Electronic and Molecular Surface Structures of Dye-Sensitized TiO2 Interfaces

(1)

UNIVERSITATISACTA UPSALIENSIS

Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 753

Electronic and Molecular Surface Structures of Dye-Sensitized TiO ₂ Interfaces

MARIA HAHLIN

ISSN 1651-6214

(2)

Dissertation presented at Uppsala University to be publicly examined in Polhemsalen, Ångström Laboratory, Lägerhyddsvägen 1, Uppsala, Friday, September 10, 2010 at 13:15 for the degree of Doctor of Philosophy. The examination will be conducted in English.

Abstract

Hahlin, M. 2010. Electronic and Molecular Surface Structures of Dye-Sensitized TiO2

Interfaces. Acta Universitatis Upsaliensis. Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 753. 77 pp. Uppsala.

The dye-sensitized solar cell is a promising solar cell technology. In these systems the key process for light to electricity conversion is molecular in nature and is initiated in dye molecules adsorbed at a semiconducting surface. This thesis focuses on the electronic and molecular surface structure of the dye/TiO2 interface, and the experimental results were obtained from surface sensitive X-ray based electron spectroscopic methods.

Two families of dyes, triarylamine based organic dyes and ruthenium based inorganic dyes, were investigated. The effect of dye structural modications on the interfacial properties was studied, such as the surface concentrations, dye molecular surface orientation, molecular interactions, and energy level matching. Also, the impact of additional parameters such as the incorporation of coadsorbents and the solvents used for dye sensitization were studied and complementary photoelectrochemical characterization was used to demonstrate functional properties corresponding to changes in the molecular layers.

The experiments provided information on how specic structural modications change the frontier electronic structure. The results also showed that the adsorption of the organic dye leads to submolecular electronic changes, and that the dye surface orientations in general favor effcient energy conversion. Moreover, effects of solvents and coadsorbents, on both energy level matching between the dye and the TiO2 substrate and the surfacemolecular structure were quantied.

Keywords: Photoelectron spectroscopy, Electron spectroscopy, Solar Cells, Dye, TiO2, Surface, Interface, Electronic structure, Molecular surface structure, coadsorbent

Maria Hahlin, Department of Physics and Astronomy, Molecular and Condensed Matter Physics, 516, Uppsala University, SE-751 20 Uppsala, Sweden

urn:nbn:se:uu:diva-127166 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-127166) ISBN 978-91-554-7847-6.

(3)

I’m sitting here writing my thesis on a sunny day in April. All the snow has just melted and the ground is still brown, but on a closer look some green grass can be seen in what shall in a months time be a green carpet. My husband and two daughters are enjoying the sunny outside, cleaning up old leafs that fell to the ground last fall, and looking for the sign of new flowers popping up. What great things the sun gives!

to Anders

(4)

(5)

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

I Electronic and Molecular Structures of Organic Dye/TiO₂ Interfaces for Solar Cell Applications, a Core Level Photoelectron Spectroscopy Study Maria Hahlin, Erik M.J. Johansson, Stefan Plogmaker, Michael Odelius, Daniel P Hagberg, Licheng Sun, Hans Siegbahn, Håkan Rensmo

Physical Chemistry Chemical Physics, 2010, 12, 1507 - 1517

II Mapping the frontier electronic structures of triphenylamine based organic dyes at TiO₂interfaces

Maria Hahlin, Michael Odelius, Martin Magnusson, Erik M.J. Johansson, Stefan Plogmaker, Daniel P Hagberg, Licheng Sun, Hans Siegbahn, Håkan Rensmo

Submitted to Physical Chemistry Chemical Physics

III Rhodanine dyes for dye-sensitized solar cells : spectroscopy, energy levels and photovoltaic performance

Tannia Marinado, Daniel P. Hagberg, Maria Hedlund, Tomas Edvinsson, Erik M. J. Johansson, Gerrit Boschloo, Håkan Rensmo, Tore Brinck, Licheng Sun and Anders Hagfeldt

Physical Chemistry Chemical Physics, 2009 (11) 1, 133-141

IV Surface molecular quantification and photoelectrochemical characterization of mixed organic dye and coadsorbent layers on TiO₂ for dye-sensitized solar cells

Tannia Marinado, Maria Hahlin, Xiao Jiang, Maria Quintana, Erik. M. J.

Johansson, Erik Gabrielsson, Stefan Plogmaker, Daniel P. Hagberg, Gerrit Boschloo, Shaik M. Zakeeruddin, Michael Grätzel, Hans Siegbahn, Licheng Sun, Anders Hagfeldt, and Håkan Rensmo

Journal of Physical Chemistry C 2010, 114, 11903–11910

V Surface compostions of dye/TiO₂ interfaces formed from ethanol, acetonitrile, and dichloromethane based solutions

Maria Hahlin, Tannia Marinado, Maria Quintana, Erik M.J. Johansson, Hans Siegbahn, Anders Hagfeldt, Håkan Rensmo

In Manuscript

(6)

VI Electronic and Molecular Surface Structure of Ru(tcterpy)(NCS)(3) and Ru(dcbpy)(2)(NCS)(2) Adsorbed from Solution onto Nanostructured TiO₂: A Photoelectron Spectroscopy Study

Johansson E.M.J., Hedlund M, Siegbahn H, Rensmo H, Journal of Physical Chemistry B, 2005, 109 (47), 22256-22263

VII Frontier electronic structures of Ru(tcterpy)(NCS)(3) and Ru(dcbpy)(2)(NCS)(2): A photoelectron spectroscopy study

Johansson E.M.J., Hedlund M, Odelius M, Siegbahn H, Rensmo H Journal of Chemical Physics, 2007, 126 ( 24), 244303

VIII The influence of water on the electronic and molecular surface structures of Ru-dyes at nanostructured TiO₂

Maria Hahlin, Erik Johansson, Rebecka Schölin, Hans Siegbahn, Håkan Rensmo

Submitted to Journal of Physical Chemistry C

IX A photoelectron spectroscopy study of Z-907 co-adsorbed with DPA on nanostructured TiO₂surfaces

Maria Hahlin, Erik Johansson, Stefan Plogmaker, Hans Siegbahn, Håkan Rensmo

In Manuscript

Reprints were made with permission from the publishers.

The following is a list of papers and communications to which I have contributed but that are not a part of this thesis.

