this temperature. Therefore, the as-made sample without annealing is the only thin
film that has a full coverage. For solution processed perovskite films [44], the film morphology
has different behavior from sample made in air and inert gas environment. Low temperature
annealing in an inert gas environment also prevents surface coverage reduction [3, 14].
The micrograph of 100 ºC annealed thin film is shown in Fig.3.5b. The thin film started
losing surface coverage, and a portion area of silicon substrate exposure can be observed. During
the heating process, many small pores formed rapidly, then either disappearing or joining
together to form a void area that can be observed from the image. In addition, the thin film
annealed at 100 ºC was homogenous, since the film was likely pieced together by crystal grains.
As XRD verified two mixed crystal structure, perovskite and PbI2, the film displayed in the SEM
44
image seems not clear enough to distinguish one crystal from the other. Nevertheless, it is
notable that the surface coverage is slightly dropped to 94% from the fully covered thin film.
Figure 3.5 SEM micrographs of perovskite samples: (a) Sample without annealing treatment; (b) Sample annealed at 100 ºC.
Upon heating the thin film to 140 ºC, the pore kept enlarging in size, and surface
coverage reduced to around 87% (Fig. 3.6c). Besides this main coverage change, the image
shows two types of crystal morphology: one kind is big crystal grains similar to those observed
45
in 100 ºC annealed sample; the other is the smaller-grain groups gathering at the gap area
between the big grains. According to the XRD results, this film still consists of two crystal
structures, perovskite CH3NH3PbI3 and PbI2, mixing together. The emergence of the smaller-
grain groups suggests that one kind of the crystal may not mix as well as in previous samples.
But further research is needed to determine and confirm which crystal is separating from the
mixture.
Figure 3.6 SEM micrographs of perovskite samples: (c) Sample annealed at 140 ºC; (d) Sample annealed at 180 ºC.
46
The thin film sample that annealed at 180 ºC is shown in Fig 3.6d. The materials lost half
of their surface coverage over the whole substrate surface, and 51 % of the area is still covered
with the remaining materials. The crystal morphology of this sample is small crystal grain groups
that look similar to those small crystal grain groups between big crystal grains (Fig 3.6c). If the
small grain groups observed in Fig. 3.6c are one pure crystal, then this morphology verifies it as
PbI2. However, it is hard to identify the chemical composition from the morphological
observation.
Figure 3.7 SEM micrographs of perovskite samples: (e) Sample annealed at 260 ºC; (f) Sample annealed at 300 ºC.
47
Surface coverage kept decreasing as annealing temperatures increases; the thin film
sample annealed at 260 ºC has a surface coverage of 44%. The surface morphology is mainly
small crystal particles with sizes around 1 μm (Fig 3.7e). XRD results support the idea of only
PbI2 crystal left in this sample.
At very high temperature, the material only remains spherical crystals with sizes smaller
than 500 nm. Shown in Fig 3.7f, the particles only cover 7 % of the surface. According to the
XRD result, the remaining crystal is mainly PbI2 at high annealing temperatures. Again, we were
not able to detect exact chemical composition, so more study is needed to address and explain
the particle composites. As shown in the images of samples annealed over 260 ºC, no significant
morphology appears as a new crystal form. So the PbCl2 cannot be assigned as any morphology
features.
Surface coverage is one of the parameters we can use to describe the thin film formation
over a surface area. To calculate the coverage, we enlarged the film, and divided part of the
image in to 20×20 small squares. By counting the squares covered with thin film, we can
calculate the thin film morphology coverage. Table 1 shows the detailed numbers of squares
filled with materials, and also the coverage percentage of thin films. The data were then plotted
in a diagram that shows the decreasing trend of the surface coverage (Fig. 3.8).
Combining both the XRD and SEM results, we determined that the annealing
temperature greatly affects the crystallinity and morphology of perovskite. The ideal annealing
temperature assists a thin film sample for better crystallinity. In XRD results, perovskite
completely disappeared beyond 180 ºC and the major component of the remaining film is PbI2.
Since the PbI2 has been presented in the film before the perovskite is gone, it is hard to conclude
if the PbI2 is from the film preparation or is from the dissolution of perovskite. From the SEM
results, even chemical components cannot be detected; here, a study of surface coverage sheds
light on the quantitative information. The surface still had more than 50% coverage at 180 ºC, so
that at least half of the thin film’s volume was PbI2. Along with temperature increases, most of
the perovskite disappears from the sample. Our study can be referenced for the future fabrication
of mixed perovskite annealing process, although interesting questions remain open for study.
