three-dimensional transparent conducting oxide based dye … · 2013-10-31 · ii three-dimensional...
TRANSCRIPT
Three-Dimensional Transparent Conducting Oxide Based
Dye Sensitized Solar Cells
By
Eric Arsenault
A thesis submitted in conformity of the requirements
For the degree of Masters in Science
Graduate Department of Chemistry
University of Toronto
© Copyright by Eric Arsenault 2011
ii
Three-Dimensional Transparent Conducting Oxide Based
Dye Sensitized Solar Cells
Eric Arsenault
Masters of Science, Department of Chemistry
University of Toronto
2011
Abstract
Electron transport and recombination are two competing factors within Dye-Sensitized Solar-
Cells (DSSCs) which have a great influence on their performance. By drastically increasing the
speed of electron transport to the electrode, it is believed that these cells could reach new record
efficiencies.
To achieve this result, an all-in-one integrated DSSC was attempted, in which the electrode
material is extended into the active area of the solar cell material. The research conducted can be
separated into two stages. The first stage is the production of a three-dimensional macroporous
electrode. The second stage is the production of an all-in-one DSSC by a simplified co-casting
technique. The structures and materials presented were examined using electron microscopy, X-
ray Diffraction, 4-point and 2-point probe electrical measurements as well as experimentally by
the testing of solar cells. The methods of fabrication, characterization, experimental results and
future directions are also presented.
iii
Acknowledgements
I would like to thank you Geoff for the short but sweet experience in your lab. Your energy,
passion and creativity inspire me as well as others in the group. I truly enjoyed brainstorming
with you and the liberty you gave me to pursue my own interests. Your constant support and
guidance during the most challenging times of this project were much appreciated.
I would also like to thank our collaborators in Switzerland for making this project a reality.
Nicholas Tetrault and Jeremy Brillet, thank you for your help. Micheal Gratzel, thank you for
the opportunity to work in your lab, it was a great learning experience as well as an unbelievable
opportunity.
Lastly, I would like to thank our lab – Sue, thank you for your help with everything – Navid and
Bryan, it was great working with you on this challenging project – Engelbert, Danny, Eric,
Jordan, Jenny, Wendong, Laura, Melanie, Makoto, Jeff, Peter, Jon, etc... you guys are great!
iv
Table of Contents
1 Introduction to Dye-Sensitized Solar cells, DSSCs ........................... 1
1.1 DSSC History ...............................................................................................................1
1.2 Recent Developments in DSSCs....................................................................................7
1.3 Cell Fabrication........................................................................................................... 13
1.4 Operation .................................................................................................................... 17
2 Transparent Conducting Oxides and DSSCs ................................... 24
2.1 What are TCOs ........................................................................................................... 24
2.2 TCO extension into the 3rd
Dimension......................................................................... 25
2.3 Why and how to extend the TCO ................................................................................ 27
3 3D Transparent Conducting Oxide frameworks .............................. 32
3.1 Strategies .................................................................................................................... 32
3.2 Templated Co-casting ................................................................................................. 33
3.3 Opal Templating ......................................................................................................... 34
3.4 Shortcomings .............................................................................................................. 44
4 Simplified procedure for an all-in-one DSSC ................................. 50
4.1 Basic concept .............................................................................................................. 50
4.2 Emulsions ................................................................................................................... 53
4.3 Preventing shorting ..................................................................................................... 57
4.4 Results and Discussion ................................................................................................ 61
5 Conclusion ...................................................................................... 68
v
Table of Figures
Figure 1-1: Relative costs of various energy sources . ................................................................ 2
Figure 1-2: Schematic of a p-n junction solar cell ...................................................................... 3
Figure 1-3: Band diagram for early stage, single crystal DSSC .................................................. 4
Figure 1-4: DSSC cross-section ................................................................................................. 5
Figure 1-5: Energy diagram of a DSSC depicting the lifetimes of the various processes ............ 6
Figure 1-6: Z-907 dye ............................................................................................................... 8
Figure 1-7: Progress in Ruthenium based dyes and Organic D-A dyes ....................................... 9
Figure 1-8: Cross-section of a DSSC with top scattering layer ..................................................10
Figure 1-9: Titania nanotube array ...........................................................................................12
Figure 1-10: Core-Shell DSSCs with SnO2 cores (left) and TiO2 cores (right) .........................13
Figure 1-11: Diagram of the Doctor Blading technique ............................................................14
Figure 1-12: SEM image of a titania nanoparticle film .............................................................14
Figure 1-13: Components of a solar cell ...................................................................................15
Figure 1-14: Cross-sectional view of DSSC cell assembly ........................................................17
Figure 1-15: Solar Simulator setup ...........................................................................................17
Figure 1-16: Typical experimental results obtained from a solar simulator ...............................18
Figure 1-17: Equivalent circuit of a DSSC................................................................................19
Figure 1-18: Effect of RS and RSH on the shape of an I-V curve ................................................19
Figure 2-1: All-in-one DSSC initially presented by Zaban et al. ...............................................26
Figure 2-2: ITO/TiO2 core/shell brush structure .......................................................................26
Figure 2-3: Continuity equation and solution for a DSSC under short circuit conditions ...........28
Figure 3-1: 3D TCO via templated and non-templated approaches ...........................................32
Figure 3-2: Optimized TCO framework produced via a co-casting template approach ..............33
Figure 3-3: STEM image of ATO nanocrystals ........................................................................36
Figure 3-4: Self-assembly steps for i-ATO-o ............................................................................37
Figure 3-5: Structure and optical reflectivity spectra of i-ncATO-o films..................................38
Figure 3-6: Crystal structure of the ATO nanocrystals (a) and sol (b) .......................................38
Figure 3-7: Resistance and porosity measurements of the ATO materials .................................39
Figure 3-8: Diagram of the improved interconnectivity before and after ATOsol ......................40
Figure 3-9: Periodic macroporous transparent oxide i-ncATO-o electrode ................................40
vi
Figure 3-10: SEM image of an inverse ATO opal with bismuth electrodeposited within. ..........41
Figure 3-11: XPS atomic percentages before and after electrodeposition of Bi on the i-o ..........42
Figure 3-12: A “floating” i-o showing the poor adhesion to the substrate .................................44
Figure 3-13: Inverse opal with an overlayer ..............................................................................45
Figure 3-14: Poor interconnectivity of the i-o framework .........................................................46
Figure 3-15: Necking of opal template resulting in a covered substrate .....................................46
Figure 3-16: Infiltration of the TCO i-o with TiO2 nanoparticles ..............................................47
Figure 4-1: 3D TCO via a two-component approach ................................................................50
Figure 4-2: Film with a 2:1 ratio of FTO nanoparticles (light) to TiO2 spheres (dark) ...............51
Figure 4-3: Co-cast TCO/titania sphere film with and without a blocking layer ........................51
Figure 4-4: SEM image of nanoparticle spheres .......................................................................55
Figure 4-5: Size distribution of the nanoparticle spheres ...........................................................55
Figure 4-6: Pore size distribution as determined by BET ..........................................................56
Figure 4-7: SEM Image of FTO nanoparticles (light) and a TiO2 nanoparticle sphere (dark) ....56
Figure 4-8: PXRD of FTO nanoparticles ..................................................................................57
Figure 4-9: TiCl4 ALD system .................................................................................................58
Figure 4-10: Image of TiCl4 solution before and after 25 minutes at 100˚C ..............................59
Figure 4-11: Resistance of FTO substrates as a function of TiCl4 treatment time ......................59
Figure 4-12: Method used to perform the conductivity measurements ......................................59
Figure 4-13: PXRD of the TiO2 powder, spheres and TiCl4 treated spheres ..............................60
Figure 4-14: TiO2 nanocrystals resulting from TiCl4 treatment .................................................60
Figure 4-15: Sphere film after TiCl4 treatment..........................................................................61
Figure 4-16: Efficiency, Fill Factor, Short Circuit Current and Open Circuit Voltage for TiO2
sphere films..............................................................................................................................61
Figure 4-17: Efficiency, Fill Factor, Short Circuit Current and Open Circuit Voltage for
FTO/TiO2 sphere films .............................................................................................................62
Figure 4-18: Efficiency of FTO-TiO2 nanoparticle-sphere DSSC with post-treatments of ALD
and TiO2 sol .............................................................................................................................63
Figure 4-19: IV curves of the TiO2 spheres (solid) and the FTO/TiO2 films (dashed) before and
after TiCl4 treatment .................................................................................................................64
1
1 Introduction to Dye-Sensitized Solar cells, DSSCs
1.1 DSSC History
The Sun has fascinated humans since the beginning of time. It has literally shaped our world
into what it is; from the evolution of life to religion, culture and geography. It is our clock, our
morning breeze, our fuel. Once our only form of energy, the sun has played an increasingly
small part of our energy needs since the discovery and addition to cheap and easy sources of
fuel. Not only are these sources of fuel beginning to run low, but the effects of global warming,
air, water and land pollution are finally beginning to catch up to us.1 We are currently faced with
one of the largest endeavours ever taken on by man – converting from cheap, abundant,
polluting energy sources to cleaner, more expensive ones. Once again we are looking up to the
Sun. On a yearly basis, there is approximately 3x1024
Joules of energy hitting the earth – about
10 000 times more than our current energy needs.2 Converting solar energy into electricity
might seem like an easy solution to our problems – but significant subsidies and investments
into these technologies are currently required to make them competitive due to the abundance
and cheap price of non-renewable energy (see Figure 1-1). Thus, many in the field believe
reducing the costs associated with producing solar cells (or alternatively increasing their
efficiency and maintaining their costs) is the key to abundant and clean energy. Although there
are a number of photovoltaic technologies, on a cost/watt basis few technologies compare to
Dye Sensitized Solar cell (DSSCs).3 The cost/watt of these cells is believed to be below 1$/watt,
although there have not been any confirmed (in comparison, the most efficient CdTe cells are
0.76$/Watt).4 Due to the cheap starting materials of these cells as well as their ease of
fabrication, DSSCs have generated significant interest in the scientific community as well as the
general public.5 Although progress has been made on DSSCs, for the past 20 years, the gains in
efficiencies have not been spectacular due to some fundamental issues yet to be addressed
including increasing the absorption spectra of the dyes, collecting electrons faster and reducing
recombination. This work was performed with the goal of tackling one of these problems:
collecting electron in a faster, more efficient way, while preventing faster recombination.
2
.
Figure 1-1: Relative costs of various energy sources 3
Solar energy and the ways in which it interacts with matter have fascinated scientists for
centuries. Ever since Henri Becquerel‟s discovery of the photovoltaic effect in 1839, researchers
have dreamt of creating devices able to capture solar energy to produce electricity or fuels.6 The
history of “dye sensitization” can be traced back to the early days of photography. In 1837,
Daguerre created the first photographic film and the first metal halide photographic film was
created shortly after in 1839 by Fox Talbot. Back in these early days, photographic films were
quite insensitive to light due to the semiconductor nature of the silver halide grains, resulting in
negligible light absorption past 460 nm. In 1883, Vogel discovered that silver halides could be
sensitized by the use of dyes – this extended the photosensitization to longer wavelengths, thus
enabling “dye sensitized” photographs. On the other hand, photoelectrochemistry began with
Becquerel‟s observation of current between two electrodes separated by a metal halide solution
upon illumination. In 1887, the “sensitization” concept was borrowed from photography and
extended to photoelectrochemistry, thus allowing much greater currents to be observed between
the two electrodes. The rest, as they say, is history! All the tools were at the disposal of chemists
to improve and perfect this method of solar energy conversion. The concept was simple: use a
dye to absorb light, and pass the excited electron into a circuit before it recombines. For the next
nearly 100 years, progress in DSSCs was slow and overshadowed by the efficiency and
understanding of Silicon solar cells. Although these initial research attempts to produce
functional DSSCs were very different than the solar cells produced today, they formed the
foundation of DSSC research. The evolution and creation of DSSCs as we know them really
began to change after the creation of Silicon solar cells which boomed in popularity in the late
1950s.7
3
The primary goal of photovoltaics is to absorb incoming light and separate the electron-hole pair
which is generated, facilitated by a junction between two different materials. In silicon solar
cells, a junction is formed at the interface of two different forms of doped silicon generally
referred to as a p-doped region and an n-doped region caused by dopant atoms such as boron
and phosphorous.8 At their interface, an electric field is generated causing an electric potential
difference at their interface and resulting in the separation of holes and electrons upon light
absorption (Figure 1-2).
Figure 1-2: Schematic of a p-n junction solar cell
This newly acquired knowledge from silicon solar cells inspired researchers, and soon enough
dyes started appearing in solar cells, beginning with the work of J. Mulder and J. De Jonge in
1963.7 Although the operating mechanism was not fully understood, this was the first attempt to
construct solar cells using semiconductors such as titania or zinc oxide with dyes. At this early
stage of research, the semiconductors used were single crystal and the dyes were free in solution
with only a small fraction of these dyes actually bound to the surface. The photocurrent
observed was due to dyes in close proximity to the semiconductor absorbing incoming light and
transferring the excited electron to the semiconductor‟s conduction band. In this work, the dyes
used were Rose Bengal and Flurescein and a ferric/ferrous redox shuttle as a hole transporter.
Figure 1-3 depicts the operating mechanism within the cell, as understood by the researchers at
the time. The diagram is surprisingly similar to what we now use to explain DSSCs. Although
these solar cells did produce current, they remained extremely inefficient due to the low surface
area of the semiconductor, resulting in light absorption of less than 1%. It was believed that a
4
single crystal semiconductor was required to produce the electric field required to separate
charge (a concept likely borrowed from silicon solar cells).7
Figure 1-3: Band diagram for early stage, single crystal DSSC7
As this field progressed, slowly, clues began emerging as to how to improve the efficiency of
the cells. One successful approach was to physically anchor the dyes to the semiconductor
surface.9 Another strategy was to increase the surface area of the semiconductor by texturing
them, resulting in a 1.5% monochromatic efficiency.10
The real breakthrough in the field came in 1991 by the work of Brian O‟Reagan and Micheal
Gratzel published in their paramount paper entitled „A low-cost high-efficiency solar cell based
on dye-sensitized colloidal TiO2 films‟.11
This research truly differentiated itself from previous
efforts mainly by using nanoparticles for the semiconductor phase as well the use of optimized
ruthenium dyes and electrolytes previously published by the group. The main components are
presented in Figure 1-4. Once light is absorbed by the dyes, electrons are transferred and
transported by the valence band of titania. Electrons are transported through the titania
nanoparticle matrix to the front transparent conducting oxide (TCO) electrode and passed
through a circuit connected to the counter electrode while producing work. Platinum is used as
a catalyst to transfer the electron from the counter electrode to the redox shuttle. This redox
shuttle cycles between the counter electrode and the dyes, allowing dyes to inject electrons over
and over again.