Using a molten organic conducting material to infiltrate a nanoporous semiconductor film and its use in solid-state dye-sensitized solar cells Fredin K, Johansson EMJ, Blom T, Hedlund M, Plogmaker S, Siegbahn H, Leifer K, Rensmo H

Synthetic Metals, 2009, 159 (1-2), 166-170

Modification of Nanostructured TiO₂Electrodes by Electrochemical Al3+

Insertion: Effects on Dye-Sensitized Solar Cell Performance

Hugo Alarcón, Maria Hedlund, Erik M. J. Johansson, Håkan Rensmo, Anders Hagfeldt, and Gerrit Boschloo

J. Phys. Chem. C, 2007, 111 (35), 13267-13274

Photovoltaic and interfacial properties of heterojunctions containing dye- sensitized dense TiO₂and tri-arylamine derivatives

Johansson EMJ, Karlsson PG, Hedlund M, Ryan D, Siegbahn H, Rensmo H Chemistry of Materials, 2007, 19 (8), 2071-2078

Segregation and Interdiffusion in (Fe,Co)/Pt superlattices

M. Björck, G. Andersson,B. Sanyal, M. Hedlund, and A. Wildes, Phys. Rev.

B, 2009, 79 (8), 085428

(7)

Magnetic moments in Fe– Co/Pt superlattices M. Björck, M. Hedlund, G. Andersson

Journal of Magnetism and Magnetic Materials, 2008, 320 2660– 2664 Surface-Confined Photopolymerization of pH-Responsive Acrylamide/

Acrylate Brushes on Polymer Thin Films

Gunnar Dunér, Henrik Anderson, Annica Myrskog, Maria Hedlund, Teodor Aastrup, and Olof Ramström, Langmuir 2008, 24, 7559-7564

The Role of Colloid Formation in the Photoinduced H₂Production with a RuII PdII Supramolecular Complex: A Study by GC, XPS, and TEM Pengxiang Lei, Maria Hedlund, Reiner Lomoth, Håkan Rensmo, Olof Johans- son, and Leif Hammarström, J. Am. Chem. Soc., 2008, 130 (1), 26-27

(8)

Comments on my own participation

The papers presented in this thesis are based upon a teamwork and the extent of my contribution is to some extent reflected by my position in the author list. In paper I, II, V, VIII, and IX, I had the main responsibility of the projects, including the PES experimental work, the data analysis and preparation of the manuscript. In paper VIII I also performed the photoelectrochemical experiments and analysis. In paper III and IV, I performed the PES experiments, was responsible for the PES discussions of the manuscript, and took part in the general analysis of the work. In paper VI and VII I took part in the experimental measurements, the discussions and assisted with the manuscript.

(9)

1. Introduction

1.1 Renewable energy

Our society today is dependent on easily available energy. For a long time the world has depended on fossil fuels for providing this energy, which has resulted in the rapid increase of greenhouse gases (CO₂, CH₄, and N₂O) in the atmosphere.¹The climate is a complex system and the parameters involved in determining the earths temperature are closely dependent on each other and therefore the effect of increased amount of greenhouse gases in the atmosphere is difficult to predict. The human enforced changes may result in a negative spiral in the existing temperature balance, where one parameter is feeding the other, and if let to far may be hard to break.

An important and frequently discussed topic in politics today is the climate changes and how to best minimize the discharge of greenhouse gases, with the aim of minimiz- ing global warming. The EU has set a goal that by year 2020 20% of the energy used in EU should come from renewable energy sources, where depending on the situation today different countries within EU has different goals. In addition to this 10% of the energy used in the transport sector shall also come from renewable energy sources.

The percentage of renewable energy in Sweden is currently rather high compared to the rest of EU. In year 1990 the share of renewable energy in Sweden was 34%, and in year 2008 it had increased to 44%, and the Swedish government has set the goal that by year 2020 50% of the used energy in Sweden shall come from renewable energy sources. As of today the dominating contribution of renewable energy sources originates from wood, water power, uptake from heat pump, organic waste, bio fuels and wind power.²

Solar cells is a renewable energy source that currently only delivers a small fraction of the used energy in the world. The solar energy reaching the earth each hour is however more than the yearly energy consumption in the world,³and thus this energy source has enormous potential for meeting future energy demand.

1.2 Solar Cells

Solar cells convert sunlight directly into electricity. The absorption of photons, i.e.

absorption of sunlight, is made through the excitation of electrons in the solar cell.

This process can only take place if the energy of the photon match an allowed energy transition in the cell. In order to make a well functioning solar cell the energy spectrum from sunlight and the allowed energy transitions in the solar cell need to match.

The sunlight spectrum, i.e. the blackbody radiation, is linked to the temperature of the sun (5778 K), and the emitted spectrum is schematically pictured in figure 1.1. The atmosphere on earth prevent parts of the light intensity emitted from the sun to reach

(12)

500 1000 1500 2000 2500 Wavelength / nm

Spectral Irradiance / W/m2/nm 1 2

Blackbody radiation from Sun Radiation at sea level

Figure 1.1:Schematic picture of the spectral irradiance emitted from the sun, and the spectral irradiance reaching the surface of the earth. Schematic picture drawn from other original version.⁴

the earths surface, and a schematically drawn picture of the spectral irradiance reaching the surface of the earth is also included in figure 1.1. For solar cell applications on earth this light that can be utilized for energy conversions to electricity.

The perfect solar cell should ideally absorb all incident photons and convert these to electrical energy without internal energy losses, however in reality this is impossible to achieve. The inability to absorb photons with lower energy than the band gap, i.e.

the lowest electronic energy transition, and the inability to convert all energy from the absorbed photons with higher energy than the band gap to electrical energy put theoretical limitations to maximal achievable efficiency.

The most common solar cell on the market today is the silicon based solar cell, and the record efficiency for such a silicon solar cell is about 25%,⁵which is close to the maximum achievable efficiency for single band gap solar cells. However there are currently various other types of solar technologies with properties well matching the solar spectrum and some are mentioned here. The thin film solar cells, for example CIGS (containing copper, indium, gallium and selenium) and CdTe (containing cad- mium and tellurium), have an efficiency record of 20% and 16.7% respectively.⁵The lower production costs for the thin film solar cells compared to the silicon based solar cells make them currently good competitors for market shares. Organic materials can also be used for solar cells. One type is purely organic and has reached an efficiency of about 7.9 %.⁵In these cells much effort today concerns developing materials with a gap that better match the solar spectrum.

