Conductive polymers as hole-conductors for solid-state dye sensitized solar cells.

Conductive polymers as hole-conductors for solid-state dye sensitized solar cells.

Conductive polymers as hole-conductors for solid-state dye sensitized solar cells.

Master thesis by: Niklas Wahlström

Supervisor: Nick Vlachopoulos

Subject specialist: Gerrit Boschloo

Examiner: Christer Elvingson

1

List of abbreviations

Bis-EDOT = bis-3,4 ethylenedioxythiophene

Bis-PheDOT = bis-3,4 phenylenedioxythiophene

DCM = dichloromethane

DSSC = dye-sensitized solar cell.

EtOH = ethanol

FTO = fluorine-doped tin oxide

HOMO = highest occupied molecular orbital

HTM = hole-transporting material

LiClO

4

= lithium perchlorate

LiTFSI = lithium (bis-trifluoromethanesulfonyl)-imide

LUMO = lowest unoccupied molecular orbital

MeCN = acetonitrile

MO = molecular orbital

PC = propylene carbonate

PEDOT (poly-EDOT) = poly(3,4 ethylenedioxythiophene)

PIA = photo-induced absorption.

Poly-PheDOT = poly(3,4 phenylenedioxythiophene)

SHE = standard hydrogen electrode

TBAHP = tertbutylammoniumhydrogenphosphate

Tert-BuOH = tert-butanol

TBP = 4-tert butylpyridine

TiO

2

= titanium dioxide

2

Abstract.

In order to decrease the emission of greenhouse gases, there is an urgent need to find alternative energy sources that can replace the fossil fuels. Solar energy is an alternative energy source that definitely has the potential to satisfy all our energy demands. One very common way to use the sun as energy source is by means of solar cells. A solar cell is an electric device that can convert sunlight directly into electric current by absorbing photons and releasing electrons that can do work in an electric circuit. Dye-sensitized solar cells

(DSSCs) are one of the most promising types of solar cells due to their low cost and high efficiency.

One common type of DSSC is the solid state DSSC in which a solid hole-conductor is used as a charge-transporting material. One possible type of hole-conductive material that can be used is a hole-conductive polymer. In the following project I have investigated two different holeconductive polymers and their possible application as hole-conductive material in DSSC. The first polymer PEDOT (poly-3,4 ethylenedioxythiophene) has already been used in solar cells applications giving about 6% efficiency [Lei Y et al, J Phys. Chem. Letters 2013, 4, 4026-

4031]. In this project, I have investigated if it was possible to obtain even higher efficiency with this polymer. The second polymer that was investigated, poly-PheDOT (poly-3,4 phenylenedioxythiophene) has still not been investigated for DSSC applications, so this was the first test.

By using PEDOT together with the D35 dye in a DSSC, I managed to obtain 2,56% efficiency, which is lower than has been reported in earlier studies, but it still shows that

PEDOT works quite well in solar cell applications. By using poly-PheDOT together with the

D35 dye, however, I only obtained 0,075% efficiency. Spectroscopic studies showed that the regeneration of the D35 is not very effective and this is a possible reason why the efficiency of the solar cells was low. The large size of the monomer and short-chained polymer molecules is two possible reasons for the ineffective electron transfer from the polymer to the oxidised dye and therefore also the reason for the low efficiency of the solar cells. In future research, I would try to perform the photoelectrochemical polymerisation of poly-PheDOT in another solvent. By doing this, one may obtain longer polymer molecules that can penetrate more deeply into the TiO

2

film which will result in more effective electron transfer and, possibly, better solar cells.

3

Table of content

1. Introduction..............................................................................................................................6

1.1. The global energy crisis....................................................................................................6

1.2. Solar energy.....................................................................................................................6

1.3. Solar cells…………………………………………………………………………………………………………………….7

1.3.1. Introduction to solar cell…………………………………………………………………………………….7

1.3.2. Dye sensitized solar cell...........................................................................................7

2. Aim ................................................................................................................................................ 7

3. Experimental work. ..................................................................................................................... 11

3.1. Chemicals and material ..................................................................................................... 11

3.2. Electrochemical studies of compact TiO

2 blocking layer ................................................... 12

3.2.1. Electrochemical studies of compact TiO

2

blocking layer i nonaqueous electrolyte 12

3.2.2. Electrochemical studies of compact TiO

2

blocking layer in aqueous electrolyte .... 14

3.3. Electrochemical and spectroscopic studies of dye molecules.........................................15

3.3.1. Introduction.............................................................................................................15

3.3.2. Electrochemical studies of dye in non-aqueous solution and on TiO

2

electrodes...17

3.3.3. Spectroscopic studies of dye coated TiO

2

electrodes..............................................18

3.3.3.1. UV-Vis spectroscopy.........................................................................................18

3.3.3.2. Photoinduced absorption spectroscopy (PIA)..................................................18

3.4. Electrochemical studies of hole-conductive polymer.......................................................19

3.4.1. The mechanism of electropolymerisation...............................................................19

3.4.2. Electropolymerisation of PEDOT on glassy carbon electrode using a bis-EDOT monomer electrolyte..............................................................................................20

3.4.3. Electropolymerisation of poly-PheDOT on glassy carbon using a bis-PheDOT monomer electrolyte..............................................................................................21

3.5. Solar cell fabrication.........................................................................................................22

3.5.1. Preparation of dye coated TiO

2

electrodes.............................................................22

3.5.2. Photoelectrochemical polymerisation of hole-conductive polymers on dye coated

TiO

2

electrodes........................................................................................................22

3.5.3. Post treatment of polymer coated TiO

2 electrodes................................................24

3.5.4. Photoinduced absorptio spectroscopy (PIA) of polymer coated TiO

2

electrodes...24

3.5.5. Ag counter electrode preparation...........................................................................24

3.6. Investigation of the solar cells..........................................................................................26

3.6.1. J-V characterization................................................................................................26

4. Results and discussion.................................................................................................................26

4

4.1. Results from electrochemical analysis of TiO

2

blocking layers in EtOH.............................26

4.1.1. Results fron reference electrode calibration in EtOH..............................................26

4.1.2. Results from cyclic voltametry of TiO

2

blocking layers in EtOH..............................28

4.2. Results from electrochemical analysis of TiO

2

blocking layers in PC.................................30

4.2.1. Results from reference electrode calibration in PC.................................................30

4.2.2. Results from cyclic voltametry of TiO

2

blocking layers in PC .................................. 31

4.3. Results from electrochemical analysis of TiO

2

blocking layers in aqueous electrolyt……..33

4.3.1. Results from the reference electrode calibration in H

2

O........................................33

4.3.2. Results from cyclic voltametry of TiO

2

blocking layers in H

2

O ................................ 34

4.4. Cyclic voltametry of dye coated TiO

2

electrodes and dye solutions………………………………37

4.5. Spectroscopic studies of dye coated TiO

2

electrodes ....................................................... 38

4.6. Results from electropolymerisation and characterization of hole-conductive polymers on glassy carbon electrodes....................................................................................................40

4.6.1. Results from electropolymerisation of PEDOT........................................................40

4.6.2. Results from electropolymerisation of poly-PheDOT.............................................47

4.7. Results from solar simulator measurments......................................................................51

4.7.1. Solar simulator measurements with PEDOT...........................................................51

4.7.2. Solar simulator measurements with poly-PheDOT.................................................53

4.8. Results from PIA measurements on polymer coated TiO

2

electrodes...............................56

5. Conclusions.................................................................................................................................57

6. Summary in non-scientific form..................................................................................................58

7. Aknowlegdements......................................................................................................................59

8. References..................................................................................................................................60

9. Appendix 1. ................................................................................................................................. 62

5

1. Introduction

1.1. The global energy crisis

The world is facing an increasing demand of energy due to an increasing population as well as the fast development of new technologies. In 2008, the fossil fuels (oil, coal and gas) correspond to 85% of the total energy supply of the earth

[1]

. Unfortunately, the increasing demand of energy will lead to depletion of the fossil fuel sources. It is also a well-known fact that the usage of fossil fuels causes emission of polluting gases (such as CO

2

) which leads to the greenhouse effect and to global warming. Due to the depletion of fossil fuel sources and the global warming effect, there is an urgent need to find other possible renewable energy sources that can replace the fossil fuels.

1.2. Solar energy

An alternative replacement of the fossil fuels is to use solar energy. In solar energy, we are using the sun as energy source. The sun is the primary energy source of the earth. All life and all chemical processes that create life depend of the energy from the sun. Each second, the earth receives about 174 PJ (1,74*10

17

J) from the sun. Some of the received energy is reflected back into space by clouds, the earth’s surface and by the atmosphere. In the end, the total amount of energy that is absorbed by the land and oceans of earth correspond to about 89

PJ (8,9*10

16

J) each second, see Figure 1.

Figure 1: What is happening with the incoming energy from the sun

[3]

.