5
Figure 1-4: DSSC cross-section4
One of the most puzzling phenomena presented by the Gratzel paper was the fact that electrons
could travel through the TiO2 nanoparticle framework without suffering from severe
recombination. In addition, the small particle size implied there was no electric field within the
cell.12
Although the high current density could be explained by the high surface area of the
nanoparticles, the high photovoltage was not expected and came as a surprise to most –
especially in light of the fact that electrons needed to travel through insulating particles without
suffering from severe recombination! The search was on to discover why and how these cells
work so effectively, with the promise to produce the cheapest and most efficient solar cell in the
world. This gigantic jump in DSSC efficiency (from 1.5% monochromatic efficiency to over
9%!) spurred intense interest in this new frontier. Although the understanding of DSSCs has
greatly expanded, the efficiency improvements have been mediocre, increasing to just over 11%
in nearly 20 years. Today, we have a basic understanding of the limiting factors affecting
DSSCs performance and an idea of what can be done to tackle these issues.
Electron collection mechanisms within DSSCs is still a topic of intense debate, and is simply
described as the “random walk model” wherein electrons diffuse in a convoluted, non-directed
manner through the titania nanoparticle matrix.13, 14
A number of techniques have been
developed to probe and better understand the various processes involved in these cells.
Examples of these measurements including frequency modulated and time modulated
techniques such as intensity modulated photocurrent/photovoltage spectroscopy, electrical
impedance spectroscopy as well as transient photovoltage and photocurrent measurements.
Impedance spectroscopy has been a key tool in this understanding and has allowed researchers
6
to probe various events in the solar cell independently, allowing scientists to identify the
lifetimes of various processes within the cells.15
It is no surprise that the key to the success of
DSSCs lies in the favourable lifetimes of the various processes involved. These lifetimes include
the injection lifetime, transport lifetime, recombination lifetime and excitation lifetime. The
approximate magnitudes of these processes are shown below in Figure 1-5. As can be seen,
electron excitation, electron injection from the excited dye to the titania as well as dye
regeneration are fairly fast processes. Transport and recombination are two processes which
have lifetimes within one order of magnitude apart and thus there is the potential to improve
DSSCs significantly is one can increase transport while preventing recombination.
Recombination refers to the process in which electrons previously injected into the TiO2
nanoparticles are transferred back to the oxidized redox shuttle or dye. It truly is remarkable that
electrons, separated by a nanometer of dye can travel through the titania nanoparticles without
recombination for such long periods of time. As will be discussed later on, one of the key
reasons for the slow recombination is due to the relatively non-aggressive nature of the redox
shuttle.16
The aim of this thesis is to increase the electron transport speed within DSSCs while
simultaneously decreasing the recombination by the extension of the electrode into the active
area of the DSSC.
Figure 1-5: Energy diagram of a DSSC depicting the lifetimes of the various processes18
7
1.2 Recent Developments in DSSCs
Since their discovery, progress in DSSCs can be divided into five categories: sensitizers, counter
electrodes, electrolytes, semiconductors and structural changes. Improvements in these
components has been performed to address specific aspects of DSSCs such as light absorption,
electron and hole transport, sealing and stability issues as well as preventing recombination. In
the following section, I will briefly describe some trends in these categories, followed by a more
detailed analysis of structural changes in DSSCs.
Counter electrode progress has been mainly driven by the replacement of platinum. A key
selling point of DSSCs is their low cost. At the current price of 1810$/oz, platinum substitution
is an easy way to produce more economical solar cells. Recent examples of substitutes include
carbon counter electrodes, metal counter electrodes as well as CoS counter electrodes.17
One of
the advantages of platinum is that it can be applied in very thin films, allowing transparent
counter electrodes to be made – a valuable feature which may someday be used to allow for
tandem cells or solar harvesting windows.
Electrolyte solutions have for the most part remained quite unchanged over the past decade.
Most purchased electrolytes are proprietary and there are few reports of electrolyte
optimization.18
Electrolytes are usually customized according to the valence band and
conduction band of the sensitizers used and thus as new dyes keep emerging into the market, so
do the electrolyte mixtures. Electrolyte solutions include Iodide and Iodine in Acetonitrile as
well as additives such as lithium or pyridines which have been shown to prevent recombination
as well as modify the redox potential of the electrolyte solution.18
A typical electrolyte solution
consists of 0.6M tetrabutylammonium iodide, 0.1M Lithium Iodide, 0.5M 4-tert-butylpyridine
and 100mM Iodine in Acetonitrile.18
Due to the volatile nature of the Acetonitrile electrolyte,
there has been a push to use alternative electrolyte solutions including ionic liquids, eutectic
melts, water and solid state charge transporters.19,20,21
Although ionic liquids have reached 7%
efficiency, they have been sown to be unstable under prolonged thermal stress and light
soaking.22
Eutectic melts have recently been shown to reach 8.2% by the Gratzel group which
has sparked a flurry of interest in this field.19
Sensitizers are a critical component of DSSCs and it is no surprise that they have been a subject
of intense interest. Interest in sensitizers is three-fold; to increase the absorption range, increase
8
the stability and decrease recombination of the sensitizers. In recent years, ruthenium(II)–
polypyridyl complexes, metal-free organic Donor–Acceptor (D–A) dyes, as well as
semiconductor nanoparticles have been used as sensitizers.23
Ruthenium dyes are characterized
by their high efficiency, expensive cost as well as their difficult purification. These molecules
have a general formula of RuLxL-ySCNz, where L and L
- are polypyridyl ligands although the
dyes are generally abbreviated by simple names, traditionally beginning with the initials of its
creator (Z-907 for example was developed by Zakeeruddin from the Gratzel lab, seen below).
Figure 1-6: Z-907 dye
Donor–Acceptor dyes on the other hand can be prepared rather inexpensively and can be
designed to absorb through the visible spectrum through molecular design. Progress in organic
dyes have far surpassed ruthenium sensitizers and could at some point in the near future surpass
them in terms of efficiency (see Figure 1-7).
9
Figure 1-7: Progress in Ruthenium based dyes and Organic D-A dyes4
Due to the high price of Ruthenium, other metal centers such as Copper and Zinc porphyrins are
also gaining interest.24
Pendant ligands play a huge role in the efficiency due to their affect on
the anchoring efficiency of the dyes, their stability in solution and on the semiconductor surface
(preventing aggregation) as well as playing a big role in reducing recombination. Traditionally
carboxylic groups have been used to anchor dyes to the surface but it has been recently shown
that phosphonic acid groups might be better suited for the role.25
The use of hydrophobic side
chains like that of Z-907 increases the long term stability of the dyes in solution as well as on
the surface of titania which is partly due to the prevention of water adsorption on the titania
surface.26
The use of electron donating groups on the pendant ligands such as ether groups or
thiocyanate groups has been shown to prevent recombination.27
The grand challenge as well as
the holy grail of this field is to extend the absorption spectra such that more light and thus more
current is generated. Efforts to out-do the “black dye” first produced in 1997 have thus far
resulted in only slightly better improvements.28
Recently, groups have begun anchoring
thiophene moieties to the dyes which has been shown to increase their extinction coefficient and
absorption spectra to the red.29
Although semiconductor nanocrystals have found quite a bit of
interest, complications associated with long term stability, low efficiencies and anchoring have
resulted in short-lived fame.30
10
Semiconductors used in DSSCs have mainly included TiO2, ZnO2, SnO2, Nb2O5 or
chalcogenides such as CdSe.31, 32
Although various semiconductors have been attempted, titania
is still the material of choice due to its relative inertness and ideal electronic structure. Zinc
oxide, the runner-up, has poor long term stability in DSSCs and thus although they have been
quite efficient in certain cells, are not as widely used.33
Conduction in SnO2 is much faster than
in TiO2 and impedance measurements suggest that transport and recombination are 2-3 orders of
magnitude higher in SnO2 resulting in low fill factors and voltages.34
Structural modifications in DSSCs have played a small but increasingly important role in
devices. Shortly after the first DSSCs, scattering layers were introduced to improve the light
absorption of the dye within the active region of the cell.35
This layer typically consists of a 5
micron film of 100-800 nm titania particles, layered above active titania film of 15-25 nm
nanoparticles (see Figure 1-8). The scattering layer reflects nearly 90% of incoming light in all
directions, allowing multiple passes through the active region of the solar cell as well as
increasing the residence time of light within the cell through low-angle reflection.36
This
technique has produced current densities with over 80% improvements.37
The resulting cell is
not transparent and thus not suited for solar windows or tandem cells. Other techniques to
increase scattering within the cell have also been reported including mixing nanoparticles with
slightly larger scattering titania particles.38
Figure 1-8: Cross-section of a DSSC with top scattering layer 26
11
Apart from the scattering and its improvement on current generation, morphological changes
have had the primary goal of improving charge collection. Vertical structures have been a
constant focal point due to the fact that electrons can travel vertically to the electrode in a more
direct manner, resulting in improved charge transport properties. Titania nanotubes have been
reported, produced via the anodic oxidation of titanium in fluoride-based electrolytes as
presented in Figure 1-9.39
The improved charge transport characteristics of these types of cells
do not make up for their lack of surface area relative to standard DSSCs and thus result in much
lower current densities. Zinc oxide nanorods have also been reported which have also resulted in
faster collection times.40
Other interesting vertical structures with improved charge transport
characteristics include fused titania nanowires as well as hollow fibers templated using natural
cellulose fibers.41,42
Both of these findings showed remarkably enhanced electron transport
properties as compared to the nanoparticle films. Many of these structures have a more “open”
structure than the conventional high surface area DSSCs and thus have clear advantages for
solid state solar cells as well as non-volatile electrolytes where infiltration of the hole transporter
is often troublesome.
Photonic structures such as Bragg Stacks and inverse-opals have also been incorporated in
DSSCs to obtain enhanced efficiency cells due to the slowing of light and creation of localized
resonance modes in these materials as well as the enhanced diffusion of electrolyte within a
macroporous structure.43,44
Mesoporous materials have also been created from surfactant
templates, allowing the precise control of titania pore size and their distribution.45
12
Figure 1-9: Titania nanotube array 46
Post treatments of standard DSSCs have also produced higher currents due to the improved
percolation pathways of electrons to the FTO electrode. Examples of this include TiCl4 as well
as atomic layer deposition (ALD) which have been found to reduce the surface area of the films,
improve the electronic contact within the film and to the FTO substrate as well as enhancing the
binding of N719 with the TiO2 surface.47
In yet another effort to increase the charge transport times, core-shell structures have been
introduced. These structures are a relatively new advance in DSSCs, in which a core material
(usually nanoparticles) is coated with a shell material (ALD or some other means). The goal of
these core-shell materials is to adjust the conduction and valence bands of the core and shell of
the DSSCs such that electron recombination is lowered and electron transport to the electrode is
more efficient. By using a core with better charge transport properties, transport lifetimes can be
significantly reduced. The core can also have a lower valence band resulting in electrons getting
“trapped” within this core material (such as Al2O3 shells on a titania core) thus reducing
recombination. Additionally the basicity of the “shell” can be selected to enhance dye
adsorption. This technique was initially reported by Zaban et al. in 2000, in which a TiO2 core
was uniformly coated with Nb2O5.48
Shortly after this study, Gratzel published a paper in which
TiO2 and SnO2 were coated with various semiconductors and insulators, including MgO, Y2O3,
Al2O3 and ZrO2 (Figure 1-10).49
SnO2 coated with various semiconductors resulted in
remarkable improvements in photovoltage as well as current densities due to the efficient charge
13
transport properties of SnO2 as well as the prevention of recombination introduced by the
semiconductor shell. As for the TiO2 films, none of the resulting cells had efficiencies matching
that of pure TiO2 DSSCs. The general trend for these cells was an increase in photovoltage and
decrease in current relative to pure anatase nanoparticles (see Figure 1-10, right). One special
case of core-shell structures in which the core is a TCO will be presented in Section 2.2.
Figure 1-10: Core-Shell DSSCs with SnO2 cores (left) and TiO2 cores (right) 49
1.3 Cell Fabrication
For these experiments, films of titania were all produced via the powerful “Doctor-Blading”
technique. In brief, three slides are assembled in an “H” configuration as shown in Figure 1-11.
Tape is used to hold the slides together as well as to create the spacing required to produce the
nanoparticle film. A solution of nanoparticles is then applied to the top slide. A fourth slide is
then used to draw the solution down to the slide of interest in a smooth and quick motion to coat
the working electrode with nanoparticles. The film is then left to dry and the tape is removed.
The film is then cut into four sections using a glass cutter and the titania film is scratched into
5mm by 5mm squares using a razorblade. Films are treated at 450˚C for half an hour before
soaking in a dye solution. This treatment process is used to remove any porogens used as well as
to sinter the nanoparticles together and remove any water from the active region of the cell. It is
important to note that the samples should be hot (70˚C) when immersed in a dye solution to
prevent H2O adsorption.
14
Figure 1-11: Diagram of the Doctor Blading technique
DSSCs are fabricated using a high surface area nanoparticle films. Typically, particle sizes
range from 10 nm to 20 nm and the porosity of the film is carefully controlled by the addition of
PEG or cellulose (Figure 1-12). In state of the art cells, the working electrode consists of about
12 microns of titania paste along with a 5 microns scattering titania layer above the active
region.
Figure 1-12: SEM image of a titania nanoparticle film2
The dye selected was the Ruthenizer 535-bisTBA, purchased from Solaronix. Once the titania
films were removed from the oven, the samples were left to cool until approximately 70˚C and
immersed in a 0.5mM solution of the dye in Methanol. The samples were left overnight in the
dark for dye adsorption. The components for cell assembly are shown in Figure 1-13.