Another technology called dye sensitized solar cells (DSC) mixes organic and inorganic materials. The DSC, schematically drawn in figure 1.2, had a major break- through when O’Regan and Grätzel reached 7.1% efficiency in 1991, through the adsorption of dye molecules on a nanostructured oxide semiconductor network.⁶The nanostructured system dramatically increased the surface area of the TiO₂and with this the photoactive interface of the solar cell. Since 1991 advances have been made and the current record efficiency is 11.1%.⁷The DSC is a promising solar cell technology for low cost production due to both the low material cost and the potential for easy manufacturing methods.^6,8–12Much effort in DSC development today concerns 12

(13)

νν

Figure 1.2:Schematic picture of a dye sensitized solar cell.

the development of new materials for increased efficiency and stability. Moreover much of the research concern understanding the conversion at the nanoscale (e.g.

charge transport) a well as understanding the conversion at the molecular level and how the material matching and gap for light absorption (i.e. the dye absorption spectrum) limits the efficiency. At the heart of this understanding is the dye-sensitized semiconductor interface.

1.3 The aim of this thesis

The aim of this thesis is to characterize the interfacial region containing dye molecules and a TiO₂substrate. A specific focus is placed on the impact of structural modifica- tions of the dye on the electronic and molecular surface structures. Included is also the impact of additional parameters such as the incorporation of coadsorbents and solvents used for dye sensitization.

Primarily electron spectroscopic methods are used as characterization technique.

These methods are surface sensitive and well suited to study the dye/TiO₂ interface. They provide molecular level information concerning e.g. the surface concentrations, the dye molecular surface orientations, molecular interactions, and energy level matching in these photoactive interfaces. Complementary photoelectrochemical characterization is used to demonstrate functional properties corresponding to changes in the molecular layers.

(14)

(15)

2. The System

The first part of this chapter presents a general introduction to the field of dye-sensitized solar cells. The second part describes more in detail the properties important for the experimental investigations presented in this thesis.

2.1 The dye sensitized solar cell

2.1.1 Materials in the dye-sensitized solar cell

The dye sensitized solar cell (DSC) is constructed out of three main parts, a working electrode, a counter electrode, and a redox or hole conducting molecular system, see figure 2.1.

TiO2 nanoparticle Dye TiO2 nanostructure

GlasF:SnO2 Pt

TiO2 F:SnO2 F:SnO2 TiO

F:SnO2 GlasF:SnO2 F:SnO2 TiO

F:SnO2 F:SnO2 TiO

F:SnO2 TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO TiO

I- I3-

Redox system

Working Electrode Counter Electrode

Figure 2.1:A schematic picture of the dye sensitized solar cell.

The working electrode often consists of a rigid glass substrate which on one side has a thin layer of an electron conducting material, typically F:SnO2. In contact with this conducting layer is a nanoporous semiconductor network. The most extensively used semiconductor material is TiO₂, although other materials such as ZnO, SnO₂, and Nb₂O₅may also be used.^13–15TiO₂being a wide band gap semiconductor only

(16)

absorb a small fraction of the incident light. To increase light absorption in the DSC the surface of the semiconductor is covered with photoactive molecules, generally referred to as dyes. An extensive variety of dyes are used for the DSC, and some of these are further discussed in section 2.2.2. A single layer of dye molecules does however only absorb a small fraction of the sun light, and therefore the internal surface area has been enlarged by the use of nanostructured TiO₂. The size of the TiO₂particles varies between 15 to 40 nm and the porosity of the network is approximately 50%.

The counter electrode, like the working electrode, may consists of a rigid glass substrate with a F:SnO₂layer on the side facing the interior of the DSC. Here a Pt layer is generally deposited on top of the conducting layer.

The redox mediator used in the most effective DSC is a liquid based acetonitrile electrolyte containing the I⁻/I⁻₃ redox couple.⁷Long term encapsulation of the liquid based redox mediator is a problem for the DSC and therefore other systems based on ionic liquids and solid state hole conductors have been developed.^9,16,17However as of today these solid state DSCs can not match the efficiency of the liquid based I⁻/I⁻₃ electrolyte solar cells.

2.1.2 Function

All solar cells convert light into electrical energy. This energy conversion requires that the solar cell must absorb light, separate the energy rich charges, and finally collect them in an outer circuit. The system must also regenerate itself.

A full cycle in a working DSC, see figure 2.2, starts when a photon is absorbed by a dye molecule. This process is followed by a fast injection of the excited electron into the TiO₂network, leaving the dye molecule in an oxidized state.^18–21In this way two charges, the electron in the semiconductor and the hole in the dye, is separated over the dye/semiconductor interface. The injected electron moves through the nanoporous network and is transferred to the F:SnO₂layer where it is collected and used for external work. At the same time electrons enter the DSC through the counter electrode, where the electrolyte accepts electrons by reducing I⁻₃ to I⁻. The I⁻/I⁻₃ redox couple is transported via diffusion between the working electrode and the counter electrode.

The oxidized dye is subsequently reduced by the I⁻thus regenerating the system and completing the full conversion cycle.²²

Increasing the incident photon to current conversion efficiency in the DSC is generally difficult by using multilayers of dye molecules on the semiconductor surface since electrical energy is only efficiently generated at the interface between the dye and the semiconductor. Instead the size of the active interface is increased by using the nanoporous structure. The internal surface area of a 10 µm thick TiO2nanoporous network of 50% porosity is approximately 1000 times larger than the area of the photoactive device.¹³

2.1.3 Energy level matching and kinetic competition

The processes involved in converting light into electrical energy in the DSC is partly governed by the energy characteristics of the system.^23–25 In figure 2.3 the energy characteristics of the semiconductor, the dye, and the redox mediator are schematically pictured as previously done by others.²⁶ The energy level characteristics are

16

(17)

I- I3-

I- I3- I3- e-

e-

e- e-

e-

Figure 2.2:An illustration of the full electronic path when sunlight is converted to electrical energy in the DSC.

represented by the semiconductor valence band and conduction band, by the the position of the highest occupied molecular orbital (HOMO) for the dye and the lowest unoccupied molecular orbital (LUMO) for the excited dye, as well as by the redox potential for the redox mediator in the electrolyte.

Conduction band

Valence band

HOMO LUMO

Redox potential Energy

E_F ₁ ₄

2

5 6

3

Semiconductor Dye Electrolyte

Figure 2.3:The energy level diagram over the solar cell process.

Electron transfer from a lower energy level to a higher energy level can only occur through input of energy, for instance by the absorption of a photon, whereas electron transfer in the opposite direction can occur spontaneously. Different charge transfer processes are shown in figure 2.3. In the DSC conversion cycle electron transfer through input of light energy occurs when electrons are excited in the dye molecules.