In 2002, the total amount of absorbed energy from the sun in one hour was twice as high the total energy demand of the earth in one year

[2]

. Therefore, solar energy has definitely the potential to satisfy all our energy demand which makes it a very attractive alternative to fossil fuels.

6

1.3. Solar cells

1.3.1. Introduction to solar cells.

A solar cell, also called a photovoltaic cell (the latter term is applied in the general case of light-to-electricity conversion devices), is an electric device which converts sunlight directly into electricity by absorbing photons from the incoming sunlight and releasing electrons that can do work in an electric circuit. The first solar cell was made with selenium wafers in 1883 by Charles Fritts

[11]

. This solar cell had an efficiency of about 1%. In 1954, G.L. Pearson,

Daryl Chapin and Calvin Fuller were using a silicon based p-n junction to obtain a solar cell that had about 6% efficiency

[4]

. Since that time, new types of solar cells have been continually developed and, nowadays, the best-performing solar cells, including silicon-based ones, can give more than 20% efficiency

[5]

. A problem with the silicon-based solar cells is that the production costs of these solar cells are high. Even though the production cost has been recently reduced, the price is still too high for widespread application for these types of solar cells.

1.3.2. Dye sensitized solar cells

Dye-sensitized solar cells (DSSCs), also called Grätzel cells, (from the name of their inventor) are considered to be one of the most promising alternatives in solar cells applications due to their low cost in comparison to high efficient silicon-based solar cells. The first efficient

DSSC for converting solar power to electricity was described by Brian O’Regan and Michael

Grätzel in a 1991 publication in Nature(London)

[6]

.

The main component of a DSSC is the photoelectrode which is a dye-coated porous TiO

2 film deposited on a conductive fluorine-doped tin oxide (FTO) glass electrode. The most common types of dyes in DSSCs are organic dyes or organometallic compounds with ruthenium. The reason for choosing these types of dyes is that they show strong light absorption in the visiblelight region. Two of the most common types are liquid DSSC and solid-state DSSC.

In a liquid DSSC we use a liquid with a redox couple as our electrolyte. One of the most common electrolytes is acetonitrile (MeCN) containing a I

-

/ I

3

-

redox couple. Another alternative is to use MeCN containing a coordination Co

2+

/Co

3+ compound as redox couple. In order to increase the overall performance of liquid DSSC, some additives such as 4-tert butylpyridine (TBP) and lithium (bis-trifluoromethanesulfonyl)-imide

(LiTFSI) is commonly added to the electrolyte.

When a liquid DSSC is irradiated with sunlight, light will be absorbed by the dye molecules which are absorbed on the TiO

2

film, which leads to injection of an electron into the conduction band of TiO

2

. The electron will diffuse through the mesoporous TiO

2

film and reach the FTO coated glass electrode. The electron will then move to the counter electrode

(usually a platinum electrode) via an external circuit. The oxidized dye will be reduced by the species in the electrolyte, and the species in the electrolyte will be re-reduced by taking up an electron at the counter electrode. The net result will be a continuous flow of current in the electrical cell circuit and also a conversion of sunlight to electrical energy if a load is

7

interposed between the two electrodes. A schematic picture of a liquid dye sensitized solar cell containing a I

-

/ I

3

- redox couple is shown in Figure 2.

Figure 2: A schematic picture of a liquid dye sensitized solar cell.

[8].

The highest efficiency of a liquid DSSC is around 13% which was obtained in 2013 by using specially designed porphyrin dye molecules

[13]

. There are some factors that limit the performance of the liquid dye sensitized solar cell. First of all, not all sunlight is absorbed by the dye. Only photons that have energy at least equal to the HOMO-LUMO energy difference of the dye will be absorbed. Some energy will also be lost due to de-excitation of electrons from LUMO to HOMO (which leads to emission of heat or light) or to recombination of electrons injected in TiO

2

with oxidized dye molecules or with species in the electrolyte. In the recombination process, the electron that has been injected into the TiO

2

film will move back to the dye or, more commonly, to redox species in the electrolyte, instead of going to the counter electrode. Some energy will also be lost as heat due to the ohmic resistance in the solar cell. Another problem is the high cost of platinum which is used as counter electrode material. Studies have shown that platinum can be replaced by other materials with lower cost, such as graphene, carbon nanotubes or a conductive polymer, but platinum is still the counter electrode material that gives the most efficient solar cells

[27-31]

. Development of lowcost counter electrode materials for liquid DSSC which can replace platinum without lowering the performance of the solar cell is one of the most important research fields in the development of liquid DSSC suitable for large-scale production.

One problem for DSSC with a liquid electrolyte based on MeCN is the volatility of the electrolyte, which makes it hard to use these solar cells at higher temperatures. The toxicity of

MeCN is also a problem. Water has been investigated as a possible replacement of MeCN due to its lower volatility and non-toxicity. By using water as solvent in a liquid DSSC, an efficiency of 2,4% has been reported

[32]

. Another possible way to overcome the problem with volatile solvents is to replace the liquid electrolyte with a room-temperature molten salt (ionic liquid). One typical example of an ionic liquid used in solar cell application is BMIM-PF

6

(1-

Butyl-3-methylimidazolium hexafluorophosphate)

[26]

. Though, the high cost and slow diffusion of redox mediators between the counter electrode and the photoelectrodes are, however, two problems that limit the use of ionic liquids in solar cell applications.

8

Another option is to replace the liquid electrolyte by a solid charge-transporting material

(HTM), most commonly a hole-conductor. One common type of hole-conductor are holeconductive polymers. A typical example is poly-3,4-ethylenedioxythiophene (PEDOT), see

Figure 4. Related studies have shown that by using PEDOT as HTM in a solid DSSC, we can obtain efficiencies around 6%

[10]

. A schematic picture of a solid-state dye sensitized solar cell with PEDOT as hole-conductive material is shown in Figure 4.

Figure 3: The structure of PEDOT, one example of a polymer that can be used as holetransporting material in a solid state dye sensitized solar cell

[11]

. The structure of the monomer (bis-EDOT) is shown in the parenthesis

[11]

.

Figure 4: Schematic picture of a solid state dye sensitized solar cell with PEDOT electrolyte

[11]

.

The working mechanism of this kind of solar cell is similar to the working mechanism of a dye sensitized solar cell with a liquid electrolyte. The dye molecules absorb light and inject electrons into TiO

2

layer. The electrons will move through the mesoporous TiO

2

film and then through the compact TiO

2

blocking layer, separating the porous TiO

2

layer from the conductive FTO glass electrode. The reason for using the TiO

2

blocking layer will be further discussed below. Subsequently, electrons will move to the counter electrode via an external circuit. The electrons at the counter electrode will move through the polymer layer and reduce the oxidized dye (S

+

). The working mechanism of a solid state DSSC with a PEDOT hole transporting material can be described by the processes (1)-(5) below:

9

S

hv

S

*

S

*

S

 

e

(

TiO

2

)

e

(

TiO

2

)

S

PEDOT e

(

anode

)

S

e

PEDOT

(

cathode

)

PEDOT

 

e

(

cathode

)

PEDOT

( 1 )

( 2 )

( 3 )

( 4 )

( 5 )

The choice of the HTM is a very important parameter in a solid DSSC. Effective electron transfer between the HTM and the oxidized dye is necessary in order to obtain an efficient solar cell. The energy level of the oxidized dye must be lower than the energy level of the

HTM so that electron transfer from the HTM to the oxidized dye can occur. It is also important that the hole-transporting material penetrates the porous TiO

2

film to a high extent.

Effective penetration of the porous TiO

2 film will lead to large contact area between the polymer molecules and the dye coated TiO

2

particles, which will lead to effective electron transfer between the hole-transporting material and the oxidized dye.

One problem that can occur in these types of solar cells is that when the HTM penetrates the

TiO

2

film, it can come into contact with the FTO glass electrode and be oxidized or reduced

This will lead to short-circuit and loss of current through recombination (electron transfer from FTO glass to the HTM). In order to prevent the HTM from coming into direct contact with the FTO glass, a compact layer of TiO

2

is deposited on the FTO glass. This layer is often referred to as the underlayer or blocking layer. It is very important that the blocking layer is smooth and covers the whole FTO glass. The thickness of this layer is also a very important parameter. If the film is too thin, direct contact between HTM and the FTO glass is still possible, which will lead to an enhancement of the recombination currents and, ultimately, to a loss of energy. On the other hand, if the film is too thick, more energy will be lost as heat, due to that the electrical resistance in the film increases with the thickness of the film. For a solid DSSC with a conductive polymer as HTM, the thickness of the compact blocking layer should be around 200 nm. One problem is that the blocking layer is usually made by using spray pyrolysis, which makes it hard to control the thickness of the film. Other methods, such as atomic layer deposition (ALD) can be used to obtain better control over the film thickness, but the high cost of this method may be a problem.