Components include: A) The top glass slide, B and C) Surlyn (DupontTM
) seals, D) Counter
electrode (with platinum), E) Surlyn gasket, and F) Working electrode (with dyed titania paste).
15
Figure 1-13: Components of a solar cell
Counter electrodes were fabricated by first drilling a hole using a diamond tip drill through an
FTO piece of dimensions 2cm by 3cm. Tape was first applied on both sides of the slide to
prevent the breaking of the electrode. Water is applied during the drilling process to prevent
fracture of the glass. Once the electrodes have holes in them, they are washed with a Piranha
solution for 1 hour to remove any organic residues (consisting of a 3:1 solution of Sulphuric
acid: Hydrogen Peroxide). The electrodes were then removed from this solution, washed in
Ethanol and stored. Before being used in a cell, counter electrodes are coated with platinum
using a 7mM solution of Hexachloroplatinic acid in Ethanol. A few drops are placed on the
counter electrode with the excess wiped off using the end of the glass pipette and left to dry in
an inclined position. The counter electrode is then heated at 450˚C in a furnace for half an hour.
Upon cooling, the counter electrode is ready to be used in a working cell.
Surlyn spacers are cut from a large sheet of Surlyn sandwiched between two grid paper sheets.
Surlyn was usually cut in squares of 9mm by 9mm with inner square dimensions of 7mm by
7mm. The top Surlyn seals were also cut from these sheets with dimensions of 9mm by 9mm.
After soaking the working electrode in dye solution overnight, the cells were assembled. To
begin, the working electrodes are removed from the dye solution and washed in Methanol for 15
seconds waving mildly with tweezers. The electrode is then removed from the wash solution
and left to dry in a vertical position. The Surlyn gasket is placed around the titania square and
the counter electrode is then placed above the working electrode sandwiching the Surlyn gasket.
16
A smooth hot plate at 200˚C is used to seal the cell, pushing with force to melt the Surlyn
uniformly. A piece of Teflon between the hot plate and the cell eases the process. During this
process, it is important to place the counter electrode on the hotplate and not the dyed titania
electrode since the dye might degrade under the heat. Using your finger to apply pressure
ensures that the cell will not become too hot. Once the Surlyn is melted (easily seen visually),
the cell is left to cool in the dark.
The next phase of cell assembly is to fill the solar cell with electrolyte and seal the cell. To do
this, first a square piece of Surlyn is used to seal the counter electrode hole, again using the
hotplate and a piece of Teflon (Figure 1-14, B). This process should be very quick since the
polymer is directly exposed to the heat (~5 seconds). Once the cell is sealed, a small hole is
punctured through the Surlyn by a syringe needle. A drop of electrolyte is then placed above the
piercing (Figure 1-14, D). The electrolyte solution was also purchased from Solaronix. At this
stage, a vacuum is applied to the cell, thus causing the liquid to enter the interior. To do this, a
high vacuum pump connected to an inverted funnel is used, the bottom of which was coated in
PDMS to provide a good seal during the vacuum infiltration. Excess electrolyte is then wiped
off the cell and partially removed from the interior of the cell. It is important to have an air
pocket within the cell since this will prevent electrolyte leakage (Figure 1-14, F). A second
Surlyn piece is then used to seal the top of the cell but this time a round top microscope slide is
used to seal the hole permanently by applying heat (Figure 1-14, G). This step is performed
quickly to prevent electrolyte from escaping the cell. Electrical contacts are then made on the
front and counter electrodes using silver paste or solder.
17
Figure 1-14: Cross-sectional view of DSSC cell assembly
1.4 Operation
Solar cell testing is performed using a solar simulator obtained from Newport as seen in Figure
1-15. The output power of the Power Supply (model 69907) used during cell testing was of
132W. A reference silicon solar cell was used to ensure proper working conditions of 1 Sun. A
150W Xenon lamp along with the Oriel Sol1A light source was used to illuminate the cells.
Electrical contacts were made using alligator clips to the working and counter electrodes.
Testing the cell involves scanning from 0V to 1V while measuring the current, scanned at a rate
of 100mV/s.
Figure 1-15: Solar Simulator setup
18
The typical output (operating curve) is shown in Figure 1-16. Point A depicts the current
obtained at 0V bias (Short Circuit Current) and point B is the point with zero current (Open
Circuit Voltage). The power obtained from 1 Sun is of 1000W/m2. To find the maximum power
delivered from the solar cell, the values of current and voltage along the operating line can be
multiplied to find the maximum value. The fill factor is the ratio of this calculated power to the
maximum cell power (A*B). The overall efficiency is the maximum cell power divided by the
incident power of the light. Although there are many techniques which provide a much deeper
understanding of the internal processes of DSSCs (such as impedance spectroscopy and IPCE)
the only equipment available to us for characterization for these experiments was the solar
simulator and the resulting I-V curves.
Figure 1-16: Typical experimental results obtained from a solar simulator
During operation, the efficiency of solar cells is reduced by the dissipation of power due to
various internal resistances. Using impedance spectroscopy, it is possible to pin down exactly
why and where these resistances are, for example, due to slow electrolyte diffusion using
viscous electrolytes, recombination at the FTO/electrolyte interface or electron transport
resistances. It is possible to lump all of the losses together into two terms, as a parallel shunt
resistance (RSH) and as a series resistance (RS) depicted in Figure 1-17. This figure is a
simplified equivalent circuit diagram of a DSSC, depicting the power source (Il), the load
(dotted), the two types of resistances as well as a diode. In an ideal cell, RSH would be infinite –
and thus no electrons would find alternate pathways (such as recombination). On the other hand,
RS would be zero – resulting in no voltage drop before the load. Both of these values can be
crudely determined by the slopes of an I-V curve at the ISC and VOC as seen in Figure 1-18.
19
Figure 1-17: Equivalent circuit of a DSSC
Figure 1-18: Effect of RS and RSH on the shape of an I-V curve
In this chapter, a brief overview of DSSCs was presented, including their history, their
development, the challenges they now face as well as the methods to assemble and measure
solar cells. Among these challenges, collecting electrons more efficiently (i.e., faster transport
and slower recombination) is among the most pressing. The remainder of this thesis describes
the approach we have taken to address this issue.
As the title of the thesis might suggest, the general strategy to decrease the electron collection
time lies in extending the TCO into the active region of DSSCs. In the following chapter a brief
overview of TCOs will be presented as well as their use in DSSCs thus far as well as the
underlying reasons behind this new architecture.
20
References
(1) Aleklett, K.; Campbell, C. J. The peak and decline of world oil and gas production. Minerals
and Energy - Raw Materials Report 2003, 18, 5-20.
(2) Grätzel, M. Photoelectrochemical cells. Nature 2001, 414, 338-344.
(3) Chiappafreddo, P.; Gagliardi, A. The photovoltaic market facing the challenge of organic
solar cells: Economic and technical perspectives. Transition Studies Review 2010, 17, 346-
355.
(4) Grätzel, M. Recent advances in sensitized mesoscopic solar cells. Acc. Chem. Res. 2009, 42,
1788-1798.
(5) Smestad, G. P.; Grätzel, M. Demonstrating electron transfer and nanotechnology: A natural
dye-sensitized nanocrystalline energy converter. J. Chem. Educ. 1998, 75, 752-756.
(6) M, R. M. Becquerel on solar physics. Nature 1874, 11, 133.
(7) Gerischer, H.; Michel-Beyerle, M.; Rebentrost, F.; Tributsch, H. Sensitization of charge
injection into semiconductors with large band gap. Electrochim. Acta 1968, 13, 1509-1515.
(8) Shah, A.; Torres, P.; Tscharner, R.; Wyrsch, N.; Keppner, H. Photovoltaic technology: The
case for thin-film solar cells. Science 1999, 285, 692-698.
(9) Tsubomura, H.; Matsumura, M.; Nomura, Y.; Amamiya, T. Dye sensitised Zinc oxide:
Aqueous electrolyte: Platinum photocell. Nature 1976, 261, 402-403.
(10) Matsumura, M.; Matsudaira, S.; Tsubomura, H.; Takata, M.; Yanagida, H. Dye
sensitization and surface structures of semiconductor electrodes. Industrial & Engineering
Chemistry Product Research and Development 1980, 19, 415-421.
(11) O'Regan, B.; Grätzel, M. A low-cost, high-efficiency solar cell based on dye-sensitized
colloidal TiO2 films. Nature 1991, 353, 737-740.
(12) Grätzel, M.; Frank, A. J. Interfacial electron-transfer reactions in colloidal semiconductor
dispersions. Kinetic analysis. J. Phys. Chem. 1982, 86, 2964-2967.
(13) Hagfeldt, A.; Grätzel, M. Molecular photovoltaics. Acc. Chem. Res. 2000, 33, 269-277.
(14) Nelson, J. Continuous-time random-walk model of electron transport in nanocrystalline
TiO2 electrodes. Physical Review B - Condensed Matter and Materials Physics 1999, 59,
15374-15380.
(15) Frank, A. J.; Kopidakis, N.; Lagemaat, J. V. D. Electrons in nanostructured TiO2 solar
cells: Transport, recombination and photovoltaic properties. Coord. Chem. Rev. 2004, 248,
1165-1179.
21
(16) Gregg, B. A.; Pichot, F.; Ferrere, S.; Fields, C. L. Interfacial recombination processes in
dye-sensitized solar cells and methods to passivate the interfaces. J Phys Chem B 2001,
105, 1422-1429.
(17) Wang, M.; Anghel, A. M.; Marsan, B.; Ha, N. -. C.; Pootrakulchote, N.; Zakeeruddin, S.
M.; Grätzel, M. CoS supersedes Pt as efficient electrocatalyst for triiodide reduction in dye-
sensitized solar cells. J. Am. Chem. Soc. 2009, 131, 15976-15977.
(18) Haque, S. A.; Palomares, E.; Cho, B. M.; Green, A. N. M.; Hirata, N.; Klug, D. R.;
Durrant, J. R. Charge separation versus recombination in dye-sensitized nanocrystalline
solar cells: The minimization of kinetic redundancy. J. Am. Chem. Soc. 2005, 127, 3456-
3462.
(19) Bai, Y.; Cao, Y.; Zhang, J.; Wang, M.; Li, R.; Wang, P.; Zakeeruddin, S. M.; Grätzel, M.
High-performance dye-sensitized solar cells based on solvent-free electrolytes produced
from eutectic melts. Nature Materials 2008, 7, 626-630.
(20) Law, C.; Pathirana, S. C.; Li, X.; Anderson, A. Y.; Barnes, P. R. F.; Listorti, A.; Ghaddar,
T. H.; ORegan, B. C. Water-based electrolytes for dye-sensitized solar cells. Adv Mater
2010, 22, 4505-4509.
(21) Snaith, H. J.; Schmidt-Mende, L. Advances in liquid-electrolyte and solid-state dye-
sensitized solar cells. Adv Mater 2007, 19, 3187-3200.
(22) Wang, P.; Zakeeruddin, S. M.; Humphry-Baker, R.; Grätzel, M. A binary ionic liquid
electrolyte to achieve 7% power conversion efficiencies in dye-sensitized solar cells.
Chemistry of Materials 2004, 16, 2694-2696.
(23) Mishra, A.; Fischer, M. K. R.; Buerle, P. Metal-Free organic dyes for dye-Sensitized solar
cells: From structure: Property relationships to design rules. Angewandte Chemie -
International Edition 2009, 48, 2474-2499.
(24) Alibabaei, L.; Wang, M.; Giovannetti, R.; Teuscher, J.; Di Censo, D.; Moser, J. -.; Comte,
P.; Pucciarelli, F.; Zakeeruddin, S. M.; Grätzel, M. Application of Cu(ii) and Zn(ii)
coproporphyrins as sensitizers for thin film dye sensitized solar cells. Energy and
Environmental Science 2010, 3, 956-961.
(25) Bae, E.; Choi, W. Effects of anchoring groups (carboxylate vs phosphonate) in ruthenium
complex-sensitized TiO2 on photoelectrochemical behaviors and hydrogen production
under visible light; ACS, Division of Environmental Chemistry; 2005; Vol. 45, pp 747-749.
(26) Wang, P.; Zakeeruddin, S. M.; Moser, J. E.; Nazeeruddin, M. K.; Sekiguchi, T.; Grätzel, M.
A stable quasi-solid-state dye-sensitized solar cell with an amphiphilic ruthenium sensitizer
and polymer gel electrolyte. Nature Materials 2003, 2, 402-407.
(27) Chen, C. -.; Wu, S. -.; Li, J. -.; Wu, C. -.; Chen, J. -.; Ho, K. -. A new route to enhance the
light-harvesting capability of ruthenium complexes for dye-sensitized solar cells. Adv
Mater 2007, 19, 3888-3891.
22
(28) Nazeeruddin, M. K.; Pachy, P.; Grätzel, M. Efficient panchromatic sensitization of
nanocrystalline TiO2 films by a black dye based on a trithiocyanato-ruthenium complex.
Chemical Communications 1997, 1705-1706.
(29) Chen, C. -.; Wu, S. -.; Wu, C. -.; Chen, J. -.; Ho, K. -. A ruthenium complex with superhigh
light-harvesting capacity for dye-sensitized solar cells. Angewandte Chemie - International
Edition 2006, 45, 5822-5825.
(30) Wang, B.; Kerr, L. L. In In Nanostructured TiO2 and ZnO solar cells using CdS as
sensitizer: Stability investigation; Conference Record of the IEEE Photovoltaic Specialists
Conference; 2010; pp 1819-1822.
(31) Rensmo, H.; Keis, K.; Lindstroem, H.; Sadergren, S.; Solbrand, A.; Hagfeldt, A.;
Lindquist, S. -.; Wang, L. N.; Muhammed, M. High light-to-energy conversion efficiencies
for solar cells based on nanostructured ZnO electrodes. J Phys Chem B 1997, 101, 2598-
2601.
(32) Snaith, H. J.; Ducati, C. SnO2-Based dye-sensitized hybrid solar cells exhibiting near unity
absorbed photon-to-electron conversion efficiency. Nano Letters 2010, 10, 1259-1265.
(33) Ke, L.; Dolmanan, S. B.; Shen, L.; Pallathadk, P. K.; Zhang, Z.; Ying Lai, D. M.; Liu, H.