This is depicted as a transfer from the HOMO to the LUMO (1), or often as the ground state and excited state redox potentials. In the conversion cycle two spontaneous elec-

(18)

tron transfers without the input of energy thereafter follows, the injection of electrons from the excited dye to the conduction band of the TiO₂(2) and the reduction of the oxidized dye by the redox mediator (3). In the DSC the requirements for electron transfer put restrictions on the material combinations used. A general picture is that the oxidation potential of the excited dye must be higher compared to the conduction band edge of the TiO₂and that the redox potential of the oxidized dye molecules must be lower than the redox potential of the electrolyte in order to allow for these electron transfers to occur.

In the DSC also unwanted spontaneous electron transfer processes can occur and these processes prevent absorbed photons to generate a current or add to the photopo- tential. Examples of unwanted processes are direct de-excitation of the excited dye molecule (4) or recombination of the generated electron hole pair through electron transfer from the conduction band of the semiconductor to either the HOMO level of the oxidized dye (5) or to I⁻₃ species in the electrolyte (6), see figure 2.3.

To secure an efficient solar cell the photocurrent generating processes must be sufficiently faster compared to the unwanted processes.^27,28 In this respect the charge transfer process between the excited dye and the semiconductor competes with the direct de-excitation of the excited dye. Similarly, the reduction of the oxidized dye competes with electron hole pair recombination through electron transfer from the conduction band in the semiconductor and the HOMO level of the oxidized dye, and the transport of electrons in the semiconductor competes with both recombination mechanisms (5, 6).²⁸

2.1.4 The efficiency of the DSC

The efficiency of a solar cell is often used for comparing different solar cells. This efficiency is defined as the ratio between the electrical power generated and the power of the incident light. The generated power is described by the product between the photovoltage over the cell at open circuit (V_oc), the generated current density at short circuit (J_sc), and the fillfactor (ff) (explained in section 3.2).

Partly determining the generated electrical power is thus the electrical potential difference between the working electrode and the counter electrode. This potential depend on the difference between the quasi fermi level of electrons in the TiO₂semiconductor (E_F) and the redox potential of the electrolyte, see figure 2.3. Additional resistance losses in the DSC will decrease the potential over the cell. The two basic parameters limiting maximum photovoltage are the position of the conduction band edge relative to the redox mediator and the rate constant constraints (section 2.1.3).²⁹ Important for a high photovoltage is low recombination rate which depend on the spatial separation of the TiO₂and the electrolyte/holeconductor material. At areas where the TiO₂is completely exposed to the electrolyte the local potential over this interface will move towards equilibrium, and this will result in an reduction of the photovoltage over the DSC. Thus the formation of a rather dense molecular layer is important for maintaining high voltages in the DSC.

The generated photocurrent density is dependent on the light harvesting efficiency, the injection efficiency, the collection efficiency, and the regeneration efficiency.

These parameters in turn depends on factors like the dye spectral response, the ability of the dye to form monolayers on the TiO₂ surface, the internal electronic

18

(19)

structure of the dye, the dye coverage, the mobility of the electrons in the TiO₂, the recombination rate of electrons in the TiO₂, and the adsorption configuration of the dye at the TiO₂surface.

Most of these parameters described above are much dependent on the interfacial molecular structure between the dye and the TiO2, and a fundamental understanding of such properties is important for understanding and optimizing the energy conversion in a DSC.

2.2 The dye/TiO

₂

interface

2.2.1 TiO

₂

The energy characteristics of the TiO₂semiconductor are schematically shown in figure 2.3. TiO₂ is considered a wide band gap semiconductor implying a large gap between the valence band and conduction band, and with the fermi level positioned in this band gap. The valence band of TiO₂has predominantly O 2p character whereas the conduction band has predominantly Ti 3d character.

Anatase and rutile are two crystal phases of TiO₂, where the latter is thermody- namically more stable. The band gap for TiO₂in the anatase phase is 3.2 and in the rutile phase 3.0 eV. The nanostructured TiO₂studied here is dominated by the anatase phase, where the most common surfaces are the (101) followed by (100)/(010) in approximately equal amount.^30,31

The position of the conduction band edge is dependent on the surface charge on the TiO₂. This property can be used to shift the position of the conduction band edge, and thereby the energy level matching and open circuit voltage.^32–35The surface charge is altered by changing the composition of the electrolyte, e.g. the pH.

2.2.2 Dyes

The dye play a central role in the DSC. To be considered a well functioning DSC dye it must be able to absorb a large fraction of the incident light, efficiently inject electrons to the conduction band of the semiconductor, accept electrons from the redox medium, and be stable enough to sustain 10⁸turnovers.³⁶Designing dyes with these abilities is complex, however some general properties of the dye which has proven successful in DSC are discussed below.

The energy levels involved in the electron transfer at the photoactive interface are the outermost energy levels, and the occupied and unoccupied orbitals are referred to as the HOMOs and the LUMOs respectively. The spatial position of these orbitals with respect to the TiO₂ surface and the redox mediator should favor the desired electron transfer processes. The position of the LUMO in the dye molecules should ideally be located close to the TiO₂surface to favor electron injection. Similarly the position of the HOMO should ideally be located close to the redox mediator to favor electron transfer from the redox mediator and at the same time far away from the TiO₂ surface to suppress recombination between electrons in the conduction band and the dye. In figure 2.4 a theoretical picture of the HOMO and LUMO are shown for the organic D5L2A1 dye, showing the spatial position of the HOMO located dominantly

(20)

HOMO LUMO

Figure 2.4:Theoretical picture of the HOMO and LUMO in D5L2A1.

around the triphenylamine moiety, and the LUMO dominantly located around the cyanoacrylic acid moiety.

As briefly discussed in section 2.1.3 the energy level positions of the frontier electronic orbitals are also important. With respect to each other, the occupied and unoccupied energy levels of the adsorbed dye determines the absorption characteristics of the solar cell. In designing dyes there is currently a desire to redshift the absorption spectrum, i.e. reduce the gap between occupied and unoccupied energy levels, to increase the overall light harvesting ability.³⁷This can be achieved by raising the position of the occupied energy levels, or by lowering the position of the unoccupied energy levels in the dye. However, a consequence of for example a raised occupied energy levels is a reduced driving force for reduction of the oxidized dye, and therefore a trade off exists between improving spectral response and keeping efficient electronic transfer kinetics.