.
48
Table 1: The surface coverage analysis of samples annealed at different temperatures.
Annealing
temperature
(ºC)
Surface area
(Counts of squares)
Occupied area
(Counts of squares)
Surface
coverage
(%)
Non-annealed 400 399 99.75
100 400 377 94.25
140 400 349 87.25
180 400 206 51.50
220 400 177 44.25
300 400 31 7.75
49
Figure 3.8 Perovskite surface coverage as a function of annealing temperatures.
3.3.4 Degradation Study of the Perovskite Thin Films (in Air)
Perovskite is known as an unstable material that can degrade in air [44]. During the
fabrication of a solar cell, the exposure of perovskite to air can hardly be avoided, so
understanding the change of the material is very important. Here we studied the degradation
behavior of perovskite sample by exposing it in air, and monitoring the changes in crystal
structure and morphology.
Figure 3.9 shows the XRD patterns of the same sample in two stages. The first stage is
right after annealing, and the second stage is after exposure in air for 14 days. The relative
humidity is around 30% on average. Fig. 3.9a shows the XRD pattern of the sample after
annealing at 140 ºC. The characteristic peaks whose position at 14.1º, 28.5º, and 43.2º
correspond to (110), (220) and (330) reflections belong to perovskite structure. However, peaks
of PbI2 are also shown at 14.1º, 27.0º and 39.0º, specifying that samples made from our
technique were mixture of CH3NH3PbI3 perovskite, and PbI2. After the XRD measurement, the
samples were kept in a container that was not airtight for 14 days, which allowed the sample to
50
gain some exposure to air, but would not be contaminated by dust. Degradation took place
during the 14-day process, and we measured the XRD pattern after 14 days, which is shown in
Fig 3.9b. Comparing the XRD pattern of 14-day degraded sample with the one just after
annealing, we found that three characteristic peaks of perovskite all completely disappeared. The
remaining peaks were all due to PbI2, indicating that perovskite degraded in air, and the
remaining material is PbI2 that exists in the film from beginning. Therefore, the result
demonstrates that perovskite CH3NH3PbI3 degrades in air spontaneously.
Figure 3.9 XRD patterns of perovskite thin films: (a) Perovskite film made by vapour deposition technique and annealed at 140 ºC; (b) The same sample of (a) exposed to air for
14 days.
51
According to the XRD results, the only degraded component of the annealed film was the
perovskite. Through observation of morphology changes during the process, it is possible to
identify the morphology composition of perovskite and other components of the thin film. In this
study, SEM was used to examine the morphology in the nano scale level. The information
provided by SEM is more straightforward than a structural analysis of XRD in determining
surface morphology. The SEM images of these two stages are as shown in Fig. 3.10. Both
images were taken immediately after the XRD measurement. At the first stage (Fig. 3.10a), the
majority of the image is the material-covered film, while only a small portion of area is the
substrate exposure. By analyzing the area of the covered surface and uncovered surface, we
determined that the thin film covered area is 87.3%. In this image, we observed that the thin film
is not totally uniform with a certain degree of coverage. Moreover, there are two kinds of
morphology features on the film. One morphological feature, circled with red on the left, has
solid thin film fully covered in a certain area. This feature is identical with those of samples that
had not been annealed. On the other hand, morphology as circled in green, on the right, has a
structure of connected small grains sitting at the edge of a big solid film. Correlated to the XRD
results, these two morphological features may strongly relate to crystal of perovskite and PbI2.
Figure 3.10b shows the second stage of the degradation process. It highlights that after exposure
in air for a 14-day period, the thin film coverage drops to 33.8%, which is less than half of its
original coverage. Also, the thin film morphology consists of small grain sized crystals with
disappearance of the bigger area solid thin film. This is indicative of the fact that the solid thin
film pieces degraded in air. We may be able to confirm that the perovskite structure that changed
in XRD and the bigger solid thin film in morphology are closely related. It is possible that the
disappeared solid crystal pieces are perovskite, and the remaining smaller sized grains are PbI2.