The main factors that limit the efficiency of a solid DSSC with a HTM are similar to these effective in a liquid DSSC. First of all, not all the sunlight is absorbed by the dye. Another possible problem is recombination with the dye molecules or with the HTM (the electron moves back from the TiO

2

layer to the dye or to a hole in the HTM). The recombination processes can be described by reactions (6)-(7) below, in the case of PEDOT as HTM

S

e

PEDOT

(

TiO

2

e

)

S

(

TiO

2

)

PEDOT

(

(

6

7 )

)

If the regeneration of the oxidized dye (S

+

) is efficient, the concentration of S

+

is so low that

(7) will be the dominant recombination process.

10

Another problem is the low conductivity of the HTM. The conductivity of this type of material is lower than the conductivity of liquid electrolytes, resulting in a higher resistance.

This is the main reason why solid DSSCs may have a lower efficiency than liquid DSSCs, especially for light intensity values around the light intensity of the sun.

2. Aim

The aim of this thesis is to study solid DSSC that are based on an electronically conductive polymer as HTM. Two different conductive polymers (see Figure 5) will be tested and compared. PEDOT has already been tested in solid DSSC and studies have shown that we can obtain 6% efficiency by using PEDOT as a HTM in a solid DSSC

[10]

. In this project I will investigate whether even higher efficiency can be obtained with this polymer. An alternative conducting polymer, Poly-PheDOT has not been investigated in solar cell applications up to now; the first test for this polymer in solar cell applications is presented here. Poly-PheDOT has many desirable properties for a polymer in solar cell applications. It has high conductivity and high transparency. It is also chemically stable and easily synthesized

[12]

.

Figure 5: Chemical structure of the hole conductive polymers which were studied in this project a) PEDOT, b) poly-PhEDOT. The corresponding monomer (bis-EDOT and bis-

PheDOT) of each polymer is shown in parenthesis.

The first part of the project is devoted to electrochemical studies of the components in a solid state dye-sensitized solar cell with a hole-conductive polymer as charge-transport medium. By performing electrochemical studies of different dyes, both in the dissolved and in the TiO

2

adsorbed state, of the compact TiO

2

blocking layer (underlayer), and of the hole-conductive polymers, we can obtain useful information that can be used in the optimization to obtain highly efficient solar cells.

The next and most important part of the project is the use of conducting polymers in dye solar cells. The solar cells will have the same structure as shown in Figure 5 (a thin film of a holeconductive polymer between a dye coated TiO

2

film and an Ag counter electrode). The

11

polymer will be deposited on dye coated TiO

2

electrodes by using photoelectrochemical polymerization, the mechanism of which will be discussed in section 3.

3. Experimental work

3.1. Chemicals and material

All chemicals used in this thesis were purchased from Sigma Aldrich except the dye molecules which were obtained from Dyenamo (Sweden). The conductive FTO glass was purchased from Pilkington Glass (Tokyo).

3.2. Electrochemical studies of a compact TiO

2

blocking layer.

3.2.1. Electrochemical studies of a compact TiO

2

blocking layer in non-aqueous electrolytes.

All electrochemical measurements were made by using an IviumStat potentiostat (Ivium

Technologies B.V., Eindhoven, The Netherlands).

As discussed in Section 1.3.2, one problem which can occur in solar cells with holeconductive polymer electrolytes is that the polymer can penetrate the mesoporous TiO

2 film and come into contact with the FTO glass electrode, which will lead to short-circuit of the solar cell and loss of energy trough recombination (electron transfer from the FTO glass to the conductive polymer). To avoid this problem, a thin compact blocking layer, or underlayer, of

TiO

2

is deposited on the FTO glass. The blocking layer can be prepared by using different method such as spray pyrolysis, atomic layer deposition (ALD) and also by putting the electrodes into an aqueous solution of TiCl

4 and heat the solution to 70 o

C for a certain time.

The stability of the compact TiO

2

layer can be investigated by using cyclic voltammetry.

Cyclic voltammetry is an electrochemical method that can be used to study electrode reactions with reactants either dissolved in solution or attached on an electrode surface. In cyclic voltammetry, one measures the current as a function of the applied potential between a working electrode (the electrode where the redox reactions take place) and a reference electrode. In a typical cyclic voltammetry experiment, usually we start the measurement by applying a potential at which no redox reactions in the electrolyte take place at the working electrode. The potential is then linearly changed with time until a potential at which oxidation or reduction of species in the electrolyte occurs at the working electrode. Thereafter, the potential is reversed and goes back to its initial value. By plotting the current as a function of applied potential on the working electrode, we obtain a cyclic voltammogram. A typical example for a reversible redox reaction is shown in Figure 7.

12

Figure 6: A typical cyclic voltammogram for a reversible redox reaction. E pc

is the potential where oxidation of the species in the electrolyte occurs and E pa

is the potential where reduction occurs. I pc

and I pa

are the peak currents from the oxidation and reduction. The redox potential of the redox couple is defined as the mean value of E pc and E pa

.

By recording the cyclic voltammogram of ferrocene in an organic solvent or ferrocyanide in an aqueous electrolyte for a system in which we use FTO glass coated with blocking layer as a working electrode, one can obtain important information about the blocking layer. If the blocking layer is good, no oxidation or reduction peak for ferrocene or ferrocyanide will be obtained in the voltammogram, since the TiO

2

blocking layer covers the whole FTO glass so that to prevent the oxidation and reduction of ferrocene from taking place.

The blocking layers were prepared by first cutting FTO glass into 30 mm×10 mm pieces. The electrodes were heated step-wise to 500 o

C. The blocking layer was prepared by using spray pyrolysis. The composition of the spraying solution was 0.2 M Ti-isopropoxide and 2 M acetylacetone in 2-propanol. Blocking layers with varying thicknesses were prepared by doing

5, 8, 10 and 12 spray cycles on different electrodes.

In the cyclic voltammetry measurements, the potential is measured with respect to an

Ag/AgCl reference electrode. In a cyclic voltammogram, however, we want to plot the potential vs. the standard hydrogen electrode, SHE (which is always 0 V per definition) because the redox potential for an Ag/AgCl electrode is different for different electrodes and the redox potential for an Ag/AgCl can also change with time. In the case of experiments in aqueous electrolytes we use the Ag/AgCl electrode with an aqueous chloride solution (3 M

NaCl in H

2

O) and in the experiments with non-aqueous electrolytes we use an Ag/AgCl electrode with 2 M LiCl in EtOH.

We next need to convert the potentials with respect to the Ag/AgCl electrode to potentials with respect to the SHE. We can determine the redox potential of the non-aqueous Ag/AgCl reference electrode by recording the cyclic voltammogram of in-situ added ferrocene in the same solution as we perform our measurements. Since the redox potential of ferrocene is

0.624 V vs. the SHE

[21]

, we can then calculate the redox potential of the non-aqueous

Ag/AgCl reference electrode and thereafter we can plot all cyclic voltammograms vs. the

SHE. The potential range for the calibration is chosen so that the oxidation of ferrocene and reduction of the oxidized form, ferrocenium, can be recorded. A typical potential range for these type of calibration is 0 V to 1 V.

The calibration of the Ag/AgCl reference electrode with 2 M LiCl in EtOH was made by recording the cyclic voltammogram of an electrolyte containing 5 mM ferrocene and 0.2 M

13

LiTFSI in EtOH. The working electrode was a 0.070 cm

2

glassy carbon electrode and the counter electrode was a stainless steel plate. The scan range was from 0 V to 1 V and the scan rate was set to 0.050 V/s.

The cyclic voltammograms of the electrodes with blocking layers were recorded by using 1.5 cm

2

FTO glass coated with blocking layers of varying thicknesses as the working electrode, together with a stainless steel counter electrode and the Ag/AgCl electrode with 2 M LiCl in

EtOH as reference electrode. The electrolyte was 0.2 M LiTFSI and 5 mM ferrocene in EtOH.

A cyclic voltammogram of a pure FTO glass was also recorded and compared with the cyclic voltammograms of the electrodes with a blocking layer. Finally, the stability of the blocking layers was measured by performing 20 scans between 0 V and 1 V on the samples with 10 and 12 cycles.

In order to find out how the solvent of the electrolyte affects the stability of the blocking layer, all experiments were repeated for an electrolyte containing 0.2 M LiTSI and 5 mM ferrocene in polycarbonate (PC). All experiments with the PC electrolyte were performed in the same way as the experiments made on the electrolyte with EtOH.

3.2.2.

Electrochemical studies of the blocking layer in an aqueous electrolyte

.

The stability of the TiO

2

blocking layers were also investigated in an aqueous electrolyte. The composition of the electrolyte was 0.5 M KCl and 2 mM K

4

Fe(CN)

6

in H

2

O. The pH of the solution was then adjusted to 4.5 by adding 0.5 M HCl (aq).

When we use an aqueous electrolyte, we must use a different Ag/AgCl reference electrode compared to the experiments with the non-aqueous electrolytes. The new reference electrode will have a different redox potential vs. the SHE. If we want to plot our voltammograms made in an aqueous electrolyte with respect to the SHE, we must start by determining the redox potential of the aqueous Ag/AgCl electrode.