The anchoring to the TiO₂is also of importance for the DSC. Generally a strong chemical bond is desired between the dye and the TiO₂surface. This should ensure a good electronic coupling between the materials, which in turn favor efficient electron injection kinetics. A good anchoring also ensures that the adsorbed dyes do not desorb and is therefore also important the long term stability of the solar cell. The chemical bond between the dye and TiO₂is accomplished by incorporating anchoring groups in the dye molecular structure. The most successful anchoring group when it comes to DSC efficiencies has been -COOH.

Also of importance is the ability of the dye to form a rather close packed molecular monolayer on the surface of the TiO2upon sensitization. Multilayers and aggrega- tions of dye molecules on the TiO₂ surface results in dye molecules not in direct contact with the surface thus preventing efficient electron injection from these dyes.

A dye monolayer on the TiO₂surface allow the dyes present in the solar cell to inject electrons directly into the conduction band of the TiO₂. Also, a closely packed dye monolayer ensures that the redox mediator is sufficiently far away from the TiO₂surface to suppress the unwanted electron transfer between these two materials. In this context, dyes that contain long chains have proven to be successful.^38,39

Ruthenium based dyes, see example in figure 2.5(left), have from an early start proven to be efficient dyes in the DSC. Generally these dyes has a Ru center often connected to some bi-pyridine ligands, where modification of one of the ligands with more bulky units have yielded efficient dyes while also improving the long term stability.^37,38As of today these dyes hold the record efficiency for the DSC and has been

20

(21)

extensively studied in the DSC research community. The ruthenium based dyes N3, N719, and Z-907 dyes investigated in this is generally believed to adsorbs to the TiO₂ surface through two or three carboxylic groups on the bpy ligands.^40–43

N N

N N Ru NCS

NCS COOH

COOH COOH

COOH

N3

N S

CN COOH

D5L2A1

Figure 2.5:Example of the molecular structure of ruthenium based dye (left) and organic based dye (right).

In recent years the interest for organic dyes has increased due to newly synthesized organic dyes showing respectable efficiencies when used in the DSC, see example in figure 2.5 (right).^44–60 Specifically the organic dye competes well with the more traditional ruthenium based dyes in the solid state DSC. This is because these systems currently benefit from the high extinction coefficients of the organic dyes for efficient light harvesting, making it possible to reduce the thickness of the nanostructured semiconductor network in the working electrode, thus reducing the limitations in charge transport.⁶¹ The relatively easy synthetic routes of the organic dyes also makes them easy to modify when it comes to energy level matching in the DSC.

2.2.3 Coadsorbents

Two different types of molecules can be used on the DSC semiconductor surface. The first and most important molecule is the photon harvesting dye molecule discussed above. The second type of molecule is not photoactive and is referred to as coadsorbent.

In many studies it has been shown that the photovoltaic efficiencies, specifically of DSCs based on novel organic dyes, can be improved by the employment of coadsorbents in the dye bath solution.^62–65However increased efficiencies due to the use of coadsorbent has also been reported for the ruthenium based DSCs.^7,66–72

The role of coadsorbents has been reported to improve the solar cell efficiency due to enhanced short circuit current, J_sc, and/or open circuit voltage, V_oc. Increases in J_sc due to the adsorption of coadsorbents have been suggested to remove or alter the formation of dye aggregates on the TiO₂surface.63,64,73,74If the consequence of such removal is a relative increase in the number of electron injecting dye molecules on the surface, an enhanced photocurrent will be attained. Increase in V_oc when using coadsorbents have been addressed to effects from enhanced photocurrents, conduction band edge shift of the TiO₂ and/or reduced interfacial electron recombination.62,63,68,75–77

(22)

The use of coadsorbents in DSCs has shown somewhat conflicting results concerning the effect on the V_ocand J_sc, indicating that the role of coadsorbents varies with the nature of the dye, the coadsorbent and the examined conditions.68,69,76,77 Gen- erally there are difficulties to identify and quantify the coadsorbents on a sensitized surface, which partly limit further understanding of the changes in the DSC functional properties due to the presence of coadsorbents.

2.2.4 Dye sensitization and formation of the molecular layer

The adsorption of the dye is performed by immersion of the TiO₂surface in a dye bath solution and, after some time, an equilibrium exists between the dyes at the surface and the dyes in the solution. This equilibrium is dependent on the solvent, nature of the dye and on nature of the surface.

The adsorption of a dye molecular layer implies an organization of the dyes on the surface. As discussed above, in most dyes an anchoring unit is incorporated in the molecular structure. The anchoring physically limit the dye molecular arrangement relative to the surface and relative to each other. This implies that intermolecular interactions, favorable or non favorable, between the adsorbed dye molecules may be enhanced.

When the sensitized surface is still in the dye solution, different solvents will screen the adsorption induced molecular interactions differently. After sensitization at the dry dye sensitized surface however, the solvent has evaporated and such screening is no longer present. The intermolecular interactions between the adsorbed dye molecules as well as interaction between the solvent and the surface, due to different wetting, may therefore also influence the dye sensitized surface structure. Moreover, different solvents generally has different abilities to host protons, and this will affect the amount of dye molecules which are already deprotonated in the solution. After sensitization and solvent evaporation the surface may, as a consequence of the different abilities of the solvent to host protons, have a different surface charge. This can effect the energy level matching between the dye and the TiO₂.

Dye-sensitization of TiO₂from solutions containing coadsorbents may change the equilibrium between the dye at the surface and in the solution. Most coadsorbents has an anchor unit incorporated in the molecular structure, implying that these molecules may compete with the dye for the specific adsorption sites on the TiO₂surface. The presence of coadsorbents may also change the nature of the dye/solvent interaction and also the intermolecular interactions for the adsorbed dye at the surface, and in this way also the preference of the dye to be adsorbed on the TiO₂surface relative to in the solution. As discussed above the nature of each dye, coadsorbent, and solvent differ, thus implicating that the composition of the sensitized TiO₂surface may vary for each new dye/coadsorbent/solvent composition.

For the DSC it has been demonstrated that the absorption spectra for dyes in solution as well as for the dyes adsorbed at the TiO2surface is dependent on the solvent, showing also solvent induced changes in the electronic and/or molecular surface structure.^78,79Solvent dependent photocurrents, photovoltages and recombination kinetics in the DSC has also been reported. Often higher photocurrent was accompanied with lower photovoltage.⁸⁰Furthermore, the injection kinetics of the dye at the TiO₂surface has also been shown to be influenced by the solvent.^79–82Theoretical studies has

22

(23)

demonstrated the importance of including solvent effects.⁸³Specifically, calculations on the deprotonated organic dye D5L2A1 showed that the incorporation of water solvent molecules was accompanied with a changed geometry at the anchor unit.