Based on the XRD and SEM results, we are able to calculate the degradation rate of
perovskite in air over time. Over 14 days, the surface coverage percentage dropped from 87.5%
to 33.8%, equivalent to a daily average drop of 5.37% on the surface area. We believe that the
degradation rate is related to the specific humidity condition and thus may vary each day.
Nevertheless, we conducted an efficient method to quantitatively evaluate the degradation rate of
perovskite thin film, which may contribute to the manufacturing control in large scale
production.
52
Figure 3.10 SEM images of perovskite films: (a) Perovskite film made by vapour deposition technique and annealed at 140 ºC; (b) The same sample from (a) exposure in air for 14
days.
53
3.3.5 High Resolution Study of Perovskite Thin Film with AFM
To investigate the surface features in high resolution, we used AFM to gain surface
information that explained the annealing process and growth mechanisms. AFM is a powerful
characterization method used to investigate surface morphology, and is designed for scanning on
small-scale surfaces with height distribution within its capable z-direction adjustment. The image
quality heavily relies on the sample surface conditions and the lateral limitation is very small
(100 μm × 100 μm). For our study, only suitable samples were selected to analyze surface
morphology with AFM.
As discussed in the XRD and SEM results sections (chapter 2), we only observed a surface
coverage loss of 12 % during the heating process, but the crystal morphology remained almost the
same. In addition, the film that showed grain size change in XRD cannot be characterized using SEM.
The detail morphological information on the crystal is studied by one of high resolution scanning
probe microscopy, AFM.
Figure 3.11 AFM images of perovskite films annealed at different temperature from 85 ºC to 140 ºC.
Here, AFM provided additional high-resolution information of surface features. Figure
3.11 is a collection of AFM images of perovskite, showing a significant small region (1 μm×1
μm) of perovskite surface, with different annealing temperatures. The samples annealed at 85 ºC,
100 ºC and 140ºC, as well as the sample that was not annealed, present morphological
differences. The non-annealed and 85 º C annealed samples are both fully covered by materials.
The thin film surface consists of uniform spherical grains with an average grain size of around
100 nm. The 100 ºC annealed sample shows a more flat shape in each grain, tending to have a
sheet-like shape. The 140 ºC annealed sample tends to increase this effect. The grains are even
flatter and exhibit a trend of amalgamating all grains into bigger crystal sheets. This phenomenon
54
indicates a recrystallization of the thin film during the annealing treatment. Grain sizes of the 140
ºC annealed sample become larger due to the recrystallization. This observation is consistent
with the XRD results. The recrystallization effects of annealing can only be observed in high
resolution AFM images, whereas SEM cannot differentiate morphological change at this scale.
We can also use AFM morphological images to study the growth process by comparing
images for films with different thickness, as shown in Fi. 3.12. The left panel shows the top-
down-view of the surface of a 10-nm thick sample, while the right panel depicts a 100 nm thin
film. The enlarged images (1 μm × 1 μm) are inserted for a detailed observation in high image
resolution. The image shows the grain sizes of the 10-nm thick sample range from 10 nm to 200
nm. At the same magnification, the surface of the 100-nm thick sample shows much more
uniform thin film structures, and all the grains have a sized of 50 nm (± 5 nm). The difference
between the two images indicates that the film formation is not always uniform along the
deposition process. The grain sizes change along the deposition, but eventually become uniform
when the film is deposited with enough thickness..
Figure 3.12 AFM images of perovskite films with10 nm thickness (left) and 100 nm thickness (right).
55
3.4 Summary
Organometal halide perovskite thin films were prepared using the vapour deposition
technique. First, this study optimized the conditions for perovskite preparation in a vapour
deposition system. The ideal deposition temperatures of PbCl2 and CH3NH3I are 125 °C and
330 °C, respectively, and the ratio of the organic source to the inorganic source is 8:1. In addition,
a standard for quickly checking the crystal film by its colour assisted this study in deciding if a
batch of deposition was successful. The XRD shows perovskite contained in most of our samples,
even though PbI2 was mixed in with the film. The XRD results compare the effects of different
annealing temperatures, and show that perovskite is a thermally unstable material. The films
demonstrated lack of tolerance for high temperature, and the characteristic peaks of CH3NH3PbI3
perovskite in the XRD disappeared after 180 °C annealing. The XRD results also support the
finding that the remaining part of the thin film is PbI2. Further observation of SEM images
compared the morphological change along with annealing temperature increases. Here, we found
that the surface coverage of the thin film decreased with the increase in annealing temperature.