Degradation mechanism of ZnO-based dye-sensitized solar cells. Solar Energy Mater.
Solar Cells 2010, 94, 323-326.
(34) Green, A. N. M.; Palomares, E.; Haque, S. A.; Kroon, J. M.; Durrant, J. R. Charge transport
versus recombination in dye-sensitized solar cells employing nanocrystalline TiO2 and
SnO2 films. J Phys Chem B 2005, 109, 12525-12533.
(35) Barbé, C. J.; Arendse, F.; Comte, P.; Jirousek, M.; Lenzmann, F.; Shklover, V.; Grätzel, M.
Nanocrystalline titanium oxide electrodes for photovoltaic applications. J Am Ceram Soc
1997, 80, 3157-3171.
(36) Wang, Z. -.; Kawauchi, H.; Kashima, T.; Arakawa, H. Significant influence of TiO2
photoelectrode morphology on the energy conversion efficiency of N719 dye-sensitized
solar cell. Coord. Chem. Rev. 2004, 248, 1381-1389.
(37) Hore, S.; Vetter, C.; Kern, R.; Smit, H.; Hinsch, A. Influence of scattering layers on
efficiency of dye-sensitized solar cells. Solar Energy Mater. Solar Cells 2006, 90, 1176-
1188.
(38) Ferber, J.; Luther, J. Computer simulations of light scattering and absorption in dye-
sensitized solar cells. Solar Energy Mater. Solar Cells 1998, 54, 265-275.
(39) Greene, L. E.; Yuhas, B. D.; Law, M.; Zitoun, D.; Yang, P. Solution-grown zinc oxide
nanowires. Inorg. Chem. 2006, 45, 7535-7543.
(40) Gonzalez-Valls, I.; Lira-Cantu, M. Vertically-aligned nanostructures of ZnO for excitonic
solar cells: A review. Energy and Environmental Science 2009, 2, 19-34.
23
(41) Tetreault, N.; Horvath, E.; Moehl, T.; Brillet, J.; Smajda, R.; Bungener, S.; Cai, N.; Wang,
P.; Zakeeruddin, S. M.; Forro, L.; Magrez, A.; Grätzel, M. High-efficiency solid-state dye-
sensitized solar cells: Fast charge extraction through self-assembled 3D fibrous network of
crystalline TiO 2 nanowires. ACS Nano 2010, 4, 7644-7650.
(42) Ghadiri, E.; Taghavinia, N.; Zakeeruddin, S. M.; Grätzel, M.; Moser, J. -. Enhanced
electron collection efficiency in dye-sensitized solar cells based on nanostructured TiO2
hollow fibers. Nano Letters 2010, 10, 1632-1638.
(43) Huisman, C. L.; Schoonman, J.; Goossens, A. The application of inverse titania opals in
nanostructured solar cells. Solar Energy Mater. Solar Cells 2005, 85, 115-124.
(44) Colodrero, S.; Mihi, A.; Haggman, L.; Ocana, M.; Boschloo, G.; Hagfeldt, A.; Miguez, H.
Porous one-dimensional photonic crystals improve the power-conversion efficiency of dye-
sensitized solar cells. Adv Mater 2009, 21, 764-770.
(45) Choi, S. Y.; Mamak, M.; Speakman, S.; Chopra, N.; Ozin, G. A. Evolution of
nanocrystallinity in periodic mesoporous anatase thin films. Small 2005, 1, 226-232.
(46) Ghicov, A.; Schmuki, P. Self-ordering electrochemistry: A review on growth and
functionality of TiO2 nanotubes and other self-aligned MOx structures. Chemical
Communications 2009, 2791-2808.
(47) Zeng, L. -.; Dai, S. -.; Wang, K. -.; Pan, X.; Shi, C. -.; Guo, L. Mechanism of enhanced
performance of dye-sensitized solar cell based TiO2 films treated by titanium tetrachloride.
Chinese Physics Letters 2004, 21, 1835-1837.
(48) Chen, S. G.; Chappel, S.; Diamant, Y.; Zaban, A. Preparation of Nb2O5 coated TiO2
nanoporous electrodes and their application in dye-sensitized solar cells. Chemistry of
Materials 2001, 13, 4629-4634.
(49) Kay, A.; Grätzel, M. Dye-sensitized core-shell nanocrystals: Improved efficiency of
mesoporous tin oxide electrodes coated with a thin layer of an insulating oxide. Chemistry
of Materials 2002, 14, 2930-2935.
24
2 Transparent Conducting Oxides and DSSCs
2.1 What are TCOs
Transparent conducting oxides (TCOs) are the cornerstone of optoelectronic devices, allowing
light in the visible region to transmit with minimal loss. Together with their low electrical
resistivity and high thermal stability these properties render them ideally suited to meet a
multitude of optoelectronic device requirements wherein electrical contacts need to be made
without obscuring photons from entering or escaping devices. The first use of TCOs on a large
scale was during World War II to heat aircraft windshields although TCOs were produced from
CdO as early as 1907 by Badeker.1 The growth of the TCO industry only really took off during
the 1980s to 1990s with the large scale production of LCD displays. From 2003 to 2006, prices
of indium climbed from less than 100$/kg to over 1000$/Kg due to the widespread adoption of
portable displays, and to a much smaller extent, solar cells.2 Although ITO is the most
conductive of TCOs, the cost of this material has been slowly changing the TCO landscape. The
large infrastructure investments required for depositing ITO on an industrial scale has resulted
in a very slow transition from conventional TCOs to cheaper alternatives such as aluminum
doped zinc oxide, AZO, although one can be assured that the transition will occur at some point
in the near future.
The most commonly used TCOs for electrodes are fluorine-doped tin oxide and tin-doped
indium oxide (FTO and ITO) although there are numerous others including doped ZnO2, Ga2O3,
CdO and multiple other tertiary oxides.3 The vast majority of these conductors are n-type, and
much effort has been directed towards developing p-type conductors.4 TCOs can be produced in
a variety of ways, including spray pyrolysis,5 sol-gel techniques,
6 chemical vapour deposition,
7
and chemical bath deposition,8 and sputtering.
9
SnO2 is the focus of this study. SnO2 in its pure form is insulating. Thank goodness for
imperfection! Non-stoichiometry in SnO2 creates oxygen vacancies allowing the material to
become conductive. To increase conductivity, Sb can be added as a cation dopant and F can be
added as an anion dopant (although even these materials are not able to achieve the
conductivities achieved by ITO or doped ZnO). There are a number of reviews which provide an
in-depth review of the physics involved in these materials.10,11
Briefly, the reason for optical
transparency and electrical conductivity in these materials is due to four reasons: (i) TCOs
25
possess a wide optical bandgap prohibiting interband transitions in the visible region. (ii)
Intrinsic dopants (oxygen) or impurity dopants can donate electrons to the conduction band. (iii)
The conduction band is a single band of s-type character possessing a uniform distribution of the
electron charge density and relatively low scattering, allowing high electron mobility within the
material. (iv) The materials possess a large gap above the conduction band prohibiting inter-
conduction-band adsorption of photons in the visible specrum.11
Currently TCOs are used in DSSCs in the form of a dense planar 2D substrate. The following
section presents attempts to use TCO building blocks to extend the 2D electrode vertically.
2.2 TCO extension into the 3rd Dimension
The extension of the current collector into the matrix of a dye sensitized solar cell was as first
proposed and presented by Zaban et al. whose concept is depicted in Figure 2-1.12
The paper
reported the use of a TCO matrix covered in TiO2 via Atomic Layer Deposition (ALD), thus
resulting in a very short electron transport path equivalent to the thickness of the titania coating
(1-10 nm!). By creating a three-dimensional porous electrode, the electrons simply need to be
shuttled through the titania layer to reach the electrode. There have been a number of similar
concepts since this study all consisting of core-shell structures wherein a conductive inner phase
is coated in an insulating semiconductor. In this work, 6 nm of titania was required to achieve
maximal open circuit voltages. It was found using voltage decay measurements that the low
voltages achieved at titania thicknesses less than 6 nm were due to recombination occurring
through the insulating titania layer. No efficiencies were reported for this work. More recently,
the group has published ITO/TiO2 core-shell architectures with efficiencies reaching 1.6%.13
The extension of the current collector has also been attempted by Hupp et al. using an anodized
alumina membrane as a template.14
In this work, ITO and TiO2 are deposited within an anodized
alumina membrane (AAM) via ALD with the primary aim of drastically reducing electron
transport times to the electrode. In this study, it was shown that a minimum of 5 nm was
required to produce maximum photocurrent, consistent with the report from Zaban. These cells
produced current densities approximately 60X higher than the pure TiO2 analogue (AAM with
an ALD layer of TiO2) as well as the pure ITO analogue (AAM with an ALD layer of ITO).
Compared to state of the art cells, these cells performed rather poorly. The relatively low surface
26
areas of the AAM (400cm2/cm
3) as well as the backscattering within these materials are the
likely causes for the poor currents observed. Overall efficiencies were of 1.1% due to poor fill
factors emerging from shunt resistances. Possible explanations for this include poor connectivity
to the FTO substrate as well as inhomogeneous titania coatings on the ITO surface.
Figure 2-1: All-in-one DSSC initially presented by Zaban et al.12
Other techniques which have been reported to produce an extended current collector include
vertically aligned carbon nanofibers coated with TiO2 reaching up to 1.1% efficiency,15
TCOs
mats produced via electrospinning,16
as well as vertical ITO brushes obtained via pulsed laser
ablation, both covered with TiO2 via ALD to produce cells up to 0.15% efficiency (see Figure
2-2).17
The overall low efficiencies observed by all of these attempts is surprising and more
work needs to be done to determine their source. Impedance spectroscopy was not performed on
any of these materials. Titania deposited via ALD is amorphous and the effects of this on
DSSCs has not been well reported, presenting a possible explanation for their low efficiencies. \
Figure 2-2: ITO/TiO2 core/shell brush structure 17
27
2.3 Why and how to extend the TCO
In today‟s most efficient DSSCs, anywhere from 5 to 20 microns of titania nanoparticles are
used as the front electrode. This allows a very large amount of dye to anchor to its surface,
resulting in very high current densities. Most record-breaking cells achieve nearly 100%
incident photon to current efficiency (IPCE). The implications of this are simply that nearly all
of the current which could be collected is collected. Although there is room for improvement in
terms of photon to electron collection (particularly in the low energy region of the spectrum),
increasing the photovoltage presents a simpler way to increase the efficiency of DSSCs. There
are two ways in which one can increase the photovoltage within DSSCs, both resulting from
reducing recombination. Extending the TCO has the potential to dramatically decrease the
electron collection lifetime to the electrode and thus increase the photovoltage of a cell.
Secondly, faster and more effective charge transport should allow different redox shuttles to be
used with potentials approaching the dye potential which should also increase the photovoltage
in DSSCs.18
From a structural standpoint, the best way to understand the improvements resulting in an all-in-
one integrated electrode is mathematically. The mathematical description for the IV
characteristics in a DSSC can be solved using the continuity equation, shown below in Figure
2-3.18
Before we begin, I would like to make the distinction between the film thickness, d, and
the apparent film thickness for which there is no abbreviation. The apparent film thickness for
the image depicted in Figure 2-1 would be about 6 nm while the true film thickness would be of
several microns. In these explanations, it is assumed that if the current collector is extended as is
the case in Figure 2-1, then the value of “d” is simply the distance to the extended current
collector.
In the equations presented below, Uphoto represents the open circuit voltage and jsat represents the
reverse saturation current. As can be seen, jsat is directly proportional to d (the film thickness)
and thus as the film thickness decreases, the voltage will increase as long as the overall light
absorption as well as the recombination within the film remains the same. Thus, if one can find
a way to achieve the same current while reducing the “apparent” film thickness, the resulting
voltage should increase.18
Extending the current collector thus has the ability to decrease the
distance electrons must travel to the electrode (due to the extended collector) as well as the
28
ability to achieve the same current densities (due to the maintained high surface area), resulting
in higher voltages.
Figure 2-3: Continuity equation and solution for a DSSC under short circuit conditions18
The tortuous path in which electrons must travel to the FTO substrate often results in
recombination. As presented in Figure 1-5, electron transport to the electrode is on the order of
hundreds of microsecond.19
Due to this, aggressive redox shuttles have been avoided since these
would effectively prevent electron collection to the electrode.20
To date, there are only a handful
of other redox shuttles that have been used due to their limited success. Although these more
aggressive redox shuttles do not work well in conventional DSSCs, they have the potential to
work very well in systems where electron lifetime is reduced. These new redox shuttles have
been shown to substantially improve the photovoltage in DSSCs (1.3V vs. 700mV). Thus, not
only will the structure of the extended current collector have the ability to improve the observed
photovoltage, but the potential to use alternative redox shuttles opens an entirely new avenue to
produce potentially record breaking solar cells.
One of the biggest challenges when extending the current collector is to prevent “shorting”.
Shorting is simply a synonym for recombination, also called the dark current, and can occur at
various points in the cell as was presented in Figure 5. In conventional cells, it has previously
been reported that the bulk of recombination occurs near the FTO/TiO2 interface, and thus has
led to the development of a dense TiO2 blocking layer between the bare FTO substrate and the
mesoporous TiO2 particles.21
The overall influence of this blocking layer has been an increase in
efficiency by around 10% - quite high considering that the FTO/electrolyte interface has a very
low surface area.22
If one is to have a three dimensional TCO framework, its entirety must be
29
uniformly covered with a semiconductor to prevent shorting. It is postulated that due to the
much higher TCO surface area in a 3D TCO framework relative to the planar analogue, the
elimination of dark current at the TCO/electrolyte interface would result in much greater
improvements.
There is yet another challenge to produce an all-in-one integrated electrode. It is well known
that recombination occurs where there is a build-up of charge. It is thus important to have a
continuous conducting framework since any discontinuity would result in immediate
recombination. The fact that electrons are “trapped” within a TCO core, surrounded by a
semiconductor matrix, would only make matters worse since they physically could not reach the
FTO substrate.