The redox potential of the aqueous Ag/AgCl electrode with 3 M NaCl in H

2

O was measured by determining the redox potential of the ferrocyanide ( Fe(II)(CN)

6

-6

/Fe(III)(CN)

6

-6

) redox couple vs. the Ag/AgCl electrode. The calibration was made by using a glassy carbon working electrode with an area of 0.070 cm

2

, a stainless steel counter electrode and a

Ag/AgCl with 3 M NaCl in H

2

O reference electrode. The composition of the electrolyte was

0.5 M KCl and 2 mM K

4

Fe(CN)

6

in H

2

O. The experiments were performed by scanning from

0 V to 0.8 V and then scan back to 0 V. The scan rate was set to 0.050 V/s.

The cyclic voltammograms of the electrodes with blocking layers were recorded by using 1.5 cm

2

FTO glass coated with a blocking layer with varying thicknesses as the working electrode, together with a stainless steel counter electrode and a Ag/AgCl reference electrode.

The electrolyte was 0.2 M LiTFSI and 5 mM K

4

Fe(CN)

6

. A cyclic voltammogram of a pure

FTO glass electrode was also recorded and compared with the cyclic voltammograms of the electrodes with a blocking layer.

14

As we have discussed earlier, another method to obtain a compact blocking layer of TiO

2

FTO glass is to immerse the electrodes into an aqueous solution of TiCl

4

and heat the solution to 70 o

C for a certain time. In order to compare how the quality of the blocking layer varies with the preparation method, FTO glass electrodes were immersed into a 2 M aqueous solution of TiCl

4

. The solution was heated to 70 o

C for 2 h. The electrodes were washed with distilled water and sintered at 500 o

C for 30 min. The cyclic voltammograms were recorded by using a 0.5 M KCl and 2 mM K

4

Fe(CN)

6

in water electrolyte with pH=4.5 and scanning from 0 V to 0.8 V and then to -0.8 V and finally back to 0 V with a scan rate of 0.050 V/s.

Cyclic voltammograms of pure FTO glass were also recorded and compared with these for the electrodes with a blocking layer

3.3. Electrochemical and spectroscopic studies of dye molecules.

3.3.1. Introduction

One of the essential components in a DSSC is the dye (or sensitizer). Strong light absorption in the visible region (sunlight) is important, so that the dye can absorb as much sunlight as possible. The dye most also bind strongly TiO

2

.

Furthermore, the LUMO level of the dye must also have a higher energy than the conduction band edge of TiO

2

in order for the injection of electrons from the excited dye into TiO

2 to be efficient. The dyes must further be chemically stable and not decompose at higher temperatures; in this respect the long-time stability of the dye is an essential requirement. It is also preferable if the dyes are non-toxic and easily synthesised.

Different dyes (see Table 1 and 2) were investigated by using UV-Vis spectroscopy, PIA

(photoinduced absorption) and cyclic voltammetry.

Table 1: The chemical structure of the organic dyes

Dye

D35

[15]

Structure formula

15

LEG-4

[11]

MK2

{16]

D21

[18]

DN-F01

(L0)

[14]

16

Table 2: The structure of the Ru dyes

N3

[19]

Z-907

{20]

K77

[17]

3.3.2. Electrochemical studies of dye molecules in non-aqueous solution and on TiO

2 electrodes.

The electrochemical properties of the dyes are of great importance. The dye is supposed to give electrons to the TiO

2

upon illumination which means that the LUMO level of the dye must have a higher energy than the conduction band edge of TiO

2

. The dye should have a high redox potential (the oxidised form of the dye should be easily reduced by the holetransporting material). By using dye coated TiO

2

electrodes as working electrodes in an electrochemical cell and perform cyclic voltammetry, we can measure the redox potential of the dyes.

17

A mesoporous TiO

2

film was prepared by using colloidal Dyesol TiO

2

paste with an average particle size of 25 nm. The TiO

2

films were prepared by adding a small amount of the TiO

2 paste on pre-cleaned FTO glass electrodes (25 x 15 mm) and then smearing it out uniformly on the surface by using a glass rod. This method is known as doctor blading. The thickness of the film was measured to 5 μm. The electrodes were sintered at 500 o

C for 30 min, after which they were cooled down to room temperature. The TiO

2

electrodes were then immersed into a 20 mM aqueous solution of TiCl

4

and the solution was heated to 70 o

C for 30 min. The electrodes was carefully washed with distilled water and dried under a strong nitrogen gas flow. Thereafter, the electrodes were sintered at 500 o

C for another 30 minutes, cooled down to 90 o

C, and finally immersed into different dye solutions containing the dyes given in Table

1 and 2. The concentration of the dye solutions was 0.2 mM and the solvent was a 1:1 v/v mixture of MeCN and tert-BuOH (prepared by mixing equal volumes of MeCN and tert-

BuOH).

The setup for the cyclic voltammetry experiments on dye coated TiO

2

electrodes was a three electrode system using the dye coated TiO

2

electrode as working electrode, a stainless steel counter electrode, and a Ag/AgCl reference electrode with 2 M LiCl in EtOH. The electrolyte was 0.1 M LiTFSI in MeCN. The cyclic voltammograms were recorded at different scan rates

(0.05 V/s, 0.2 V/s and 0.5 V/s.).

Cyclic voltammetry experiments were also performed in dye solutions. Each dye solution contained 0.2 mM dye and 0.1 M LiTFSI in a 1:1 mixture of MeCN and tert-BuOH. The working electrode was a glassy carbon electrode with an area of 0.070 cm

2

, the counter electrode was a stainless steel electrode and the reference electrode was a Ag/AgCl electrode containing 2 M LiCl in EtOH. The cyclic voltammograms were recorded by scanning from 0

V to 1.5 V and then back to -0.3 V. Three different scan rates were used (0.05 V/s, 0.2 V/s and 0.5 V/s).

3.3.3. Spectroscopic studies of dye coated TiO

2

electrodes

3.3.3.1. UV-Vis spectroscopy

The light absorbing properties of the dyes shown in Table 1 & 2 were investigated by using

UV-Vis spectroscopy. We irradiate the sample with light with a certain intensity (I

0

) and measure the intensity of the out-going light from the sample (I). By comparing the difference in intensity of the incoming and out-going light intensity, we can measure how many photons which are absorbed by the dye. The relation between intensity and concentration is given by the Lambert-Beers law

[11]

: log

10

I

0

I

c

l

( 8 )

18

where c is the dye concentration, l is the thickness of the sample and ε is the molar extinction coefficient; the latter is a characteristic for a particular dye at a given wavelength. A typical set-up for a UV-Vis spectroscopy measurement is shown in Figure 7:

Figure 7: A typical set-up for a UV-Vis spectroscopy experiment

[33]

.

The dye coated TiO

2

films were prepared as described in Section 3.3.2. The UV-Vis spectroscopy measurements were performed on dye coated TiO

2

electrodes by using a HR-

2000 Ocean Optics fibre optics spectrometer. The light source was a deuterium lamp, and a

TiO

2

film without any absorbed dye was used as a reference.

3.3.3.2. Photoinduced absorption spectroscopy (PIA)

PIA is a suitable method to study the electron transfer processes in a DSSC, by investigating if the dye molecules inject electrons into TiO

2 upon illumination and also whether the holeconductive polymer can regenerate the oxidised dye. By recording PIA spectra of dye coated

TiO

2 electrodes, we can investigate whether the dye can inject electrons into TiO

2

upon illumination. Later in this work, we will also do PIA measurements on TiO

2

electrodes coated with both dye and hole-conductive polymer to investigate if the dye can be regenerated by the hole-conductive polymer. The setup for the PIA experiments on dye coated electrodes is shown in Figure 8.

Figure 8: Setup used for the PIA measurement on dye coated TiO

2

electrodes

[11]

.

19

PIA measurements were made on dye coated TiO

2

electrodes (the same samples that were used in the UV-Vis spectroscopy measurements). The sample was excited by a 465 nm blue

LED which is switched on and off with a frequency of 9.33 Hz. A 20 W white light tungstenhalogen lamp was used as a continuous background light source. The PIA spectrum was recorded between 300 nm and 1000 nm by using a Si detector.

3.4. Electrochemical studies of hole-conductive polymers

3.4.1. The mechanism of electropolymerisation.

Electropolymerisation is an electrochemical method that can be used to deposit a polymer film on a conductive substrate. In this method, we apply an external potential to an electrolyte containing the monomer. This potential is often referred to as the monomer oxidation

potential.

The electropolymerisation is initiated by oxidation of a monomer which is caused by the applied oxidation potential. The oxidation of the monomer forms a reactive monomer radical.