The preparation of the DSC is generally performed under normal atmospheric pres- sure and in this environment the electrodes are exposed to water present in the air.

Under these circumstances it is inevitable that all DSC will contain certain amount of water molecules. Several groups have deliberately added small amounts of water to the DSC, either by addition of H₂O in the DSC or by pretreatment of the TiO₂with H₂O, and by doing so improved the efficiency of the DSC.^84–90 An increase in the V_ocand a decrease in J_schas generally been observed resulting in an overall increase in efficiency. The result is explained by a decrease in the rate of the back reaction pos- sibly due to blocking of the TiO₂surface by adsorbed H₂O or by a reduced amount of I⁻₃ on the TiO₂surface due to higher solubility in the electrolyte.^84–86The decrease in J_sc is explained by the detachment of the adsorbed dye or by a weakening of the dye/TiO₂interaction.⁸⁶ There has also been suggestion on the depletion of the thiocyanate groups due to the presence of water, however the H₂O induced changes in the working electrode has been found reversible thus suggesting little depletion of the NCS groups.⁸⁴

In summary, the surface molecular arrangement for a dye/TiO₂ surface depend on the nature of the materials as well as the assembly process. Since the function of the DSC largely depend on the precise electronic and molecular surface structure means to obtain such structural information is important. Photoelectron spectroscopy techniques are well suited to study such characteristics, and has been the main tool in the present thesis.18,45,91–99

(24)

(25)

3. Experimental

In this chapter a short presentation of the main techniques and the related analysis concepts used in this thesis are described. This chapter does not present a complete description of all aspects of the different techniques and for such information the reader is directed to more comprehensive textbooks for further details.¹⁰⁰

3.1 Electron spectroscopy techniques

3.1.1 Photoelectron Spectroscopy, PES

Photoelectron spectroscopy (PES) is used for surface characterization of materials.

The technique is based on the photoelectric effect, experimentally first observed by Heinrich Rudolf Hertz in 1887 and later explained by Albert Einstein in 1905, which describes the emission of electrons from a material under illumination of photons with energy above a certain threshold.^101,102 The conservation of energy makes it possible to express the energy relationship for the photoelectric process as described in equation 3.1,

hν = EB+ E_K (3.1)

where the hν is the photon energy, EBis the binding energy of the photoelectron, E_Kis the kinetic energy of the outgoing photoelectron. In the case of metals the binding energy is most commonly referred to the Fermi level, and thus an introduction of a fourth term in equation 3.1 is necessary. This term is known as the work function of the material, φ , and the energy value of this term is the difference between the Fermi level and the vacuum level. The energy relationship is in this case described by equation 3.2.

hν = EB+ E_K+ φ (3.2)

In this thesis many measurements are performed on the semiconducting TiO₂. In this case a natural fermi level is more difficult to determine and the energy calibration is made by positioning the Ti2p_3/2core level on 458.56 eV. This value is obtained by referencing the the C1s to the same value as for C1s found on Au sample when it has been exposed to air.

An illustration of the PES technique is found in figure 3.1 where a division of PES is made into core level PES and valence level PES depending on the origin of the photoelectron.

(26)

Core Level PES Valence Level PES

Figure 3.1:Illustration of the PES process, divided into core level and valence level PES.

O1s Ti2p

N1s

C1s

S2p Ti3p

10 8 6 4 2

Figure 3.2:PES spectrum of a D29 sensitized TiO₂sample.

In PES the sample is illuminated with high energy photons and the kinetic energy of the emitted photoelectrons are measured. From equation 3.1 (or 3.2) the binding energy of the photoelectron is calculated. It is important to realize that the obtained binding energy is generally best described by the energy difference between the initial state and the final state energies in the PES process. The initial state is the total energy of the system containing N electrons and the final state in the PES process is thus the total energy of the system containing N-1 electrons.

In PES the amount of photoelectrons are measured as a function of their kinetic energy and by using a well defined monochromatic photon energy a PES spectrum can be obtained. A PES spectrum is displayed as the amount of photoelectrons (intensity) as a function of their binding energy, see figure 3.2.

Each element has a characteristic core level pattern, i.e. the energy positions of the core levels are different for all elements. The energy positions of the core levels are also spread out over a large energy range and thus the measured PES core level lines are normally well separated from each other, as can be observed in the PES spectrum of a dye sensitized TiO2sample in figure 3.2. Element specific information is therefore gained by using core level PES.

26

(27)

The valence levels on the other hand are made up of the electrons forming bonds between the different atoms and these are found in the lowest binding energy region in a PES spectrum, see enlarged region in figure 3.2. The valence levels normally form complex band structures and to analyze these structures much information is gained by comparison to theoretical calculations.

3.1.1.1 Chemical shifts

The core levels in an atom are localized and do not participate in the formation of chemical bonds, and therefore a specific core level (for a particular element) can be found at a certain binding energy. In detail however, variations in the binding energy of a specific core level can be observed and these are known as chemical shifts.

The chemical shifts give information on the chemical state of the materials. In the total energy model the chemical shift, ∆E, between atom A and atom B is dependent on the total initial and final state energies as described in equation 3.3.

∆E =

E_A^{f inal}− E_B^{f inal}

−

E_A^initial− E_B^initial

(3.3) Differences in the initial state energies arises from external factors, such as adsorption on different materials or at different sites, whereas the difference in final state energies arises from the different systems formed after core-ionization. Including initial and final state effects is important for a full understanding of the chemical shift.

Understanding of the chemical shifts can be made using the Z+1 approximation, see figure 3.3.

Often however a qualitative estimation of a chemical shift is made using potential models. In this simplified picture the chemical shift originates from ground state changes in the electronic structure in the atom due to different bond formations and can to a first approximation be related to changes in the Coulomb interaction between the electrons and the nucleus.^100,103

An example of an experimentally measured chemical shift is shown in figure 3.4, which shows the N1s core level for the N3 dye adsorbed on TiO₂. The N3 dye contains six nitrogen atoms, which are found in two different chemical environments. Two equivalent nitrogen atoms are found in the thiocyanate ligands and four equivalent nitrogen atoms are found in the bipyridine ligands. In the N1s spectrum two signals from these nitrogen atoms can be identified with a chemical shift of approximately 2 eV between them.