Knowing these potential advantages and requirements, how could one design a system in which
these potential improvements can be achieved? First, high surface area must be maintained. The
reason for this is due to the fact that current must not suffer at the expense of voltage. Secondly,
shorting must be prevented to the greatest extent as possible to prevent recombination, achieved
by having a continuous conducting framework as well as a dense blocking layer over its
entirety. Lastly, one of the most important features of DSSCs is their low cost. Ideally,
extending the current collector can be produced in an economical and scalable way.
In this section we have reviewed TCOs and their use in DSSCs as well as the reasons for
wanting to use them as platforms for next-generation solar cells. In the next section, attempted
strategies to build these cells will be presented.
30
References
(1) Brehm, J. U.; Winterer, M.; Hahn, H. Synthesis and local structure of doped nanocrystalline
zinc oxides. J. Appl. Phys. 2006, 100.
(2) O'Neill, B. Indium market forces, a commercial perspective; Conference Record of the IEEE
Photovoltaic Specialists Conference; 2010; , pp 556-559.
(3) Minami, T. New n-type transparent conducting oxides. MRS Bull 2000, 25, 38-44.
(4) Kawazoe, H.; Yanagi, H.; Ueda, K.; Hosono, H. Transparent p-type conducting oxides:
design and fabrication of p-n heterojunctions. MRS Bull 2000, 25, 28-36.
(5) Shanthi, E.; Dutta, V.; Banerjee, A.; Chopra, K. L. Electrical and optical properties of
undoped and antimony-doped tin oxide films. J. Appl. Phys. 1980, 51, 6243-6251.
(6) Terrier, C.; Chatelon, J. P.; Roger, J. A. Electrical and optical properties of Sb:SnO2 thin
films obtained by the sol-gel method. Thin Solid Films 1997, 295, 95-100.
(7) Kim, K. -.; Yoon, S. -.; Lee, W. -.; Ho Kim, K. Surface morphologies and electrical
properties of antimony-doped tin oxide films deposited by plasma-enhanced chemical
vapor deposition. Surface and Coatings Technology 2001, 138, 229-236.
(8) Luo, W. -.; Tsai, T. -.; Yang, J. -.; Hsieh, W. -.; Hsu, C. -.; Fang, J. -. Enhancement in
conductivity and transmittance of zinc oxide prepared by chemical bath deposition. J
Electron Mater 2009, 38, 2264-2269.
(9) Klein, A.; Körber, C.; Wachau, A.; Säuberlich, F.; Gassenbauer, Y.; Schafranek, R.; Harvey,
S. P.; Mason, T. O. Surface potentials of magnetron sputtered transparent conducting
oxides. Thin Solid Films 2009, 518, 1197-1203.
(10) Chopra, K. L.; Major, S.; Pandya, D. K. Transparent conductors-A status review. Thin
Solid Films 1983, 102, 1-46.
(11) Batzill, M.; Diebold, U. The surface and materials science of tin oxide. Prog Surf Sci 2005,
79, 47-154.
(12) Chappel, S.; Chen, S. -.; Zaban, A. TiO2-coated nanoporous SnO2 electrodes for dye-
sensitized solar cells. Langmuir 2002, 18, 3336-3342.
(13) Grinis, L.; Ofir, A.; Dor, S.; Yahav, S.; Zaban, A. Collector-shell mesoporous electrodes
for dye sensitized solar cells. Isr. J. Chem. 2008, 48, 269-275.
(14) Martinson, A. B. F.; Elam, J. W.; Liu, J.; Pellin, M. J.; Marks, T. J.; Hupp, J. T. Radial
electron collection in dye-sensitized solar cells. Nano Letters 2008, 8, 2862-2866.
31
(15) Li, J.; Liu, J.; Kuo, Y. -.; Klabunde, K.; Rochford, C.; Wu, J. Vertically aligned carbon
nanofiber array coated with TiO2 shell as new architectures for solar energy conversion;
ACS National Meeting Book of Abstracts 2009.
(16) Ostermann, R.; Rudolph, M.; Schlettwein, D.; Smarsly, B. M. Novel nanostructured
photoelectrodes - Electrodeposition of metal oxides onto transparent conducting oxide
nanofibers; Materials Research Society Symposium Proceedings 2010; Vol. 1211, pp 1-5.
(17) Joanni, E.; Savu, R.; de Sousa Góes, M.; Bueno, P. R.; de Freitas, J. N.; Nogueira, A. F.;
Longo, E.; Varela, J. A. Dye-sensitized solar cell architecture based on indium-tin oxide
nanowires coated with titanium dioxide. Scr. Mater. 2007, 57, 277-280.
(18) Martinson, A. B. F.; Hamann, T. W.; Pellin, M. J.; Hupp, J. T. New architectures for dye-
sensitized solar cells. Chemistry - A European Journal 2008, 14, 4458-4467.
(19) Wang, Q.; Ito, S.; Grätzel, M.; Fabregat-Santiago, F.; Mora-Sera, I.; Bisquert, J.; Bessho,
T.; Imai, H. Characteristics of high efficiency dye-sensitized solar cells. J Phys Chem B
2006, 110, 25210-25221.
(20) Nusbaumer, H.; Moser, J. -.; Zakeeruddin, S. M.; Nazeeruddin, M. K.; Grätzel, M.
CoII(dbbip)2
2+ complex rivals tri-iodide/iodide redox mediator in dye-sensitized
photovoltaic cells. J Phys Chem B 2001, 105, 10461-10464.
(21) Zhu, K.; Schiff, E. A.; Park, N. -.; Van, d. L.; Frank, A. J. Determining the locus for
photocarrier recombination in dye-sensitized solar cells. Appl. Phys. Lett. 2002, 80, 685.
(22) Ito, S.; Liska, P.; Comte, P.; Charvet, R.; Pechy, P.; Bach, U.; Schmidt-Mende, L.;
Zakeeruddin, S. M.; Kay, A.; Nazeeruddin, M. K.; Grätzel, M. Control of dark current in
photoelectrochemical (TiO2/I --I3-) and dye-sensitized solar cells. Chemical
Communications 2005, 4351-4353.
32
3 3D Transparent Conducting Oxide frameworks
3.1 Strategies
During the course of this project, three general approaches were attempted to produce an all-in-
one integrated electrode, presented below in order of progression. The original approach to
produce an extended electrode was to construct a TCO framework with structural integrity,
porosity and interconnectivity. In this section, I will describe the attempts made to develop an
extended current collector using a templated approach seen in Figures A and B below. Both of
these techniques required a “template”, usually silica or polystyrene spheres, to create porosity
and interconnectivity. As will be discussed, numerous attempts to produce an all-in-one DSSC
failed using a template approach and thus required the development of non-templated approach,
presented in Figure C below, which will be discussed in the next chapter.
B) Opal Templating
Figure 3-1: 3D TCO via templated and non-templated approaches
A) Templated Co-Casting C) Non-templated co-casting
33
3.2 Templated Co-casting
In the initial stages of this research project, a simple two-step process was attempted to produce
the 3D-TCO framework. This was performed by mixing purchased ATO nanoparticles and
silica spheres in appropriate ratios, doctor blading this solution on an FTO slide, calcining the
film at 450˚C and finally etching of the silica spheres to produce a porous framework. The
challenges in this technique were mainly achieving a stable framework which was porous
enough to accommodate titania. At lower ATO weight fraction, the structure would collapse
upon etching due to the lack of interconnectivity. Alternatively, at higher ATO weight fractions,
the structure would not be porous enough. The optimized structure (Figure 3-2) was still not
porous enough to allow the infiltration of titania within the structure due to the lack of
interconnectivity between the pores.
Figure 3-2: Optimized 3D TCO framework produced via a co-casting template approach
To make the films, a 20wt% solution of silica spheres suspended in Ethanol was used as stock,
as well as a 50wt% solution of ATO nanoparticles in water purchased from Alfa Aesar. The
ATO solution was diluted to 10wt% using Ethanol and sonicated. The ratios of these two
solutions were mixed together to obtain films ranging from 80% spheres to 20% spheres by
weight. It is important to have a high concentration of Ethanol in solution to produce uniform
films (water beads on the surface of FTO). If the films were too thick, cracking could be
34
observed under the microscope and dilution with Ethanol was performed to reduce the film
thickness. Controlling cracking was challenging and restricted the thickness of films to about 5
microns. In addition, a higher weight percent of ATO nanoparticles also prevented cracking.
3.3 Opal Templating
Using an opal film as a template to produce an inverse opal (i-o) has been a quite common
method to produce macroporous frameworks (76% porous) from various materials including
nanoparticles.1 This approach, presented below, had the potential to produce a very well
interconnected macroporous TCO framework which would be ideally suited for titania
infiltration in a DSSC. In this work, a macroporous TCO framework is constructed along with
an example showing its use as an electrode material in which bismuth is reversibly
electrodeposited within the opal structure.
To exploit the full potential of TCOs as electrode materials for various device applications,
besides dense planar forms of the material, it is also necessary to find means to make them
porous at length scales traversing the microscopic (< 2 nm) through the mesoscopic (2-50 nm)
to the macroscopic (> 50 nm).2 Nothing currently exists in the open or patent literature on any
kind of TCO fashioned with arrays of macropores, either periodically or aperiodically arranged.
Structuring TCO electrodes into novel architectures over different length scales is emerging as a
powerful way to boost current densities and shorten electron transport distances for a range of
device applications.3-6
To date, structured electrode materials have been produced using bottom-
up techniques such as electroplating,7 de-alloying,
8 as well as surfactant or colloidal crystal
templating.9, 10
Periodic macroporous electrodes made by template replication of opals have
been described for carbon and silicon for use as anodes in lithium ion batteries,6, 11
and metallic
inverse opals for applications in electrocatalysis.12
The preparation of (electro)chemically stable
macroporous TCO has however remained elusive due to materials synthesis challenges
associated with the doping requirements to achieve desired electrical and optical properties. In
addition, maintaining structural integrity has remained a challenge due to adverse mechanical
effects associated with chemically induced shrinkage during structure formation.
35
We have managed to surmount these obstacles thanks to recent progress in the synthesis of
colloidally stable TCO nanocrystals.13
Described herein we present the synthesis details,
structure determination, properties measurements and demonstration of electrode function for
the first example of a periodic macroporous TCO film.
In this study, a bottom-up synthetic pathway is presented to produce the first example of a
periodic macroporous TCO electrode. The synergism of optical, electrical and structural
properties of the material is exploited to create a novel electrochemically actuated optical
switch, which relies on the reversible electrodeposition and electrodissolution of bismuth metal
within the macropores to modify in a predetermined cyclic manner the photonic properties of the
material from light diffracting to light reflecting in the visible wavelength range. This proof-of-
concept demonstration highlights the ability of a periodic macroporous TCO light scale material
to host another material, while exploiting its electrical, optical and photonic properties to
perform a specific function. This capability is impossible for either periodic microporous or
mesoporous TCO length scales due to their inability to diffract visible light. This key point is
amplified upon below.
Why is the control of pore size and crystalline porosity in TCO film important? First of all, in
contradistinction to known periodic mesoporous TCO film, a periodic macroporous TCO film
can function as a photonic crystal displaying optical diffraction tunability and slowing of light,
enabling it to be utilized for example as an optoelectronic chemical sensor or to enhance photon
harvesting in solar cells. Secondly, the large pores of this TCO material allow efficient mass
fluid flow through the open structure permitting their contact with the large internal surface of
the electrode. Furthermore, a macroporous TCO has the ability to host another material within
the voids including various nanoparticles for solar cell, light emitting diode and sensor
applications. Because of the attributes of this genre of TCO, in this study it has proven possible
for a periodic macroporous TCO film to be used as a reversible electrochemically actuated
photonic crystal optical switch, illustrating thereby its distinction to all other forms of TCO,
dense or porous, reported in the open or patent literature.
In this study, 6 nm ncATO was used as a building block for synthesizing an inverse
nanocrystalline antimony-doped tin oxide opal denoted i-ncATO-o, seen in Figure 3-3. The
small size of these nanoparticles enabled a straightforward and scalable way to create a
transparent conducting periodic macroporous film able to function as an electrode for the
36
operation of a prototypical electrochemical redox reaction and demonstration of a proof-of-
concept electro-optical light valve.
Figure 3-3: STEM image of ATO nanocrystals
The assembly of the periodic macroporous TCO framework is outlined in Figure 3-4. In brief, a
silica opal film was grown by evaporation induced self assembly and infiltrated with ncATO
which was synthesized according to a modified literature preparation.13
The particle size of as-
synthesized ncATO was shown to be 6 nm by Dynamic Light Scattering (DLS), 5-7 nm by
Scanning Electron Microscopy (SEM) and 5 nm by powder X-ray diffraction (PXRD). The very
small size of ncATO was crucial for its controlled infiltration into the void spaces of the silica
opal lattice.
37
Figure 3-4: Self-assembly steps for i-ATO-o
The as-synthesized ncATO solution was spin-coated into the silica opal film, resulting in an
infiltrated opal with little to no excess nanoparticles covering the template. The resulting
ncATO-SiO2-o film was then heated to 450˚C in air to provide mechanical strength to the
ncATO framework. The effectiveness of the infiltration was judged from the optically
determined volume filling fraction determined from the Bragg-Snell equation. The infiltration of
ATO within the opal structure reduces the refractive index contrast between the spheres and the
surrounding matrix, resulting in a red shift of the Bragg peak as well as a decrease in
reflectivity.14
Figure 3-5b (solid line) shows the progression of the ncATO infiltration into the
silica opal template with increasing number of spin coatings of the ATOnc solution.
At this stage of the synthesis, etching of SiO2 spheres resulted in a collapsed structure. To
further improve the mechanical integrity of the structure, ATO sol was infiltrated into the voids
between ncATO in the ncATO-SiO2-o film and thermally treated to yield an ATO
nanocomposite with improved structural stability and higher conductivity.15
The silica spheres
were then etched with 1% HF. The filling fraction of this mechanically enhanced i-ncATO-o
was calculated to be 65% percent as determined by the wavelength of the optical Bragg
reflection.