The monomer radical can further react either with another monomer radical or a neutral monomer and form an intermediate which is oxidised to a dimer by loss of protons. The next step is oxidation of the dimer. The dimer has a lower oxidation potential because the electrons in the HOMO in the dimer have a higher energy than these in the HOMO of the monomer, which makes the dimer more easily oxidised than the monomer. The oxidised dimer can now react with another dimer or a monomer to form a trimer or a tetramer. The polymer chain can continue to grow by adding more monomers or dimers. When the polymer molecules grow, they will rapidly become insoluble in the solvent and they will eventually deposit as a polymer film on the electrode. A schematic picture of the mechanism of electropolymerisation is shown in Figure 9:

Figure 9: A schematic picture of electropolymerisation

[25]

.

20

3.4.2. Electropolymerisation of PEDOT on a glassy carbon working electrode using a bis-EDOT monomer electrolyte.

The electropolymerisation of PEDOT on a glassy carbon working electrode was made by using cyclic voltammetry. By using this method, we can find the oxidation potential of bis-

EDOT. When we record the cyclic voltammogram of a bis-EDOT solution, we will see a peak in the cyclic voltammogram that corresponds to the oxidation of the bis-EDOT monomer. By doing repeated scans between the monomer oxidation potential and a lower potential, we will obtain a polymer coating on the glassy carbon electrode.

The electropolymerisation was performed by using a three-electrode system (see Figure 10).

The working electrode was a glassy carbon electrode with an active area of 0.070 cm

2

(a circular electrode with 3 mm diameter), the counter electrode was a stainless steel electrode and the reference electrode was Ag/AgCl electrode with 2 M LiCl in EtOH. Two different electrolytes were tested in order to find out if the solvent has any effect on the polymerisation.

The first electrolyte contained 0.1 M LiTFSI and 5 mM bis-EDOT in MeCN. The second electrolyte contained 0.1 M LiTFSI, 0.05 M Triton X-100 and 1 mM bis-EDOT in H

2

O. The solubility of bis-EDOT in water is very low, so Triton-X100 (a surfactant) is added to increase the solubility of bis-EDOT in H

2

O.

Figure 10: The setup for the electropolymerisation (1), (2) and (3) represent the three electrodes. (1) is the working electrode (glassy carbon), (2) is the reference electrode

(Ag/AgCl electrode) and (3) is the counter electrode (stainless steel).

First, the cyclic voltammograms of the 0.1 M LiTFSI in MeCN (without monomer) were measured by scanning from -0.5 V to 1.0 V and back to -0.5 V using a scan rate of 0.050 V/s.

Thereafter, bis-EDOT was added to the electrolyte and the measurement was repeated. The electropolymerisation was performed by cycling between -0.5 V and 0.9 V 20 times. After 20 cycles, a black surface coating on the glassy carbon electrode was visible. The electrode was then carefully washed with acetonitrile. The washed electrode was put back into an electrolyte

21

containing 0.1 M LiTFSI in MeCN (without monomer) and the cyclic voltammetry experiment was repeated by doing 10 scan cycles between -0.5 V and 0.9 V.

The electropolymerisation in the aqueous electrolyte was performed in a similar way. First, the cyclic voltammogram of 0.1 M LiTFSI, 0.05 M Triton-X100 in water (without monomer) was recorded by cycling from -0.5 V to 1.2 V. Thereafter, bis-EDOT to obtain a concentration of 1 mM was added and the electropolymerisation was performed by cycling between -0.5 V and 1.2 V 20 times. After 20 cycles, a black surface coating was visible on the glassy carbon electrode. The electrode was then carefully washed with acetonitrile The washed electrode was put back into an electrolyte containing 0.1 M LiTFSI and 0.05 M Triton-X100 in H

2

O and the cyclic voltammetry experiment was repeated by doing 10 scan cycles between -0.5 V and 1.2 V.

3.4.3. Electropolymerisation of poly-PheDOT on a glassy carbon electrode using a bis-PheDOT monomer electrolyte.

The electropolymerisation of poly-PheDOT was performed with the same setup as for the electropolymerisation of bis-EDOT (see Figure 11). The electropolymerisation was tested in two different electrolytes. The composition of the first electrolyte was 1 mM bis-PheDOT and

0,1 M TBAHP in DCM and the composition of the second electrolyte was 0,1 M tertbutylammoniumhydrogenphosphate (TBAHP) and 1 mM bis-PheDOT in an 1:1 mixture of MeCN and dichloromethane (DCM). No electropolymerisation were performed in aqueous electrolyte due to the very low solubility of bis-PheDOT in water.

3.5. Solar cell fabrication

3.5.1. Preparation of dyed coated TiO

2

electrodes

The transparent FTO glass was purchased from Pilkington Glass (Tokyo). The FTO glass electrodes were prepared by cutting FTO glass into 25 mm×15 mm pieces. The electrodes were etched with zinc powder and 2 M HCl. This step is made to prevent short-circuit of the solar cell (direct contact between FTO glass and the Ag counter electrode). The electrodes were washed in an ultrasonic bath with 2% soap solution, distilled water, acetone and ethanol for 30 min.

A compact blocking layer of TiO

2

was deposited on the glass electrodes by heating the electrodes step-wise to 500 o

C and spraying a solution containing 0,2 M Ti-isopropoxide and

2 M acetylacetone in 2-propanol on the electrodes. Ten spraying cycles were made on each electrode in order to obtain a film thickness around 200 nm. After spraying, the electrodes were kept at 500 o

C for 30 min and were thereafter cooled down to room temperature.

A thin film of mesoporous TiO

2

Dyesol paste (25 nm, average particle size) was applied on the electrodes by using doctor blading. The TiO

2

electrodes were then sintered at 500 o

C for

22

30 min. After sintering, the electrodes were cooled down to room temperature and then placed into an 20 mM aqueous solution of TiCl

4

. The solution was heated to 70 o

C for 30 minutes.

The reason for doing this TiCl

4

treatment is to increase the roughness of the TiO

2

films which will lead to better adsorption of the dye. Thereafter, the electrodes were washed with water and dried by N

2

gas. The electrodes were sintered at 500 o

C for 30 min and cooled down to

90 o

C. The electrodes were then placed into dye solution overnight (approximately 15 h).

Three different dye solutions were tested, the first one contained 0.2 mM LEG-4 in a 1:1 mixture of MeCN and tert-BuOH, the second one contained 0.2 mM D35 in MeCN and the third one contained 0.2 mM D35 in EtOH.

3.5.2. Photoelectrochemical polymerisation of hole-conductive polymers on dye coated TiO

2

electrodes

The hole-conducting polymers were deposited on the dye coated TiO

2

by using photoelectrochemical polymerisation. The mechanism of photoelectrochemical polymerisation is similar to the mechanism of electropolymerisation. By irradiating a dye- coated TiO

2

surface placed in a solution containing the monomer, oxidation of the dye will occur due to injection of electrons into the conduction band of TiO

2

. The oxidized dye molecule is reduced by oxidation of a monomer molecule forming a monomer radical. When the polymer chain grows by the same chain reaction as discussed in Section 3.4.1, it becomes insoluble in the solvent so that it eventually precipitates as a polymer molecule on the dyecoated TiO

2

electrode.

The setup was a three-electrode electrochemical cell (see Figure 11). The polymerisation was made by using a constant current of 20 μA for 2500 s. The light source was a white LED with a light intensity around 1% sun. The light was shining on the backside of the dye coated TiO

2 electrode. Both aqueous and organic electrolytes were tested. The compositions of the electrolytes are shown in table 3.

Table 3: The composition of the electrolytes used for the photoelectrochemical polymerisation of PEDOT and poly-PheDOT

Polymer

PEDOT

Composition of organic electrolyte

0.1 M LiTFSI and 5 mM bis-

EDOT in MeCN

Composition of aqueous electrolyte

-

Poly-PheOT 0.1 M LiTFSI and 1 mM bis-

PheDOT in a 1:1 mixture of

MeCN and DCM

0.1 M LiTFSI, 0.05 M

Triton-X100, 0.5 mM bis-

PheeDOT in H

2

O

23

Figure 11: The setup for the photoelectrochemical polymerisation. (1): working electrode

(dye coated TiO

2

film), (2): counter electrode (stainless steel electrode), (3): reference

electrode (Ag/AgCl electrode). The light source was shining on the backside of the dye coated

TiO

2

electrodes.

3.5.3. Post treatment of polymer coated electrodes

After the photoelectrochemical polymerisation, the polymer coated TiO

2

electrodes were carefully washed with EtOH. A few drops of a treatment solution containing TBP (4-tert butylpyridine) and LITFSI (lithium bis-trifluoromethanesulfonyl-imide) was smeared out on the polymer film and the solvent was evaporated by spin-coating the samples for 30 s with a rotation speed of 2000 rotations/s. By doing this step, we can increase the efficiency of the final solar cell. The Li

+

ions in LiTFSI can intercalate into the TiO

2

and cause a positive shift in the conduction band of TiO

2

. This shift will cause a larger energy difference between the

LUMO level of the dye and the conduction band edge of TiO

2

which will give a higher driving force for electron injection into the TiO

2

. A higher driving force for electron injection will automatically give a higher current from the solar cell. The TBP can block the contact between the TiO

2

and the polymer electrolyte, which will prevent recombination and cause a negative shift in the conduction band of TiO

2

. This shift will improve the overall voltage from the solar cell.