From the results of both XRD and SEM. we concluded that perovskite was decomposed while
only PbI2 remained. We also identified that PbCl2 evaporates from the film in high temperatures.
Air exposure investigation illustrated that the perovskite was unstable when exposed to
air for long periods of time, and also only PbI2 remained after degradation. This observation
confirmed that perovskite is very sensitive to the ambient environment, and the mixed thin film
of perovskite and PbCl2 usually resulted only in PbCl2. Naturally, SEM has limitations on
observing the surface morphology in a high resolution, so AFM was used to watch closely of the
thin film surface. AFM showed detailed morphological change in high resolution.
56
Chapter 4 Conclusions and Future Work
In this study, organometal halide CH3NH3PbI3 thin films for perovskite solar cells were
prepared using the vapour deposition technique. The deposition temperature and the ratio of two
source chemicals were optimized. The resultant perovskite films still contained a certain portion
of PbI2, which indicates that PbI2 is the intermediate of the chemical reaction. The presence of
the PbI2 intermediate also verified the proposed chemical reactions for perovskite formation. The
study also discovered that controlling the deposition rate and ratio of chemical sources are key
factors in preparing uniform perovskite thin film.
The tolerance temperature of annealing treatment was examined, and an ideal
temperature for the annealing process was decided to be around 140 ºC. A series of perovskite
samples were prepared and annealed at a wide range of temperatures (100 - 400 °C). Results
showed that the perovskite has a low tolerance, and the thin film was not very stable upon
heating to high temperatures. Annealing at 180 °C eliminated all perovskite in the thin film, and
only PbI2 remained. At very high annealing temperatures, PbCl2 was observed, but it was very
hard to explain the source of the Cl. One possible reason is that the hot plate was not clean from
a previous use.
In terms of morphological control, SEM micrographs of non-annealed perovskite showed
the full coverage of perovskite thin film. It indicated that the vapour deposition technique is ideal
for making perovskite uniform thickness, and fully covered thin film. The surface kept losing the
material coverage with the increase of the annealing temperature. Prior 180 °C, it was observed
that perovskite dissolved in the temperature range of 140 – 180 °C. The remaining PbCl2 covered
more than half of the surface, even over 180 °C.
The stability and durability of perovskite thin films was studied by exposing the sample
to air for a relatively long period of time. The observations demonstrated the transformation of
perovskite to PbI2 intermediate. The use of AFM technique allowed for the detailed observation
on surface features, so the annealing effect that was unobtainable in SEM became apparent.
AFM results also showed high-resolution details, which indicated that the film gets more
uniform in thicker thin film samples.
57
There are still a number of topics worth investigation, beyond the results and findings of
this study. Regarding the control of perovskite thin film preparation, further optimization of
preparation conditions may offer perovskites with purer and higher crystallinity. Thus, the film
can get larger crystal grains, and a higher purity of perovskite. Since the crystal quality is a key
factor for photovoltaic performance, it would be helpful to obtain a device with higher
performance capabilities. Moreover, further studies of morphological controls are crucial for
optimizing the manufacturing procedure. In terms of the morphological effect in a real solar cell
device, a study of the interfaces of different functional layers in photovoltaic device may be of
great interest and importance.
Advanced microscopic technology will assist in understanding a series of questions about
surface and interface of thin film devices. By displaying the high-resolution morphology in ultra-
thin films solar cells, a number of unrevealed surface features can be captured. These features
would greatly benefit on explaining thin film properties and the device performance.
58
References:
1. Wen, W.-T.D., Hybrid Organic/Inorganic Solar Cells based on Electrodeposited ZnO Nanowire Arrays on ITO and AZO Cathodes. 2011, University of Toronto.
2. Dennler, G., N.S. Sariciftci, and C.J. Brabec, Conjugated polymer-based organic solar cells. Technology, 2006. 11: p. 2.
3. Hodes, G., Perovskite-Based Solar Cells. Science, 2013. 342(6156): p. 317-318.
4. Snaith, H.J., Perovskites: The Emergence of a New Era for Low-Cost, High-Efficiency Solar Cells. The Journal of Physical Chemistry Letters, 2013. 4(21): p. 3623-3630.