38
Figure 3-5: Structure and optical reflectivity spectra of i-ncATO-o films
The results of PXRD indicate a Cassiterite structure (SnO2) for both ncATO and ATO sol
(Figure 3-6). In this Figure, A is the PXRD for the ATO nanocrystals, B is for the ATO sol and
C is the reference PXRD reflection diagnostics for the tin dioxide Cassiterite structure. Samples
were prepared by spin coating solutions of ATO nanocrystals and ATO sol on a silicon wafer
and calcining then at 450°C for 2hr before measurement. PXRD presented in Figure 3-6 shows a
higher crystallinity in ncATO compared to ATO sol consistent with the higher conductivity
observed in ncATO samples, presented below. In addition, XPS was performed on the ATO
nanocrystals with an antimony doping of 6.6 at% compared to an antimony doping of 3.2 at% in
ATO sol which also plays a significant role in the measured conductivities.13
Figure 3-6: Crystal structure of the ATO nanocrystals (a) and sol (b)
39
Conductivity and porosity measurements were performed on the ATOnc, ATO sol and ATO
composite films. Films of pure ATO nanocrystals and sol reveal that the nanocrystals are
significantly more conductive and porous than the ATO sol (Figure 3-7). The resistance of spin
coated ATO films were measured by the four-probe Van-der Pauw method, a well established
technique for measuring the resistivity of arbitrary shape samples. The data presented consists of
the ATO nanocrystals infiltrated with ATO sol (1-4 infiltrations) by spin coating a solution of
sol on the nanocrystal film. Results indicate that ATOnc films with ATO sol post-infiltration
(termed ATO nanocomposite) has the lowest resistance. Ellipsometric porosimetry
measurements confirm that the porosity of the i-ncATO-o film is greatly reduced by ATO sol
infiltration. We attribute the low resistance of the ATO nanocomposite to consolidation of the
grain boundaries between ncATO as well as a decrease in porosity from the infiltration of ATO
sol in the void spaces between ncATO in the i-ncATO-o framework, creating percolative charge
transport pathways between the nanocrystals, as seen in the image presented in Figure 3-8.16-18
It
is interesting that with just one treatment of ATO sol, the conductivity and porosity of the film
reaches a plateau.
Figure 3-7: Resistance and porosity measurements of the ATO materials
40
Figure 3-8: Diagram of the improved interconnectivity before and after ATOsol
The use of i-ncATO-o as a platform electrode for electrochemistry was investigated by cyclic
voltametry using ferrocene as a representative redox shuttle (Figure 3-9). The redox activity of
i-ncATO-o (Figure 3-9a) is found to be slightly higher than the pristine ncATO thin film of the
same weight produced via spin coating. It is believed that the reason for this is due to the higher
accessible surface area of the electrode to ferrocene in the i-ncATO-o film compared to ncATO
in planar ncATO films.
Figure 3-9: Periodic macroporous transparent oxide i-ncATO-o electrode
41
To demonstrate functionality of the periodic macroporous TCO electrode, a prototype
electrochemical photonic crystal switch was constructed whose operation is founded on the
reversible electrodeposition of bismuth within the void spaces of the i-ncATO-o electrode to
tune its photonic crystal properties. In our demonstration cyclic voltammetry was performed
from -2V to +2V at a scan rate of 20mV/s of a bismuth (III) electrolyte contained within the i-
ncATO-o photonic crystal electrode. As a result, the optical Bragg diffraction intensity of i-
ncATO-o was dramatically reduced, resulting in a switching from a red reflective state to a
black absorbing state (Figure 3-9 b and c). The origin of this change is the absorption of light of
the electrodeposited bismuth throughout the visible wavelength range preventing photon
penetration into the photonic lattice thereby eliminating the opportunity for Bragg diffraction.19
SEM images after electrodeposition of bismuth showed the metal within and around the
macropores of the i-ncATO-o electrode (Figure 3-10). X-ray photoelectron spectroscopy (XPS)
analysis was performed on the i-ncATO-o sample before and after electrodeposition of bismuth
provides additional evidence for the deposition of bismuth on the the i-ncATO-o electrode
(Figure 3-11).
Figure 3-10: SEM image of an inverse ATO opal with bismuth electrodeposited within.
42
Element
At.% before
deposition
At.% after
deposition
Bi 0.03 19.1
Sn 37.1 24.7
Sb 2.45 2.02
O 60.4 54.2
Figure 3-11: XPS atomic percentages before and after electrodeposition of Bi on the i-o
In summary, a periodic macroporous transparent conducting oxide, denoted i-ncATO-o, is
presented which for the first time offers a unique synergistic combination of optical
transparency, electrical conductivity and photonic crystal properties. Besides the demonstration
of a prototype reversible, electrochemically activated light valve for the i-ncATO-o electrode, its
large internal surface area and easily accessible network of interconnected macropores makes
the i-ncATO electrode interesting for the development of a number of enhanced performance
photoelectrochemical devices, like dye sensitized solar cells and water splitting, light emitting
diodes and chemical sensors.
Experimental
All the chemicals for this study were purchased from Sigma-Aldrich Co. and used without
further purification.
i-ncATO-o film:
The procedure to make the film is as follows: Spheres were grown via a modified Stӧber
method.20
The as-synthesized spheres were centrifuged and re-suspended in ethanol 5 times and
diluted to a 10wt% solution. An opal film was grown by the EISA of silica spheres at 40˚C over
3 days on an FTO substrate. The opal sample was then thermally treated to 450˚C to provide the
film with enhanced mechanical stability. ATO nanocrystals and ATO sol were used for
infiltration of opal film. These solutions were spin coated above the opal film, followed by heat
treatment at 450˚C. The samples were etched in 1%HF overnight to remove the silica sphere
template. ncATO did not etch is HF.
43
ncATO preparation:
A previously published method with modifications was used for synthesis of ATO
nanocrystals.13
A mixture of 0.4010 mmol of SbCl3 and 1.2967 mmol of SnCl4 was slowly
added to 20 ml benzyl alcohol. The mixture then was placed in the oven at 140°C for 2 hours.
Nanocrystals separated from the solution were washed with acetone and ethanol several times to
separate organic residue from the nanocrystals.
ATO sol preparation:
ATO sol was prepared in two different vials. A mixed solution of 59.8 mmol SnCl2.2H2O, 50ml
acetic acid and 50 ml ethanol was prepared. The mixture was then stirred for 2 hours at 70°C. A
second mixed solution was prepared by adding 2.4 mmol of SbCl3, 25ml acetic acid and 25ml
ethanol. The mixture was stirred for 2 hours at 70°C. The ATO sol was prepared by adding the
two precursor solutions and stirring for another 2 hours at 70°C and aging for 24 hours.
Cyclic Voltametry:
A 10mM solution of ferrocene and 0.5M tetrabutylammonium hexafluorophosphate in
acetonitrile was used for cyclic voltametry measurements. Voltage ranges used were -500mV to
1200mV at a scan rate of 100mV/s. Platinum counter and reference electrodes were used.
Sample weights were measured using a Mettler TOLEDO MX5 microbalance.
Electrodeposition:
Electrodeposition of bismuth was performed on i-ncATO-o films prepared on glass substrates.
Electrical contacts to i-ncATO-o film were made using silver paste. The entire film was
immersed in an electrolyte solution and placed within a Petri dish under the eyepiece of an
optical microscope fitted with a camera and fibre optic microoptical reflectance spectroscopy
attachment. bismuth films were electrodeposited according to literature. A 1M HCl solution
containing 0.02M BiCl3, 0.5M LiBr and 3mM CuCl2 was electroplated using a BAS epsilon
potentiostat in a conventional 3-electrode electrochemical cell using the cyclic voltammetry
setting from -2V to 2V at a scan rate of 20mV/s. Counter and reference electrodes were both
platinum. Optical spectra were obtained using an Ocean Optics SD2000 fibre optic
spectrophotometer along with an optical microscope. SEM images were obtained on a Hitachi
44
HD-2000. Spectroscopic ellipsometry analyses were performed in a Sopra GES-5E ellipsometer
at a fixed incidence angle of 70.15 in the range 1.2–3.5 eV. The modeling and regression
analyses of the ellipsometric spectra were performed using the software Winelli provided by the
manufacturer. Low angle X-ray diffraction patterns were acquired with a Siemens D5000
diffractometer using Cu Kα1 radiation operated at 50 KV and 35 mA with a Kevex solid-state
detector. Wide angle X-ray scattering patterns were obtained with a Bruker D8 diffractometer
operated at 40 KV and 40 mA.
3.4 Shortcomings
Although the i-o framework might seem like the ideal candidate for a TCO framework, in reality
there were problems related to this structure which limited its practical use in DSSCs. First,
inverse opals produced from nanoparticles (even very small ones!) have inherent challenges
associated with the infiltration process. Any aggregates or larger particles clog the opal
structure, resulting in an inverse opal with a gradient structure and subsequent poor mechanical
strength. Adhesion to the substrate was found to be non-ideal and it was often observed that the
i-o structure was “floating” in SEM images as seen in Figure 3-12. In a functional DSSC, such a
problem would result in severe recombination and an inefficient cell.
Figure 3-12: A “floating” i-o showing the poor adhesion to the substrate
45
Another problem with this structure was the overlayer which was often created. An overlayer is
formed when nanocrystals accumulate on the top of the opal surface such that upon etching, the
top is non-porous, as seen in Figure 3-13 . For this reason, Reactive Ion Etching is often
required to produce an accessible macroporous surface.21
Figure 3-13: Inverse opal with an overlayer
Poor interconnectivity of the framework was another challenge which had to be overcome in
this approach as seen in Figure 3-13 and Figure 3-14. Various techniques such as SiCl4 vapour
treatments, polymer coatings, heat, as well as a nanoparticle-based necking technique were
attempted. The biggest problem with all of these approaches (other than heat) is that the FTO
substrate would get coated with a layer of “necking material” which always resulted in no
interconnectivity between the TCO framework and the substrate due to the fact that the entire
substrate surface would become coated during these procedures (see Figure 3-15). As previously
established, heating the spheres above 800˚C on a silicon wafer did produce necking but
unfortunately FTO glass was not able to survive the treatment.22
Yet another shortfall of this structure was due to the coating of titania and infiltration of
nanoparticles required after the production of the framework. After fabrication of the TCO
framework, all of the pores must be filled with titania nanoparticles to produce a functional
46
DSSC. Before infiltration, the entire structure needs to be covered by a TiO2 layer to prevent
shorting. ALD was chosen as the method of choice to coat the structure with TiO2. From
published results of all-in-one DSSCs, at least 6 nm were required to prevent shorting.23
As can
be seen from Figure 3-14, the interconnectivity of the framework ranged from 0-80 nm for 250
nm spheres. After a conformal coating of 6 nm TiO2, the resulting pore sizes would be further
reduced resulting in a very challenging infiltration process into the structure. In addition, the
tortuous path nanoparticles must take to enter the framework made it nearly impossible to
infiltrate the framework.
Figure 3-14: Poor interconnectivity of the i-o framework
47
Figure 3-15: Necking of opal template resulting in a covered substrate
Infiltration of the structure was attempted although none achieved full infiltration (see Figure
3-16). The techniques attempted included dip coating and spin coating various TiO2
nanoparticles ranging in size from 2nm to 30nm, purchased from suppliers and also synthesized
in our lab. Synthesis of titania was produced by adding titanium tetraethoxide dropwise to 0.1 M
HNO3 at room temperature under stirring. The suspension was then left to stir at 80 °C for 8 h.
The resulting dispersion was filtered to remove any agglomerates and subsequently diluted to
the desired concentration.24
Growth within the structure was also attempted using TiF62-
,
previously reported by our lab.25
Other techniques included the production of TiO2 nanoparticles
via a hydrogen peroxide metal to metal oxide technique developed in our labs as well as the
hydrolysis at 60 °C of titanium tetrabutoxide in the presence of acetylacetone and para-
toluenesulfonic acid.26
Various techniques were attempted including spin coating, dip coating,
drop casting and evaporation induced infiltration. Multiple techniques resulted in cracking and
flaking upon heat treatment. All techniques only resulted in partial infiltration similar to the
figure presented below. With this in mind, a simpler approach to extend the current collector
was sought.
Figure 3-16: Infiltration of the TCO i-o with TiO2 nanoparticles
In this section, we reviewed two strategies to produce an all-in-one integrated electrode design
for DSSCs. Unfortunately, both strategies were doomed due to the various shortcomings
48
presented above. In the following section I present a new design for an all-in-one integrated
electrode which circumvents all of the limitations presented above.
References
(1) Kwak, E. S.; Lee, W.; Park, N. -.; Kim, J.; Lee, H. Compact inverse-opal electrode using
non-aggregated TiO2 nanoparticles for dye-sensitized solar cells. Advanced Functional
Materials 2009, 19, 1093-1099.
(2) Hou, K.; Puzzo, D.; Helander, M. C.; Lo, S. S.; Bonifacio, L. D.; Wang, W.; Lu, Z. -.;
Scholes, G. D.; Ozin, G. A. Dye-anchored mesoporous antimony-doped tin oxide
electrochemiluminescence cell. Adv Mater 2009, 21, 2492-2496.
(3) Fan, Z.; Ruebusch, D. J.; Rathore, A. A.; Kapadia, R.; Ergen, O.; Leu, P. W.; Javey, A.
Challenges and prospects of nanopillar-based solar cells. Nano Research 2009, 2, 829-843.
(4) Martinson, A. B. F.; Elam, J. W.; Liu, J.; Pellin, M. J.; Marks, T. J.; Hupp, J. T. Radial
electron collection in dye-sensitized solar cells. Nano Letters 2008, 8, 2862-2866.
(5) O'Brien, P. G.; Puzzo, D. P.; Chutinan, A.; Bonifacio, L. D.; Ozin, G. A.; Kherani, N. P.
Selectively transparent and conducting photonic crystals. Adv Mater 2010, 22, 611-616.
(6) Esmanski, A.; Ozin, G. A. Silicon inverse-opal-based macroporous materials as negative
electrodes for lithium ion batteries. Advanced Functional Materials 2009, 19, 1999-2010.
(7) Soleymani, L.; Fang, Z.; Sargent, E. H.; Kelley, S. O. Programming the detection limits of
biosensors through controlled nanostructuring. Nature Nanotechnology 2009, 4, 844-848.
(8) Erlebacher, J.; Aziz, M. J.; Karma, A.; Dimitrov, N.; Sieradzki, K. Evolution of
nanoporosity in dealloying. Nature 2001, 410, 450-453.