3.1.1.2 Surface sensitivity, quantification, and depth distribution The intensity of a core level peak in a PES measurement depends on the distance the electrons travel in the material, d/sin θ , the mean free path, λ , of the electrons in the material, the differential cross section for electron emission from the specific core level, σ , the surface concentration of the specific element, ρ, and the analyzer transmission of photoelectrons in the measurement, S, as described in relation 3.4.

I ∝ I₀(σ · ρ · S) · exp

−d

λ sin θ

(3.4)

(28)

E_FreeMol+(Z_A*,Z_B) ≈ E_FreeMol(Z+1_A,Z_B)

E_FreeMol(Z_A,Z_B)

E_FreeMol+(Z_A,Z_B*) ≈ E_FreeMol+(Z_A,Z+1_B)

+ Z+1

E_b(A) Eb(B) ΔE

Figure 3.3:Schematic picture of how chemical shifts are estimated using the Z+1 approximation. The creation of a core hole results in an organization of the orbitals outside the core hole similar to that of a system where the nuclear charge is increased by one. As a starting point for the Z+1 approximation the E_FreeMol(Z_A, Z_B) is the total energy of the system before ionization. After core ionization in atom ZAthe total energy of the system will be E_FreeMol⁺(Z^∗_A, Z_B), and the total energy of this system is approximated with that of a system where the core ionized atom is replaced by an element having one extra nuclear charge, i.e. the next element in the periodic table (Z+1), and where the system instead has a hole in the valence levels. This state is refereed to as E_FreeMol+(Z+1_A, Z_B). The binding energy is thus the energy difference between the EFreeMol(ZA, ZB) and the E_FreeMol⁺(Z+1A, ZB) system. The procedure is repeated for ionization of the Z_Batom, resulting in a system with total energy E_FreeMol+(Z_A, Z+1_B). The estimated chemical shift, ∆E, is thus the total energy difference between the E_FreeMol⁺(Z+1A, Z_B) and the E_FreeMol+(Z_A, Z+1_B) systems. Similar estimations of chemical shifts can be made also for solid systems, where the energy needed to remove the Z atom and put back the Z+1 atom into the solid system is included.

N N

N N Ru NCS

NCS COOH

COOH COOH

COOH

Chemical shift

404 402 400 398 396

Binding Energy / eV

Intensity / arb.units

N1s

Figure 3.4:A N1s spectrum of N3 adsorbed on a TiO2surface illustrating chemical shifts.

28

(29)

The mean free path of electrons describes the average distance an electron travels in a material without energy loss. This distance is dependent on the material and on the kinetic energy of the electrons and is roughly described by the universal curve, where the minimum mean free path is found for electrons having a kinetic energy of around 60 eV and is around 3 Å. The mean free path increases up to around 40 Å for electrons with a kinetic energy of 3 keV. In the PES experiments presented in this thesis the kinetic energy of the photoelectrons were 1400 eV (or below) with a mean free path of approximately 20 Å or lower, implicating that the information gained originates dominantly from the outermost 20 Å of a dense sample surface. The short mean free path of the photoelectrons makes the PES technique very surface sensitive.

The element specificity of core level PES can be used to gain information about the composition of the investigated sample. The intensity of the core level peaks can in addition to this be used to quantify the surface composition of different elements on the sample surface. It is however observed in equation 3.4 that the intensity is dependent also on other factors apart from the surface concentration and therefore these factors need to be considered.

The cross section term varies for each element and each core line dependent on the photon energy used. In detail, the angle between the incident photon polarization and emitted electron is also an important factor when using polarized light. Changes in the analyzer transmission is obtained by measuring a reference sample with a well known surface composition. The cross section and the analyzer transmission needs to be considered when quantifying different elements on a surface. As a special case, when comparing the intensity of two core level peaks originating from atoms of the same element and separated only by a chemical shift, the cross section and analyzer transmission can be considered the same and these factors can be omitted. This can be observed in figure 3.4 where the intensity relationship between the two nitrogen peaks is close to 1:2, thus reflecting the relationship found in the molecular structure of N3.

The exponential term in equation 3.4 originates from the attenuation of the PES signal as the electrons travels through the material between the site of the PES process and the vacuum outside the sample. This implies that the PES signal from elements located deeper in the material has lower intensity compared to elements located closer to the surface even though their concentration are the same. For the quantitative analysis of inhomogeneous samples the attenuation dependence may be difficult to control, however it can be used to gain information on the composition of the surface. Effects from attenuation can be used to estimate the thickness of a material layer on top of a substrate by measuring the decrease of the substrate signal before and after depositing the overlayer. In equation 3.4 it is observed that since the λ is dependent on photon energy the attenuation effect can be enhanced in the outermost surface region by using a lower photon energy and thus by varying the photon energy a depth distribution can be made. An example is shown in figure 3.5 where the three O1s signals originating from the surface adsorbed D5L2A1 molecule is enhanced compared to the O1s peak originating from the TiO₂substrate when using a lower photon energy. In equation 3.4 it is also observed that different angles between the surface plane and the detection direction also provides an alternative route to change the surface sensitivities in a PES measurement.¹⁰⁰

(30)

Intensity /arb.units

538eV 536 534 532 530 528 526 Binding Energy /eV

D5L2A1

640 eV

758 eV

TiO2

Dye

Figure 3.5:The O1s signal from a dye sensitized TiO₂surface showing the enhanced molecular signal as the photon energy is decreases.

3.1.2 X-ray Absorption Spectroscopy, XAS

X-ray Absorption Spectroscopy (XAS) is a technique that can be used to probe the unoccupied energy levels in a sample. The XAS process involves the excitation from a core level to an unoccupied energy level, and a schematic picture of the XAS process is shown in figure 3.6 (top).

The excitation in the XAS process, is governed by the dipole selection rule implicating that excitation can only occur by a change of the orbital angular momentum by

±1, for example electronic transition from a s- or d-orbital to a p-orbital. The excitation is also localized to the site of the produced core hole giving the possibility to distinguish between contributions originating from different atoms of the same element. In XAS, similar to PES, a core hole is produced in the adsorption process and this will influence the energy levels in the probed atoms and thus ultimately the result of the measurement.