5. Goldschmidt, V.M., Die gesetze der krystallochemie. Naturwissenschaften, 1926. 14(21): p. 477-485.
6. Wenk, H.-R. and A. Bulakh, Minerals: their constitution and origin. 2004: Cambridge University Press.
7. Tejuca, L.G. and J.L. Fierro, Properties and applications of perovskite-type oxides. 2000: CRC Press.
8. Lemanov, V., et al., Perovskite CaTiO3 as an incipient ferroelectric. Solid state communications, 1999. 110(11): p. 611-614.
9. Luxová, J., P. Šulcová, and M. Trojan, Study of perovskite compounds. Journal of thermal analysis and calorimetry, 2008. 93(3): p. 823-827.
10. Lufaso, M.W. and P.M. Woodward, Jahn-Teller distortions, cation ordering and octahedral tilting in perovskites. Acta Crystallographica Section B, 2004. 60(1): p. 10-20.
11. Chen, Q., et al., Planar heterojunction perovskite solar cells via vapor-assisted solution process. Journal of the American Chemical Society, 2013. 136(2): p. 622-625.
12. Park, N.-G., Organometal perovskite light absorbers toward a 20% efficiency low-cost solid-state mesoscopic solar cell. The Journal of Physical Chemistry Letters, 2013. 4(15): p. 2423-2429.
13. Hao, F., et al., Lead-free solid-state organic-inorganic halide perovskite solar cells. Nature Photonics, 2014.
14. Burschka, J., et al., Sequential deposition as a route to high-performance perovskite-sensitized solar cells. Nature, 2013. 499(7458): p. 316-319.
15. Lufaso, M.W. and P.M. Woodward, Jahn-Teller distortions, cation ordering and octahedral tilting in perovskites. Acta Crystallographica Section B: Structural Science, 2004. 60(1): p. 10-20.
59
16. Johnsson, M. and P. Lemmens, Perovskites and thin films—crystallography and chemistry. Journal of Physics: Condensed Matter, 2008. 20(26): p. 264001.
17. Jahn, H.A. and E. Teller, Stability of polyatomic molecules in degenerate electronic states. I. Orbital degeneracy. Proceedings of the Royal Society of London. Series A, Mathematical and Physical Sciences, 1937: p. 220-235.
18. O’regan, B. and M. Grfitzeli, A low-cost, high-efficiency solar cell based on dye-sensitized. nature, 1991. 353: p. 737-740.
19. Lee, M.M., et al., Efficient hybrid solar cells based on meso-superstructured organometal halide perovskites. Science, 2012. 338(6107): p. 643-647.
20. Chen, G., X. Liu, and C. Su, Distinct effects of humic acid on transport and retention of TiO2 rutile nanoparticles in saturated sand columns. Environmental science & technology, 2012. 46(13): p. 7142-7150.
21. Bian, Z., et al., Superior electron transport and photocatalytic abilities of metal-nanoparticle-loaded TiO2 superstructures. The Journal of Physical Chemistry C, 2012. 116(48): p. 25444-25453.
22. Kalyanasundaram, K., Dye-sensitized solar cells. 2010: EPFL press.
23. Zeng, W., et al., Efficient dye-sensitized solar cells with an organic photosensitizer featuring orderly conjugated ethylenedioxythiophene and dithienosilole blocks. Chemistry of Materials, 2010. 22(5): p. 1915-1925.
24. Chung, I., et al., All-solid-state dye-sensitized solar cells with high efficiency. Nature, 2012. 485(7399): p. 486-489.
25. Feldt, S.M., et al., Design of organic dyes and cobalt polypyridine redox mediators for high-efficiency dye-sensitized solar cells. Journal of the American Chemical Society, 2010. 132(46): p. 16714-16724.
26. Ning, Z., Y. Fu, and H. Tian, Improvement of dye-sensitized solar cells: what we know and what we need to know. Energy & Environmental Science, 2010. 3(9): p. 1170-1181.
27. Rühle, S., M. Shalom, and A. Zaban, Quantum ‐ dot ‐ sensitized solar cells. ChemPhysChem, 2010. 11(11): p. 2290-2304.
28. Zhang, Q., et al., Highly efficient CdS/CdSe-sensitized solar cells controlled by the structural properties of compact porous TiO2 photoelectrodes. Physical Chemistry Chemical Physics, 2011. 13(10): p. 4659-4667.