(9) Yan, H.; Blanford, C. F.; Holland, B. T.; Parent, M.; Smyrl, W. H.; Stein, A. Chemical
synthesis of periodic macroporous NiO and metallic Ni. Adv Mater 1999, 11, 1003-1006.
(10) Bartlett, P. N.; Birkin, P. R.; Ghanem, M. A. Electrochemical deposition of macroporous
platinum, palladium and cobalt films using polystyrene latex sphere templates. Chemical
Communications 2000, 1671-1672.
(11) Lee, K. T.; Lytle, J. C.; Ergang, N. S.; Oh, S. M.; Stein, A. Synthesis and rate performance
of monolithic macroporous carbon electrodes for lithium-ion secondary batteries. Advanced
Functional Materials 2005, 15, 547-556.
(12) Huang, Y. -.; Lai, C. -.; Wu, P. -.; Chen, L. -. A facile approach to fabricate Ni inverse
opals at controlled thickness. Mater Lett 2009, 63, 2393-2395.
(13) Müller, V.; Rasp, M.; Śtefanić, G.; Ba, J.; Günther, S.; Rathousky, J.; Niederberger, M.;
Fattakhova-Rohlfing, D. Highly conducting nanosized monodispersed antimony-doped tin
49
oxide particles synthesized via nonaqueous sol-gel procedure. Chemistry of Materials 2009,
21, 5229-5236.
(14) Hatton, B.; Kitaev, V.; Perovic, D.; Ozin, G.; Aizenberg, J. Low-temperature synthesis of
nanoscale silica multilayers - Atomic layer deposition in a test tube. Journal of Materials
Chemistry 2010, 20, 6009-6013.
(15) Terrier, C.; Chatelon, J. P.; Roger, J. A. Electrical and optical properties of Sb:SnO2 thin
films obtained by the sol-gel method. Thin Solid Films 1997, 295, 95-100.
(16) Montes, J. M.; Cuevas, F. G.; Cintas, J. Porosity effect on the electrical conductivity of
sintered powder compacts. Applied Physics A: Materials Science and Processing 2008, 92,
375-380.
(17) Riefler, N.; Mädler, L. Structure-conductivity relations of simulated highly porous
nanoparticle aggregate films. Journal of Nanoparticle Research 2010, 12, 853-863.
(18) Benkstein, K. D.; Kopidakis, N.; Van, d. L.; Frank, A. J. Influence of the percolation
network geometry on electron transport in dye-sensitized titanium dioxide solar cells. J
Phys Chem B 2003, 107, 7759-7767.
(19) Córdoba De Torresi, S. I.; Carlos, I. A. Optical characterization of bismuth reversible
electrodeposition. J Electroanal Chem 1996, 414, 11-16.
(20) Stöber, W.; Fink, A.; Bohn, E. Controlled growth of monodisperse silica spheres in the
micron size range. J. Colloid Interface Sci. 1968, 26, 62-69.
(21) Suezaki, T.; Chen, J. I. L.; Hatayama, T.; Fuyuki, T.; Ozin, G. A. Electrical properties of p-
type and n-type doped inverse silicon opals - towards optically amplified silicon solar cells.
Appl. Phys. Lett. 2010, 96, 242102.
(22) Scott, R. W. J.; Yang, S. M.; Coombs, N.; Ozin, G. A.; Williams, D. E. Engineered
sensitivity of structured tin dioxide chemical sensors: Opaline architectures with controlled
necking. Adv. Funct. Mater. 2003, 13, 225-231.
(23) Grinis, L.; Ofir, A.; Dor, S.; Yahav, S.; Zaban, A. Collector-shell mesoporous electrodes
for dye sensitized solar cells. Isr. J. Chem. 2008, 48, 269-275.
(24) Puzzo, D. P.; Bonifacio, L. D.; Oreopoulos, J.; Yip, C. M.; Manners, I.; Ozin, G. A. Color
from colorless nanomaterials: Bragg reflectors made of nanoparticles. Journal of Materials
Chemistry 2009, 19, 3500-3506.
(25) Chen, J. I. L.; Von Freymann, G.; Choi, S. Y.; Kitaev, V.; Ozin, G. A. In In Slow photons
in TiO2 inverse opals: Optical amplification and effect of disorder on the photocatalytic
efficiency 2007; Vol. 6650.
(26) Scolan, E.; Sanchez, C. Synthesis and Characterization of Surface-Protected
Nanocrystalline Titania Particles. Chemistry of Materials 1998, 10, 3217-3223.
50
4 Simplified procedure for an all-in-one DSSC
4.1 Basic concept
A non-templated approach was developed due to the challenges associated with the infiltration
of TCO nanoparticles within the opal and the infiltration of titania within the resulting TCO
framework. Compared to the two extended current collector concepts presented in the previous
chapter, one can see that the concept illustrated in Figure 4-1 is a simpler and potentially more
effective way to achieve the same results due to the lack of etching and infiltration steps
required. In this new technique, titania nanoparticle spheres are made and mixed with TCO
nanoparticles in solution. This two-component solution can then be cast as a film, resulting in an
all-in-one DSSC. To prevent shorting, a thin insulating film must be deposited above the entire
surface of the film. The beauty of this technique lies in the fact that both components can be
tuned. Big or small particles can be chosen for either component, thus allowing a greater
flexibility for device optimization. Large spheres or small spheres can be created and the ratio of
spheres to TCO nanoparticles can also be varied. In addition, the potential scalability of this
technique is more appealing than the template approach. Figure 4-2 shows an SEM cross section
image of a film produced using this technique. The white section in this image consists of TCO
nanoparticles which clearly show a continuous framework to the FTO substrate – a key factor
for success.
Figure 4-1: 3DTCO via a two-component approach
51
Figure 4-2: Film with a 2:1 ratio of FTO nanoparticles (light) to TiO2 spheres (dark)
There is however one main challenge when producing DSSCs using this technique. As
previously mentioned, no TCO must be in contact with the electrolyte due to shorting. As a
result, a post-treatment step must be done on the composite film to cover the nanoparticles with
a blocking layer of titania (see figure below). The challenge lies in the fact that this must be
done while maintaining porosity in the film as well as high surface area.
Figure 4-3: Co-cast TCO/titania sphere film with and without a blocking layer
52
Titania spheres have been reported in the literature, including by the assembly of nanoparticles,1
by sol-gel/hydrothermal processes,2 and via solvothermal routes.
3 Published applications for
these materials include enrichment of phosphorilated proteins,1 dye sensitized solar cells,
4
magnetic separation,5 lithium ion battery electrodes,
6 photocatalysis,
7 with many other
applications surely to follow. In many of these applications, the advantages of porous titania
spheres as compared to nanoparticles were due to light scattering 8 as well as the easy diffusion
of fluids through the sphere framework. For this study, the assembly of nanoparticles into
spheres was preferred over other strategies due to the ability to control pore sizes and sphere
sizes easily.
Assembling nanoparticles into various structures have been an interest of scientists for quite
some time ranging from inverse opal structures to chains and spheres and layered materials.9, 10
Due to their ease of formation and multiple uses, spherical assemblies of nanoparticles have
gathered a lot of attention and have come to include both hollow11
and dense spheres.12
Control
over nanoparticle synthesis and most importantly their surface properties, have allowed virtually
any type of nanoparticle to be assembled into spherical structures including metal oxides,13
chalcogenides,14
and fluorides using plates, rods and dots.15
Methods to produce spheres from
nanoparticles include emulsions,4 aerosol,
16 and microfluidics.
17 In this study, an emulsion
technique was used to produce nanoparticle spheres.
The basic concept of assembling nanoparticles into spheres involves the creation of an emulsion,
either a water-in-oil emulsion (W/O) or an oil-in-water emulsion (O/W). Titania nanoparticles
must be suspended in the droplet phase to produce the spheres. Various techniques were
attempted to produce emulsions from the two-component system including vigorous shaking,
blending and stirring. Once an emulsion is created, the solvent in the droplet phase must be
evaporated until the nanoparticles form a porous sphere.
53
4.2 Emulsions
Titania nanoparticle spheres produced via an emulsion technique have previously been
described elsewhere for titania nanoparticles.13, 18
In this work, nanoparticles in an oil phase are
stabilized within an emulsion droplet suspended in a water phase. Key factors affecting stability
of the emulsions and nanoparticles include the nanoparticle ligands, the surfactant used to
stabilize the emulsion as well as the concentration of the nanoparticles and ligands as well as the
choice of solvents. Nanoparticles on their own often behave as an emulsifier and thus often
complicate the production of emulsions due to their attraction to interfaces.19, 20
The stability of
emulsions are usually simply judged by eye by assessing their “long term” behaviour ranging
from minutes to years. There are two main types of instability during emulsion formation:
creaming and coalescence. Coalescence occurs when small droplets combine to form
progressively larger droplets. Creaming occurs when two liquids of different densities separate
resulting in the less dense material rising to the top.
Trapping particles within an emulsion is challenging due to the fact that particles are often
drawn to the interface of fluids. Microparticles have been known to stabilize the interface
between immiscible fluids for quite some time.21
Pieranski noted that the activation energy for
particle detachment at interfaces scales quadratically with the colloidal radius – for example, the
detachment energy for polystyrene particles at the air/water interface is roughly 107 times higher
than thermal energy. Repulsion among particles must be retained during this process if one is to
obtain a stable system. On the other hand, nanoparticle stability at interfaces is much lower than
microparticles, on the order of 10-100 times the thermal energy. This creates additional
challenges (no vigorous stirring, no intense heating) when attempting to assemble nanoparticles
at interfaces and within emulsion droplets.22
In addition to this, particles in emulsions or two-
phase systems are drawn to the interface and thus very often produce hollow structures.23
The nanoparticles themselves can initially be suspended in a water phase or an oil phase. During
this research project, both techniques were attempted with success. One of the greatest
challenges was ensuring that nanoparticles did not migrate from one phase to the other. This was
further complicated by the fact that to produce an emulsion, surfactants are required. It was
found that these surfactants often allowed the migration of nanoparticles from the oil phase to
the water phase and vice-versa due to the capping of the nanoparticles by the surfactants. The
success of the project was thus largely determined by the selection of adequate surfactants,
54
namely surfactants which are only miscible within one phase to prevent migration of
nanoparticles to the other phase.
The HLB number (Hydrophobic/Lypophilic Balance) is a key component one must be aware of
when producing emulsion droplets.24
This number which ranges from 0-20 is an indicator of the
ratio of hydrophobic/hydrophilic head groups of the surfactant molecules and determines what
type of emulsion can be expected. For example, a lower HLB number signifies that there is a
greater proportion of hydrophobic head groups, thus these molecules are generally oil-soluble
and form large water droplet emulsions within an oil phase, and the contrary is true for high
HLB numbers. In the middle range of HLB numbers, the surfactants can solubilise in either
component or generally form very small emulsions (< 50 nm). In this study, Hypermer 2296
was used as the surfactant. It is important to note that Hypermer 2296 is a very low HLB
surfactant which is completely insoluble in the water phase, a property which was not seen using
any other surfactant. It is believed that this allowed the facile creation of nanoparticle spheres
due to the fact that the surfactant could not enter the water phase. This effectively eliminated
nanoparticle migration from the water phase to the oil phase.
In this study, emulsions were prepared by diluting Aearodisp AW 740X titania solution
purchased from Evonik Industries to 20wt%. A solution of 5wt% Hypermer 2296 in hexadecane
were mixed together to form the oil phase. The aqueous solution was then added via a pipette to
the oil phase such that the water phase resulted in a 20vol % solution. The solution was then
blended with a hand blender (Oster blender). After blending, the solution was heated to 70˚C
under intense stirring until the water phase evaporated. The as-synthesized spheres were
centrifuged and re-dispersed in ethanol (Figure 4-4). At this stage, the nanoparticle spheres are
stable and are able to withstand sonication and shaking. The resulting spheres have a fairly
uniform size distribution centered around 500 nm as can be seen in Figure 4-5.
55
Figure 4-4: SEM image of nanoparticle spheres
Figure 4-5: Size distribution of the nanoparticle spheres
A very important parameter in this study is the porosity of the spheres. In a typical DSSC film,
the surface area and average pore size of the films are approximately 193m2/g.
18 BET was
performed on the spheres and results proved that the pore size within the spheres was centered
around 40 nm and possessed a specific surface area of 59.6m2/g (see Figure 4-6). In standard
DSSCs, the much higher surface area could be due to the use of surfactants and porogens used
during the making of the films as well as the smaller particle sizes used. In these experiments,
no porogen were used since this was observed to influence the quality of sphere formation. The
particle sizes used in these experiments were of 27 nm ± 5.5 nm compared to particles of 10-20
0
5
10
15
20
25
30
Fre
qu
en
cy
Sphere size range (nm)
56
nm used in state of the art DSSCs. As previously mentioned the size of the nanoparticles play a
huge role in the efficiency of the devices and should be modified if one is to optimize the system
with respect to porosity and nanoparticle size.
Figure 4-6: Pore size distribution as determined by BET
FTO nanoparticles were chosen as the TCO phase, purchased from Keeling Walker. The size of
these particles was measured to be in the range of 30 to 100 nm as seen by the SEM image
shown in Figure 4-7 below. PXRD was performed on the particles, showing a pure SnO2
Casserite structure shown in Figure 4-8.
Figure 4-7: SEM Image of FTO nanoparticles (light) and a TiO2 nanoparticle sphere
(dark)
57
Figure 4-8: PXRD of FTO nanoparticles
4.3 Preventing shorting
As mentioned earlier, a blocking layer is required to prevent shorting within the DSSCs. This
was attempted via ALD as well as by TiCl4 treatments. ALD was performed via the alternate
exposure of TiCl4 and H2O vapours to the substrate (Figure 4-9). Nitrogen gas was used to carry
H2O vapour or TiCl4 to the sample for a period of 10 seconds using a flow of 1-2
bubbles/second. After filling the chamber with the gas, the chamber is cycled three times
between a vacuum and nitrogen gas to ensure no residual gases within the chamber. PXRD was
attempted on the ALD samples although these were too thin to produce detectable peaks. Each
ALD cycle was found to deposit around 8 nm of titania on the substrate via ellipsometry.