The treatment solution for the PEDOT electrodes was 0.01 M LiTFSI and 0.18 4-tertbutylpyridine in EtOH. The treatment solution for poly-PhEDOT electrodes was 0.01 M

LiTFSI in EtOH (for the electrodes that was polymerised in organic electrolyte) and 0.01 M

LITFSI in H

2

O (for the electrodes that was polymerised in aqueous electrolyte). No 4-tertbutylpyridine was added to the treatment solution of the poly-PheDOT electrodes because the first experiment showed that the poly-PheDOT film is (for unknown reasons) destroyed when

4-tert-butylpyridine is added.

24

3.5.4. Photoinduced absorption spectroscopy (PIA) of D35 and LEG-4 electrodes coated with polymer.

By doing PIA on polymer coated electrodes and comparing the spectra with PIA spectra for electrodes with dye only (see Section 3.2.2.2.), we can obtain information about differences in the efficiency of the regeneration of the dye. Efficient electron transfer from the holeconductive polymer to the oxidised dye is necessary to obtain an efficient solar cell.

After the post treatment of the polymer coated electrodes, the PIA spectrum of LEG-4 and

D35 electrodes coated with polymer was recorded using the setup in Figure 9. The measurements were made using the same parameters as in Section 3.2.2.2. The light was shining on the back-side of the electrode (not the side coated with polymer).

3.5.5. Ag counter-electrode preparation

The Ag counter electrode was prepared by depositing a 200 nm Ag film on the polymer coated electrode by using Ag evaporation under vacuum condition using a pressure of 5×10

-5 mbar. This step completes the solar cell making. The solar cell is now ready for characterization.

3.6. Investigation of the solar cells.

3.6.1. J-V characterization.

One of the most important parameters of a solar cell is the efficiency (η). The efficiency can be determined by first applying a voltage and measure the outcoming current density (A*cm

-

2

) on a solar cell that is irradiated with light. By changing the voltage (V) linearly over time and measure the obtained current density (J) as a function of the applied potential, we can obtain a J-V curve. A typical J-V curve is shown in Figure 12.

25

Figure 12: A typical J-V curve of a solar cell

[10]

.

Here, V

oc

is the open-circuit voltage which is the voltage over the solar cell when no current is passing through the external circuit (J=0). Theoretically, V

oc corresponds to the energy difference between the quasi-Fermi level of the TiO

2 and the Fermi level corresponding to the redox potential of the hole-conductive polymer. In reality, there is always some energy-loss due to electric resistance in the solar cell, so the experimental value of V

oc is usually lower than the theoretical value. J

sc

is the short-circuit density which corresponds to the current density when no potential is applied (V=0). FF is the fill factor which is a measure of how much energy that is lost due to electric resistance in the solar cell and P

max is the maximum power from the solar cell. V

oc

, J

sc

and P

max

can be determined from the J-V curve and by irradiating the solar cell with a light source with a known light intensity (P in

), we can determine the efficiency (η) by using equations (9) and (10) given below.

P

max

P in

FF

V oc

j sc

P in

FF

V

P

max

oc j sc

( 9 )

( 10 )

The performance of the solar cells was measured by using the Keithley 2400 computational software together with a 300 W xenon lamp with a light intensity corresponding to 1 sun

(1000 W/m

2

). The irradiated area of the solar cell was 0,2 cm

2

. The current density from the solar cell was measured as a function of the applied potential, and from the obtained J-V curve, V oc

, J sc

, FF and η were determined. It should also be mentioned that the measurements were performed immediately after the solar cells were made. The data presented in Table 4 and 5 shows the data from the overall best-performing solar cells.

26

4. Results and discussion.

4.1. Results from cyclic voltammetry of TiO

2

blocking layers in EtOH electrolyte.

4.1.1. Results from the calibration of the Ag/AgCl reference electrode in EtOH.

The cyclic voltammograms for the Ag/AgCl reference electrode calibration are shown in

Figure 13-14.

Figure 13: Cyclic voltammogram of 0.2 M LiTFSI in EtOH using a glassy carbon electrode with an area of 0.070 cm

2

. The scan rate was set to 0.050 V/s.

Figure 14: Cyclic voltammogram of 0.2 M LiTFSI and 2 mM ferrocene in EtOH using a glassy carbon working electrode with an area of 0.070 cm

2

. The scan rate was set to 0.050

V/s.

27

As can be seen in Figure 13, we obtain no peaks for LITFSI in EtOH without ferrocene . The obtained current is only a capacitive background current, which does not result from redox reactions in the electrolyte. Therefore, the peaks in Figure 14 must correspond to oxidation and reduction of ferrocene. We obtain a positive peak around 0.49 V which corresponds to the oxidation of ferrocene. We also have a negative peak at 0.39 V which corresponds to the reduction of oxidized ferrocene. The redox potential of ferrocene vs. the Ag/AgCl electrode is therefore (0.39+0.49)/2=0.44 V. From literature, we also know that the redox potential of ferrocene is 0,624 V

[24]

vs. SHE. Therefore, the redox potential of the Ag/AgCl electrode must be 0.624-0.44=0.184 V. We can now plot all our cyclic voltammograms vs. the SHE by adding 0.184 V to all potentials in the measured potential range.

4.1.2. Results from the cyclic voltammetry of blocking layers in EtOH.

The cyclic voltammograms for the blocking layers in the EtOH electrolyte are shown in

Figure 15-18.

Figure 15: Cyclic voltammogram of 0.2 M LiTFSI and 2 mM ferrocene in EtOH using a FTO glass working electrode with an area of 1.5 cm

2

and a scan rate of 0.050 V/s.

28

Figure 16: Cyclic voltammogram of FTO glass with blocking layer with varying thickness. 5 cycles (black), 8 cycles (red), 10 cycles (blue), 12 cycles (green) in 0.1 M LITFSI and 5 mM ferrocene in EtOH. The area of the electrodes was 1.5 cm

2 and the scan rate was 0.050 V/s

As can be seen in Figure 15, one obtain an oxidation peak and reduction peak when we use an

FTO glass electrode without a blocking layer as working electrode, which indicates that oxidation and reduction of ferrocene can take place on the surface of the FTO glass.

The peak separation for the FTO glass voltammogram (Figure 15) is larger than that for the peak separation of the glassy carbon electrode voltammogram (see Figure 15) primarily due to a slower electron transfer reaction on FTO and secondly due to the higher ohmic resistance of the FTO electrode. The higher resistance results in a larger overpotential for the oxidation and reduction of ferrocene, which results in larger peak separation. When we, however, use FTO glass with a blocking layer (see Figure 16), the oxidation and reduction peaks disappear and the current density becomes much lower. This behavior shows that the blocking layer can stop the oxidation and reduction of ferrocene from taking place on the electrodes. By comparing the voltammograms in Figure 16, we can see that the sample with 5 cycles (black curve) gives higher current density than the samples with 8, 10 and 12 cycles (red, blue and green curves).

The reason for this can be that the quality of the blocking layer on this sample was worse than the quality of the blocking layers on the other samples.

29

Figure 17: Cyclic voltammograms for the 20 CV cycles on a FTO glass with 10 blocking layer cycles. The working area of the electrode was 1.5 cm

2

.

Figure 18: Cyclic voltammograms for the 20 CV cycles on a FTO glass with 12 blocking layer cycles. The area of the electrode was 1.5 cm

2 and the scan rate was 0.050 V/s.

As can be seen in Figure 17 and 18, the shape of the cyclic voltammograms does not change much during the 20 cycles and no oxidation or reduction peaks appear which indicates that the blocking layers have good stability in EtOH.

30

4.2. Results from cyclic voltammetry of TiO

2

blocking layer in PC electrolyte

4.2.1. Results from Ag/AgCl reference electrode calibration in PC electrolyte.

The cyclic voltammograms for the Ag/AgCl reference electrode calibration in PC are shown in Figure 19-20.

Figure 19: Cyclic voltammogram of 0.2 M LITFSI in PC using a glassy carbon electrode

.

Figure 20: Cyclic voltammogram of 0.2 M LiTFSI and 5 mM ferrocene in PC using a glassy carbon working electrode.

As can be seen in Figure 19, no peaks for LITFSI in PC without ferrocene are obtained. The obtained current is only a capacitive background current that does not come from redox reactions in the electrolyte. Therefore, the peaks in Figure 20 must correspond to oxidation and reduction of ferrocene. We obtained a positive peak around 0.49 V which corresponds to the oxidation of ferrocene. We also have a negative peak at 0.39 V which corresponds to the reduction of ferrocene. The redox potential of ferrocene vs. the Ag/AgCl electrode is the mean value and therefore it must be (0.39+0.49)/2=0.44 V vs. the Ag/AgCl electrode. From

31

literature, we also know that the redox potential of ferrocene is 0.624 V vs. the standard hydrogen electrode

[21]

. Therefore, the redox potential of the Ag/AgCl electrode must be

0.624-0.44=0,184 V. We can now plot all our cyclic voltammograms vs. the standard hydrogen electrode by adding 0.184 V to all potentials in the measured potential range.