29. Barea, E.M., et al., Design of injection and recombination in quantum dot sensitized solar cells. Journal of the American Chemical Society, 2010. 132(19): p. 6834-6839.
30. Tang, J., et al., New starburst sensitizer with carbazole antennas for efficient and stable dye-sensitized solar cells. Energy & Environmental Science, 2010. 3(11): p. 1736-1745.
60
31. Burschka, J., et al., Tris (2-(1 H-pyrazol-1-yl) pyridine) cobalt (III) as p-type dopant for organic semiconductors and its application in highly efficient solid-state dye-sensitized solar cells. Journal of the American Chemical Society, 2011. 133(45): p. 18042-18045.
32. Mitzi, D., et al., Conducting layered organic-inorganic halides containing< 110>-oriented perovskite sheets. Science, 1995. 267(5203): p. 1473-1476.
33. Kojima, A., et al., Organometal halide perovskites as visible-light sensitizers for photovoltaic cells. Journal of the American Chemical Society, 2009. 131(17): p. 6050-6051.
34. Im, J.-H., et al., 6.5% efficient perovskite quantum-dot-sensitized solar cell. Nanoscale, 2011. 3(10): p. 4088-4093.
35. Kim, H.-S., et al., Lead iodide perovskite sensitized all-solid-state submicron thin film mesoscopic solar cell with efficiency exceeding 9%. Scientific reports, 2012. 2.
36. Heo, J.H., et al., Efficient inorganic-organic hybrid heterojunction solar cells containing perovskite compound and polymeric hole conductors. Nature photonics, 2013. 7(6): p. 486-491.
37. Noh, J.H., et al., Chemical management for colorful, efficient, and stable inorganic–organic hybrid nanostructured solar cells. Nano letters, 2013. 13(4): p. 1764-1769.
38. Service, R., Perovskite Solar Cells Keep On Surging. Science (New York, NY), 2014. 344(6183): p. 458.
39. Tsujimoto, K., et al., TiO2 surface treatment effects by Mg2+, Ba2+, and Al3+ on Sb2S3 extremely thin absorber solar cells. The Journal of Physical Chemistry C, 2012. 116(25): p. 13465-13471.
40. Qiu, Y., W. Chen, and S. Yang, Facile hydrothermal preparation of hierarchically assembled, porous single-crystalline ZnO nanoplates and their application in dye-sensitized solar cells. Journal of Materials Chemistry, 2010. 20(5): p. 1001-1006.
41. Mei, A., et al., A hole-conductor–free, fully printable mesoscopic perovskite solar cell with high stability. science, 2014. 345(6194): p. 295-298.
42. Hardin, B.E., H.J. Snaith, and M.D. McGehee, The renaissance of dye-sensitized solar cells. Nature Photonics, 2012. 6(3): p. 162-169.
43. Liu, M., M.B. Johnston, and H.J. Snaith, Efficient planar heterojunction perovskite solar cells by vapour deposition. Nature, 2013. 501(7467): p. 395-398.
44. Eperon, G.E., et al., Morphological Control for High Performance, Solution‐Processed Planar Heterojunction Perovskite Solar Cells. Advanced Functional Materials, 2014. 24(1): p. 151-157.
61
45. Bragg, W.L., The structure of some crystals as indicated by their diffraction of X-rays. Proceedings of the Royal Society of London. Series A, 1913. 89(610): p. 248-277.
46. Dann, S.E., Reactions and Characterization of Solids. 2002: Royal Society of Chemistry, USA.
47. Lennard-Jones, J.E., Proc. R. Soc. Lond. A, 1924. 106: p. 463.
48. Qin, P., et al., Inorganic hole conductor-based lead halide perovskite solar cells with 12.4% conversion efficiency. Nature communications, 2014. 5.
49. Grätzel, M., The light and shade of perovskite solar cells. Nature materials, 2014. 13(9): p. 838-842.
50. Zhou, H., et al., Interface engineering of highly efficient perovskite solar cells. Science, 2014. 345(6196): p. 542-546.
51. Green, M.A., A. Ho-Baillie, and H.J. Snaith, The emergence of perovskite solar cells. Nature Photonics, 2014. 8(7): p. 506-514.