The other technique to prevent shorting was a modified treatment often used in DSSCs – namely
the use of a heated aqueous solution of TiCl4. This treatment is often used to increase the current
in DSSCs, by providing nanometer texturing on the TiO2 nanoparticles as well as increasing the
interconnectivity of the nanoparticles.25
In these experiments, a 2M TiCl4 solution in HCl was
diluted to 0.05M before use. This solution was kept in the freezer until ready to use. The films
were placed in this solution at 95-105˚C for 0-45 minutes, resulting in a hazy solution and a
TiCl4 coating on the substrate.
58
Figure 4-9: TiCl4 ALD system
It was determined that the TiCl4 solution treatment began to deposit TiO2 around 20 minutes
after being placed in the oven. At this point, the solution turned from clear to cloudy within a
minute as seen in Figure 4-10. This growth was confirmed via SEM as well as electrically. To
determine the optimal TiCl4 treatment time, the resistance of the TiO2 films on an FTO substrate
were determined as a function of TiCl4 treatment time (Figure 4-11) using a two-point probe
technique. Measurements were performed as depicted in Figure 4-12. From these experiments, it
was determined that the ideal treatment time was between 20-32 minutes. PXRD was performed
on the TiO2 deposited by TiCl4 which indicated an Anatase phase although there was more
Rutile than the pure TiO2 nanoparticles (Figure 4-13). An SEM image of the nanoscale growth
resulting from this treatment is presented in Figure 4-14. A sphere sample with TiCl4 post-
treatment is presented in Figure 4-15.
TiO2 sol was also attempted as a blocking layer. This solution was prepared by adding a solution
of ethanol and water to tetrabutyl othotitanate and diethanolamine in ethanol.26
The resulting
solution was spin-coated onto a film of titania spheres of the FTO/TiO2 composites.
59
Figure 4-10: Image of TiCl4 solution before and after 25 minutes at 100˚C
Figure 4-11: Resistance of FTO substrates as a function of TiCl4 treatment time
Figure 4-12: Method used to perform the conductivity measurements
1
10
100
1000
10000
100000
1000000
10000000
100000000
0 10 20 30 40
Resi
stan
ce (
oh
m)
TiCl4 treatment time
60
Figure 4-13: PXRD of the TiO2 powder, spheres and TiCl4 treated spheres
Figure 4-14: TiO2 nanocrystals resulting from TiCl4 treatment
61
Figure 4-15: Sphere film after TiCl4 treatment
4.4 Results and Discussion
Preliminary experiments were performed to test whether or not a TCO nanoparticle framework
had the effect of decreasing the electron collection time. In these experiments, reference cells
were first tested using titania sphere films, the results of which are presented below. The cell #
represents the TiCl4 treatment time, wherein 0 is no treatment and cells 1 to 7 range from 20 to
32 minutes of treatment in 2 minute intervals.
Figure 4-16: Efficiency, Fill Factor, Short Circuit Current and Open Circuit Voltage for
TiO2 sphere films
62
Identical experiments were performed on the composite films with a 2:1 weight ratio of
FTO:TiO2, presented below in Figure 4-17.
Figure 4-17: Efficiency, Fill Factor, Short Circuit Current and Open Circuit Voltage for
FTO/TiO2 sphere films
Spheres composed of titania nanoparticles performed quite well in DSSCs. Upon TiCl4
treatment, it can be observed the efficiency of the cells increases from 3.8% to 5.6%. The source
of this increase is likely due to the increased current density from the higher surface area
resulting from the TiCl4 treatment process. Both the voltage and fill factor decrease as a result of
the process, most likely due to the densification of the film upon prolonged exposure. All
components other than the current density continue decreasing as more TiO2 is deposited within
the film via TiCl4.
63
As for the FTO/TiO2 films, the trends are similar with one notable exception. The voltage in
these cells jumps from 600mV to 720mV after TiCl4 treatment. A possible explanation for the
higher Voc is a faster electron collection time, the effect of which is described in Chapter 2.
Upon TiCl4 treatment, the insulating effect of TiO2 might begin to prevent recombination, thus
allowing the FTO nanoparticles to shuttle electrons to the substrate. None of these cells
outperformed the reference sphere cells.
The results of ALD and TiO2 sol treatment are presented below, with the results showing
negligible increases in efficiency due to these treatments. A more complete study of ALD
treatment including a series of cycles as well as TiO2 sol are still under investigation.
Figure 4-18: Efficiency of FTO-TiO2 nanoparticle-sphere DSSC with post-treatments of
ALD and TiO2 sol
Figure 4-19 presented below, presents the IV curves for the sphere samples and the composite
samples. The effect of shunt resistances and series resistance can be observed from this figure,
as outlined in Chapter 1.4. For the titania sphere samples, the TiCl4 treatment reduces the series
resistance in the cell remarkably. This is most likely due to the higher interconnectivity resulting
from the TiCl4 treatment (see Figure 4-15). On the other hand, the series resistance does not
show any improvement in the composite film after TiCl4 treatment. Overall, the series
resistances for the composite films were worse than the titania sphere samples. This suggests
that the efficiency of electron collection through the TCO framework is inefficient since good
electron shuttling would result in a lower series resistance relative to the titania sphere film. The
shunt resistances in all of these cells were comparable, suggesting that the problem lies in the
electron transport and not in recombination.
64
Figure 4-19: IV curves of the TiO2 spheres (solid) and the FTO/TiO2 films (dashed) before
and after TiCl4 treatment
What is next? These very early results present the basis for future work on this project. At this
stage, there are a number of parameters which need to be modified and optimized to obtain
optimal results. Three parameters of interests should be modified: film thickness, TiO2/FTO
ratio and TiCl4 or ALD treatment time. In addition, the sizes of particles used in both the TiO2
phase and the FTO phase should be modified and optimized to ensure maximum surface area as
well as porosity within the framework. The conductivity of FTO should also be determined.
Finally, impedance spectroscopy should be incorporated into to this project to gain a better
understanding of the various processes involved and actually determine the lifetimes within
DSSCs containing an extended current collector.
In this chapter, the steps and treatments for producing a scalable and reproducible all-in-one
integrated DSSC are presented. Although much work remains to be done towards the
optimization of the film, the concepts presented and the experimental techniques form the basis
for these future experiments.
65
References
(1) Lu, Z.; Ye, M.; Li, N.; Zhong, W.; Yin, Y. Self-Assembled TiO2 Nanocrystal Clusters for
Selective Enrichment of Intact Phosphorylated Proteins. Angew. Chem. Int. Edit. 2010, 49,
1862-1866.
(2) Chen, D.; Huang, F.; Cheng, Y.; Caruso, R. A. Mesoporous Anatase TiO2 Beads with High
Surface Areas and Controllable Pore Sizes: A Superior Candidate for High-Performance
Dye-Sensitized Solar Cells. Adv Mater 2009, 21, 2206-2211.
(3) Yang, W.; Li, J.; Wang, Y.; Zhu, F.; Shi, W.; Wan, F.; Xu, D. A facile synthesis of anatase
TiO2 nanosheets-based hierarchical spheres with over 90% {001} facets for dye-sensitized
solar cells. Chem. Commun. 2011, 47, 1809-1811.
(4) Chen, D.; Huang, F.; Cheng, Y.; Caruso, R. A. Mesoporous Anatase TiO2 Beads with High
Surface Areas and Controllable Pore Sizes: A Superior Candidate for High-Performance
Dye-Sensitized Solar Cells. Adv Mater 2009, 21, 2206-2211.
(5) Lu, Z.; He, L.; Yin, Y. Superparamagnetic nanocrystal clusters for enrichment of low-
abundance peptides and proteins. Chem. Commun. 2010, 46, 6174-6176.
(6) Wang, J.; Zhou, Y.; Hu, Y.; O'Hayre, R.; Shao, Z. Facile Synthesis of Nanocrystalline TiO2
Mesoporous Microspheres for Lithium-Ion Batteries. J. Phys. Chem. C 2011, 115, 2529-
2536.
(7) Li, Y.; Chen, G.; Wang, Q.; Wang, X.; Zhou, A.; Shen, Z. Hierarchical ZnS-In2S3-CuS
Nanospheres with Nanoporous Structure: Facile Synthesis, Growth Mechanism, and
Excellent Photocatalytic Activity. Adv. Funct. Mater. 2010, 20, 3390-3398.
(8) Huang, F.; Chen, D.; Zhang, X. L.; Caruso, R. A.; Cheng, Y. Dual-Function Scattering
Layer of Submicrometer-Sized Mesoporous TiO2 Beads for High-Efficiency Dye-
Sensitized Solar Cells. Adv. Funct. Mater. 2010, 20, 1301-1305.
(9) Davis, S. A.; Breulmann, M.; Rhodes, K. H.; Zhang, B.; Mann, S. Template-directed
assembly using nanoparticle building blocks: A nanotectonic approach to organized
materials. Chemistry of Materials 2001, 13, 3218-3226.
(10) Marlow, F.; Dong, W. Engineering nanoarchitectures for photonic crystals.
ChemPhysChem 2003, 4, 549-554.
(11) Kim, Y. J.; Chai, S. Y.; Lee, W. I. Control of TiO2 structures from robust hollow
microspheres to highly dispersible nanoparticles in a tetrabutylammonium hydroxide
solution. Langmuir 2007, 23, 9567-9571.
(12) Yang, W.; Wan, F.; Chen, Q.; Li, J.; Xu, D. Controlling synthesis of well-crystallized
mesoporous TiO2 microspheres with ultrahigh surface area for high-performance dye-
sensitized solar cells. J. Mater. Chem. 2010, 20, 2870-2876.
66
(13) Lu, Z.; Ye, M.; Li, N.; Zhong, W.; Yin, Y. Self-assembled TiO2 nanocrystal clusters for
selective enrichment of intact phosphorylated proteins. Angewandte Chemie - International
Edition 2010, 49, 1862-1866.
(14) Harrison, R. G.; Washburn, A. L.; Pickett, A. T.; Call, D. M. Assembly of CdSe
nanoparticles into microspheres by a liquid droplet emulsion process. Journal of Materials
Chemistry 2008, 18, 3718-3722.
(15) Bai, F.; Wang, D.; Huo, Z.; Chen, W.; Liu, L.; Liang, X.; Chen, C.; Wang, X.; Peng, Q.; Li,
Y. A versatile bottom-up assembly approach to colloidal spheres from nanocrystals.
Angewandte Chemie - International Edition 2007, 46, 6650-6653.
(16) Vasiliev, P. O.; Faure, B.; Ng, J. B. S.; Bergstroam, L. Colloidal aspects relating to direct
incorporation of TiO2 nanoparticles into mesoporous spheres by an aerosol-assisted
process. J. Colloid Interface Sci. 2008, 319, 144-151.
(17) Xu, S. Q.; Nie, Z. H.; Seo, M.; Lewis, P.; Kumacheva, E.; Stone, H. A.; Garstecki, P.;
Weibel, D. B.; Gitlin, I.; Whitesides, G. M. Generation of monodisperse particles by using
microfluidics: Control over size, shape, and composition. Angew. Chem. -Int. Edit. 2005,
44, 724-728.
(18) Yang, W.; Wan, F.; Chen, Q.; Li, J.; Xu, D. Controlling synthesis of well-crystallized
mesoporous TiO2 microspheres with ultrahigh surface area for high-performance dye-
sensitized solar cells. J. Mater. Chem. 2010, 20, 2870-2876.
(19) Binks, B. P. Particles as surfactants - Similarities and differences. Current Opinion in
Colloid and Interface Science 2002, 7, 21-41.
(20) Aveyard, R.; Binks, B. P.; Clint, J. H. Emulsions stabilised solely by colloidal particles.
Adv. Colloid Interface Sci. 2003, 100, 503-546.
(21) Pieranski, P. Two-dimensional interfacial colloidal crystals. Phys. Rev. Lett. 1980, 45, 569-
572.
(22) Bresme, F.; Oettel, M. Nanoparticles at fluid interfaces. Journal of Physics Condensed
Matter 2007, 19.
(23) Eun, T. H.; Kim, S. -.; Jeong, W. -.; Jeon, S. -.; Kim, S. -.; Yang, S. -. Single-step
fabrication of monodisperse TiO2 hollow spheres with embedded nanoparticles in
microfluidic devices. Chemistry of Materials 2009, 21, 201-203.
(24) Boyd, J.; Parkinson, C.; Sherman, P. Factors affecting emulsion stability, and the HLB
concept. J. Colloid Interface Sci. 1972, 41, 359-370.
(25) Sommeling, P. M.; O'Regan, B. C.; Haswell, R. R.; Smit, H. J. P.; Bakker, N. J.; Smits, J. J.
T.; Kroon, J. M.; Van Roosmalen, J. A. M. Influence of a TiCl4 post-treatment on
nanocrystalline TiO2 films in dye-sensitized solar cells. J Phys Chem B 2006, 110, 19191-
19197.
67
(26) Yu, H.; Zhang, S.; Zhao, H.; Will, G.; Liu, P. An efficient and low-cost TiO2 compact layer
for performance improvement of dye-sensitized solar cells. Electrochim. Acta 2009, 54,
1319-1324.
68
5 Conclusion
In this Master‟s thesis, we have laid the foundation for the next generation of Dye Sensitized
Solar Cells with improved charge transport characteristics through the development of a three-
dimensional transparent conducting oxide current collector.
This thesis documents the progress towards achieving this goal as well as the challenges
associated with the various techniques. Three main nanochemistry strategies were presented
towards achieving this goal. A templated approach was first tackled, resulting in a multitude of
challenges associated with infiltration, interconnectivity and scalability. A non-templated
approach was thus developed, resulting in an integrated DSSC in which the current collector is
extended into the working electrode.
Although the material challenges and the techniques to produce the material have been
established, much work remains towards the optimization of this system. A number of
parameters have yet to be optimized such as particle size, ratios of TCO and semiconductor and
blocking layers. In addition, the project will require the use of AC impedance spectroscopy to
determine the effect of having TCO nanoparticles within these films. These improvements will
not only lead to a much better understanding of the various parameters involved in DSSCs but
could also lead to record breaking efficiency cells. Extending the current collector has been
slowly gaining popularity and should play an ever increasing role in the production of next
generation solar energy devices.