4.2.2. Results from cyclic voltammetry of blocking layer in PC.

The cyclic voltammograms for the blocking layers in PC electrolyte are shown in Figure 21-

24.

Figure 21: Cyclic voltammogram of 0.2 M LiTFSI and 5 mM ferrocene in PC using a FTO glass working electrode

.

Figure 22: Cyclic voltammograms of FTO glass coated with blocking layer of varying thickness Black=5 cycles, Red=8 cycles, Blue=10 cycles, Green=12 cycles. 0.2 M LITFSI and

5 mM ferrocene in propylene carbonate. Scan rate 0.050 V/s. Electrode area 1.5 cm

2

.

32

As can be seen in Figure 21, one obtains an oxidation peak and reduction peak when one uses an FTO glass electrode without a blocking layer which indicates that oxidation and reduction of ferrocene can take place on the surface of the FTO glass. As discussed in Section 4.1.2, the reason that the peak separation in the cyclic voltammogram of FTO glass is larger than the peak separation in the cyclic voltammogram of glassy carbon is that the electron transfer reaction is slower on FTO glass and the ohmic resistance of the FTO glass is higher. The higher resistance gives a higher overpotential for the oxidation and reduction of ferrocene and that results in a larger peak separation. When we, however, use FTO glass as electrode substrate with a blocking layer (see Figure 22), the oxidation and reduction peaks disappear and the current density becomes much lower. Our conclusion is that the blocking layer can stop the oxidation and reduction of ferrocene. The voltammograms overlap, so it is hard to draw any conclusion about how the thickness (number of cycles) affects the quality of the blocking layer. We do not see any trend indicating that the current density decreases when the thickness of the blocking layer increases.

Figure 23: Cyclic voltammogram for the 20 CV cycles on a FTO glass with 10 blocking layer cycles in 0.2 M LiTFSI and 5 mM ferrocene in PC. Scan rate: 0.050 V/s

33

Figure 24: Cyclic voltammogram for the 20 CV cycles on a FTO glass with 12 blocking layer cycles in 0.2 M LiTFSI and 5 mM ferrocene in PC. Scan rate: 0.050 V/s.

As can be seen in Figure 23 and 24, the cyclic voltammograms does not change much during the 20 cycles and no oxidation and reduction peaks appears during the experiments which indicates that the blocking layers have good stability in PC.

4.3. Results from cyclic voltammetry in aqueous electrolyte

4.3.1. Results from Ag/AgCl reference electrode calibration in ferrocyanide.

The cyclic voltammogram för the Ag/AgCl reference electrode calibration in the H

2

O electrolyte is shown in Figure 25:

Figure 25: Cyclic voltammogram of 0.5 M KCl, 2 mM K

4

Fe(CN)

6

in water by using a glassy carbon electrode with an area of 0.070 cm

2

and a scan rate of 0.050 V/s.

34

As can be seen in Figure 26, we obtain a positive peak at 0.269 V which corresponds to the oxidation of Fe

2+

to Fe

3+

( Fe(CN)

6

4-

 Fe(CN)

6

3-

+ e

-

) and a negative peak at 0.195 V which correspond to the reverse reaction ( Fe(CN)

6

3-

+ e

-  Fe(CN)

6

4-

). The redox potential of

Fe

2+

/Fe

3+

must therefore be (0.269+0.195)/2=0.232 V vs. the Ag/AgCl electrode. According to literature, we also know that the redox potential of Fe(CN)

6

3-

/ Fe(CN)

6

4-

is 0.356 V

[24]

vs. the standard hydrogen electrode. Therefore, the potential of the Ag/AgCl electrode must be

0.356-0.232=0.124 V vs. SHE. One can now plot all our cyclic voltammograms vs. the standard hydrogen electrode by adding 0.124 V to all potentials between -0.8 and 0.8 V.

4.3.2. Results from cyclic voltammetry in aqueous electrolyte.

The cyclic voltammograms for the blocking layers in H

2

O electrolyte are shown in Figure 26-

30.

By comparing Figure 27 and 30 (black curves) we can see that one obtain no oxidation and reduction peaks for ferrocyanide when we use the FTO glass electrode with a blocking layer made by spray pyrolysis (see Figure 27), which means that this blocking layer is able to block the oxidation and reduction of ferrocyanide. The large negative peak that appears at negative potentials in the voltammograms in Figure 28 is probably caused by the reduction of dissolved oxygen superimposed on the TiO

2

substrate reduction (Ti

IV

to Ti

III

). The peak is shifted to more negative potentials when the thickness of the film (number of cycles) increases. An increasing thickness of the film gives a higher resistance which gives a higher overpotential and a more negative reduction potential for electrochemical reactions. As we can see in Figure 29 and 30, no oxidation and reduction peak for ferrocyanide appears during the 20 cycles, which indicates that the blocking layer has good stability in aqueous electrolyte.

Figure 26: FTO glass in 0.5 M KCl, 2 mM K

4

Fe(CN)

6

using a 1.5 cm

2

FTO glass electrode and a scan rate of 0.050 V/s.

35

Figure 27: Cyclic voltammogram for FTO glass with 8 (black), 10 (red) and 12 (blue) blocking layers cycles made by spray pyrolysis in 0.5 M KCl, 2 mM K

4

Fe(CN)

6

. Scan rate:

0.050 V/s.

Figure 28: Cyclic voltammogram for the 20 CV cycles on a FTO glass with 10 blocking layer cycles in 0,5 M KCl and 2 mM K

4

Fe(CN)

6

in H

2

O, pH=4.5. Scan rate: 0.050 V/s.

36

Figure 29: Cyclic voltammogram for the 20 CV cycles on a FTO glass with 10 blocking layer cycles in 0.5 M KCl and 2 mM K

4

Fe(CN)

6

in H

2

O, pH=4.5. Scan rate: 0.050 V/s.

Figure 30: Cyclic voltammogram of a 1.5 cm

2

FTO glass (black) and FTO glass with blocking layer (blue) made by 2 M TiCl

4

treatment for 2 h at 70 o

C. The scan rate was 0.050

V/s.

When we, however, use the FTO glass with a blocking layer made by 2 M TiCl

4

treatment

(see Figure 31), we obtain both an oxidation and reduction peak of ferricyanide which means that oxidation and reduction of Fe is possible on this electrode. The peak currents are lower than for pure FTO glass, but it is on the other hand much higher than the current which was

37

obtained for the blocking layer made by spray pyrolysis. Our conclusion is that the blocking layers made by TiCl

4

cannot stop the oxidation and reduction of ferro/ferricyanide, which indicates that the quality of these blocking layers are much worse than the blocking layers prepared by using spray pyrolysis. For this reason, we will use spray pyrolysis as our preparation method when we make the blocking layers in our solar cells.

4.4. Results from cyclic voltammetry of dye coated TiO

2

electrodes and dye solutions.

All cyclic voltammograms for the dye electrodes and dye solutions are shown in Appendix 1.

One can see that the voltammograms are different on TiO

2

electrodes and in solution. The reason for this behavior is that when the dye molecules are in solution, they are completely dissolved in the electrolyte, as opposed to when the dyes are adsorbed on the TiO

2

electrode,

The dye electrodes only show one oxidation peak, (except from MK2), but in most of the dye solutions (except L0), two oxidation peaks are observed which indicates that two oxidation steps are possible in solution. Theoretical studies, such as MO calculations can be made to find out exactly what electron transfer reaction that give rise to the peaks in the voltammograms for each dye, but in this work, no such studies were made

For the dye solution voltammograms, the peak potentials are independent of scan rate, but for the TiO

2

electrodes, the oxidation potential becomes more positive when the scan rate increases. For a reversible redox reaction, the peak potential should be independent of scan rate. Therefore, we can say that the dye solutions show better reversibility than the dye electrodes. One reason for this behavior is that the electron diffusion in the TiO

2

film is slow.

The oxidized dye molecule will desorb from the TiO

2

electrode surface before it can be reduced during the reverse scan. The electron diffusion on the glassy carbon is faster and therefore, reduction of the oxidized dye molecules occurs during the reverse scan before they have desorbed from the electrode surface.

The reversibility of the reactions differs between the dyes. D35 and LEG-4 show both oxidation and reduction peaks on both electrodes and in solution which indicates good reversibility of the oxidation and reduction of these dyes. On the other hand, some of the other dyes such as K77 and N3 only show an oxidation peak and no reduction peak. The reason for the irreversibility is probably fast diffusion of the oxidized dye molecules from the TiO

2

or chemical decomposition of the oxidized dye molecules.