The excited state produced in the XAS process is a short lived state which decays either with the emission of photons or the emission of secondary electrons (Auger electrons), as shown in figure 3.6 (bottom). The amount of absorbed photons are pro- portional to the number of existing core holes and thus also to the amount of emitted photons and to the amount of emitted secondary electrons. A XAS spectrum can therefore be obtained by measuring either the amount of emitted photons or the amount of electrons as a function of photon energy. An example of a XAS spectrum, obtained by measuring the amount of secondary electrons in partial yield mode, is shown in figure 3.7. The cross section for photoemission varies smoothly with photon energy and thus the background in a XAS spectrum is a smooth line and the sudden increase in the amount of electrons observed in figure 3.7 is related to the emission of secondary electrons.¹⁰⁴

30

(31)

Participator Decay Spectator Decay Radiative Decay

XAS

Figure 3.6:A schematic picture of the two step XAS process. Step one involves the excitation of a core electron to an unoccupied energy level, and step two involves the decay of this excited state.

(32)

420 415

410 405

400 Photon Energy /eV

Intensity / arb.units

Figure 3.7:N1s-XAS spectrum of D5L2A1 adsorbed on TiO₂.

3.1.3 Resonant Photoelectron Spectroscopy

Resonant photoemission spectroscopy (RPES) implies photoemission spectroscopy using a photon energy that coincides (resonance) with the excitation energy of a core orbital to an unoccupied orbital in the system, see figure 3.6. The excited states formed from the X-ray absorption process primarily decay in an Auger type process leading to electron emission. In resonant PES repeated PES valence band spectra are thus measured as the photon energy is scanned in a range where excitation to an unoccupied energy level is allowed. A typical resonant PES spectrum is shown in figure 3.8.

The data is a three dimensional matrix with the scanned photon energy, the investigated binding energy region, and intensity of photoelectrons on the three axes. In the resonant PES spectrum an increase of photoelectrons will occur at binding energy positions where the kinetic energy of the secondary electrons are the same as the kinetic energy of the photoelectrons in the normal PES process.

In analyzing the various features it is useful to classify the RPES transitions into spectator and participator processes, see figure 3.6. The former imply that the decay occurs without direct involvement of the resonantly excited electron, leading to final states having two holes among the valence orbitals in addition to the excited electron and a signal that is constant in kinetic energy. The participator decay, on the other hand, implies that the resonantly excited electron participates in the decay of the core hole to the final state with an electron in the continuum and one valence hole. Since the final state in the resonant participator decay is the same as in nonresonant photoemission, the participator decay signal is constant in binding energy.

The participator decay in resonant PES can be used to map the partial contribution to the valence band from specific atoms. In figure 3.7 the highest peak originates from a N1s-XAS resonance in the cyanate unit, and at the same photon energy in the RPES spectrum, see figure 3.8, a clear resonance is also observed. Assuming that this feature in the resonant PES spectrum dominantly originates from participator decay, enhanced intensity implies large partial contribution from the nitrogen in the cyano moiety at the corresponding binding energy scale.

It is of interest to follow in detail the intensity of a particular valence orbital hole state as a function of photon energy over the energy interval of the XAS spectrum.

This procedure may allow conclusions to be drawn as to the time development of the

32

(33)

(34)

Beamline

θ

Sample

hν

E

Electron Analyser

Figure 3.9:A schematic picture of the experimental set up at the end station of BL I411.

turn bends the photoelectrons in a circular trajectory, only allowing electrons with a kinetic energy coinciding with the pass energy to pass through the hemispherical system. At the end of the trajectory the multichannel plate is positioned and counts the amount of arriving photoelectrons. The resolution of the analyzer is set by the pass energy, the slit of the entrance to the analyzer, and the radius of the hemispheres.

The photons used for PES experimental work at BLI411 at MAX-lab are generated in the MAXII storage ring. In the storage ring electrons travel at close to the speed of light, and as the electrons are subjected to a force, accelerated, synchrotron light is emitted. Due to the relativistic speed of the electrons the light is emitted in a narrow cone in the forward direction, where also the beam line is positioned. The energy of the synchrotron light ranges between 50-1500 eV and in the beam line a particular photon energy is selected using a modified SX-700 monochromator with 1220 l/mm grating and a plane-elliptical focusing mirror. To increase the intensity of the emitted light at BLI411 an undulator is inserted in the storage ring. An undulator consist of a series of magnets which repeatedly accelerates the electrons. The distance between the magnets in the undulator can be varied in order to achieve constructive interference between the light pulses emitted at different magnets at a chosen photon energy.

The main advantage of working at a synchrotron radiation facility is the ability to tune the photon energy used in the experiment. Different photon energies enables for example depth profiling, highlighting of particular elements, and probing of the unoccupied energy levels.

3.1.4.2 In house ESCA300 Setup

The quantitative PES characterization in this thesis, comparing amount of the different elements at the surface, are performed with an in house ESCA 300 spectrometer, see figure 3.10. This is a commercial instrument (Scienta) with a hemispherical electron analyzer with a radius of 300 mm. The photon source in the ESCA 300 instrument is monochromated Al Kα radiation (1486.7 eV). The photons are generated by ac- celerating electrons, produced in an two step high power electron gun, on a rotating aluminum anode which subsequently emits Al Kα radiation. The radiation is further monochromated using quarts crystals. The electron take off angle is variable but for the measurements included in this thesis an angle of 90^◦was used.

34

(35)

Figure 3.10:A picture of the ESCA300 instrument.

3.2 Photoelectrochemical techniques

Photoelectrochemical measurements were performed with the aim of linking the PES surface analysis to solar cell properties.

3.2.1 Current Voltage measurements, IV

In IV measurements the current and voltage characteristics of the solar cell is investigated under illumination and from the IV curve, see example in figure 3.11, the efficiency, η, of the solar cell is obtained. The efficiency is expressed as eqn 3.5, where J_sc is the current generated at short circuit condition, V_oc is the voltage at open circuit condition, and ff is the fillfactor. The fillfactor is the ratio between the maximum power generated by the DSC, illustrated by the smallest dashed rectangle in figure 3.11, and the product between J_scand V_oc, illustrated by the largest dashed rectangle in figure 3.11, and reflects the degree of squared shape of the current voltage curve.

η = J_sc·V_oc· f f

P_In (3.5)

In IV measurements the working electrode and counter electrode are attached to a variable external load and the voltage over the solar cell is scanned between 0 and V_oc while measuring the generated photocurrent. The standard light source in IV measurements is a solar simulator with intensity 1000 W/m².

3.2.2 Incident photon to current conversion efficiency, IPCE

The IPCE technique measures how efficiently light is convert to current, i.e. the amount of useful electrons per incident photon, for each wavelength. The IPCE is calculated according to relation 3.6

Electronic and Molecular Surface Structures of Dye-Sensitized TiO2 Interfaces

Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 753