38

4.5. Results from spectroscopic studies of dye coated electrodes.

The UV-Vis spectra of the dye coated electrodes are shown in Figure 31-32 and the PA spectra for dye coated electrodes are shown in Figure 33-34.

By comparing the UV-Vis spectra for the organic dyes and the Ru dyes (Figure 31 and 32), we can see that the Ru-dyes have generally broader absorption spectra than the organic dyes.

The absorbance of the Ru dyes is, however, lower. Since the absorbance is directly proportional to the extinction coefficient (ε), we can conclude that the Ru dyes have lower extinction coefficients than the organic dyes.

4

3

2

1

0

350 450 550 650

Wavelength (nm)

750

D35

LEG-4

MK2

L0

D21

Figure 31: UV-Vis spectra for the organic dyes

3

2

1

0

350 450 550 650

Wavelength (nm)

750

K77

Z-907

N3

Figure 32: UV-Vis spectra for the Ru dyes.

39

0,0004

0,0002

0

500

-0,0002

-0,0004

-0,0006

600 700 800 900 1000

D21

D35

L0

LEG-4

MK2

-0,0008

Wavelength (nm)

Figure 33: PIA spectra for the organic dyes.

0,00015

0,0001

0,00005

0

-0,00005

450 550 650 750 850 950

K77

N3

Z-907

-0,0001

-0,00015

-0,0002

Wavelength (nm)

Figure 34: PIA spectra of the Ru dyes.

As can be see in Figures 33-34, all dyes in the PIA spectra show a negative peak somewhere between 540 nm and 630 nm. This peak is caused by the Stark effect

[23]

. The Stark effect is a shift of the spectral lines of an atom or molecule caused by an external electric field. When the dye absorbs light upon illumination, electrons are injected into the TiO

2

. The injected electrons and the positively charged oxidized dye molecules form an electric field that leads to the Stark shift of the neutral dye molecules in the dye film. The dyes also show positive absorption (ΔA > 0) for higher wavelengths (λ > 650 nm). This signal originates from the oxidized dye

[23]

, which indicates that all the tested dyes can inject electrons into TiO

2

.

On the basis of the spectroscopic and the electrochemical measurements on the dyes, I decided to use LEG-4 and D35 in the solar cells. The reason for choosing these dyes is that their UV-Vis spectra show broad and high absorption in the UV-Vis range, and the PIA

40

measurements indicate that electrons can be injected into the TiO

2

film. Finally, the electrochemical measurements show that the redox potentials for these dyes are sufficiently positive for regeneration by the hole-conductor (around 1,1 V vs. SHE). Even though, the

MK2 dye has both broader absorption spectrum (Figure 32) and also a more positive redox potential (Figure 66) compared to D35 and LEG-4, it was not investigated in the solar cells due to very low amounts available in the lab.

4.6. Results from electropolymerisation and characterization of conductive polymer on glassy carbon electrode

4.6.1. Electropolymerisation and characterization of PEDOT on glassy carbon electrode.

The cyclic voltammograms for the electropolymerisation of PEDOT in EtOH are shown in

35-38.

No oxidation or reduction peak is observed in the cyclic voltammogram for a 0.1 M LiTFSI solution (Figure 35). The obtained current is only a capacitive background current, which does not originate from redox reactions in the electrolyte. This indicates that the electrolyte is stable in the potential range used for the electropolymerization (-0.5 V to 1.0 V).

Figure 35: Cyclic voltammogram of 0.1 M LITSFI in acetonitrile (without bis-EDOT monomer). The reference electrode was an Ag/AgCl (2 M LiCl in EtOH) electrode and the scan rate was 0.050 V/s.

41

Figure 36: Cyclic voltammogram of 0.1 M LiTFSI and 5 mM bis-EDOT in acetonitrile. The reference electrode was an Ag/AgCl (in 0.1 M LiTFSI in acetonitrile) electrode and the scan rate was 0.050 V/s.

As can be seen in Figure 36, a positive peak is obtained around 1.1 V. Since no peak is obtained in the cyclic voltammogram of LITFSI in acetonitrile in the absence of added electroactive species, we can conclude that the peak does not come from oxidation of acetonitrile or LiTFSI. Therefore, the peak must correspond to the oxidation of precursor bis-

EDOT. No reverse peak is obtained, so we can conclude that the oxidation of bis-EDOT is an irreversible reaction. The reason for the irreversibility is probably caused by that the oxidation product of bis-EDOT is a radical, which will react with a new monomer, before it can be reduced back to the precursor.

42

Figure 37: Cyclic voltammograms for the electropolymerisation of PEDOT (20 cycles) using a glassy carbon electro with an area of 0.070 cm

2

and a scan rate of 0.050 V/s.

Figure 38: Cyclic voltammograms for the 20 cycles electropolymerisation of bis-EDOT using an 0.1 M LiTFSI and 5 mM bis-EDOT in MeCN electrolyte. The figure shows the cyclic voltammograms for the 1 st

(black), 10 th

(red) and 20 th

(blue) cycle. The scan rate was set to

0.050 V/s

By comparing the cyclic voltammograms for the 1 st

, 10 th

and 20 th

cycle in Figure 38, one can see that the current density increases with the number of cycles, which indicates that the thickness of generated polymer increases upon prolonged cycling. We can also see that a very broad oxidation peak with a maximum around 0.4 V appears when we increase the number of scans together with a reduction peak at -0.2 V when we increase the number of cycles. These peaks corresponds to the oxidation and reduction of the formed polymer respectively

43

Figure 39: Cyclic voltammograms for the polymer coated glassy carbon electrode in 0.1 M

LiTFSI in MeCN. The blue curve corresponds to the first scan cycle.

By comparing the voltammograms in Figure 39 we can see that the voltammogram for the first scan (blue curve) has a different shape than the voltammogram for the 9 last scans (black curves). The reason is probably that the polymer layer is not fully equilibrated with the inert electrolyte during the first scan cycle. The peak height also decreases with the number of scans. This is probably caused by slow dissolution of the polymer in the electrolyte.

However, the position of the peaks in the voltammogram does not change upon continuous cycling, which indicates the formation of an insoluble and stable conductive polymer on the surface of the glassy carbon electrode.

The cyclic voltammograms for the electropolymerisation of PEDOT in H

2

O are shown in

Figure 40-42.

No oxidation or reduction peak is observed in the cyclic voltammogram for a 0.1 M LiTFSI solution (Figure 40). The obtained current is only a capacitive background current, which does not originate from redox reactions in the electrolyte. This indicates that the electrolyte is stable in the potential range used for the electropolymerization (-0.5 V to 0.9 V).

44

Figure 40: Cyclic voltammogram of 0.1 M LITSFI, 0.05 M Triton-X100 in water. The reference electrode was an Ag/AgCl electrode with 3 M NaCl in H

2

O. The scan rate was

0.050 V/s.

Figure 41: Cyclic voltammograms of the polymerization of bis-EDOT in a 1 mM bis-EDOT

0.1 M LiTFSI, 0.05 M Triton X-100 in H

2

O electrolyte. The scan rate was set to 0.050 V/s.

As can be seen in Figure 41, the cyclic voltammogram for the polymerization in water looks different from the cyclic voltammogram from of the polymerization in MeCN (Figure 38).

The peak currents for the oxidation and reduction of the polymer that appears around 0.2 V are much lower than the peak currents in the voltammogram for the polymerization in MeCN.

The reason is that it is more difficult to form a polymer layer in an aqueous electrolyte One possible explanation is that the bis-EDOT precursor concentration in the polymerization

45

electrolyte is lower in this case due to the lower solubility in water. The presence of Triton X-

100 molecules in the electrolyte may also affect the efficiency of the polymerization.

Figure 42: Cyclic voltammograms of the polymerization of bis-EDOT in a 1 mM bis-EDOT

0.1 M LiTFSI, 0.05 M Triton X-100 in H

2

O electrolyte . The figure shows the cyclic voltammogram for the 1 st

(black), 10 th

(red) and 20 th

(blue) cycle. The reference electrode was an Ag/AgCl 3 M NaCl in H

2

O) electrode and the scan rate was 0.050 V/s.

If we compare the voltammograms for the 1 st

, 10 th

and 20 th

cycle för the electropolymerisation in H

2

O (Figure 42), we can see that when the number of polymerization cycles increases, we obtain an oxidation and a reduction peak for the formed polymer around 0.2 V. The height of the peaks increases with the number of cycles, which indicates that the thickness for generated polymer progressively increases

Figure 43: Cyclic voltammograms for the polymer coated glassy carbon electrode in 0.1 M

LiTFSI in acetonitrile. Scan rate 0.050 V/s

46

We can observe the same behavior in Figure 43 as we saw in Figure 39. The peak current decreases with number of cycles due to slow dissolution of polymer in the electrolyte.

However, the position of the peaks in the voltammogram does not change, which indicates the formation of an insoluble and stable conductive polymer on the surface of the glassy carbon electrode.

The final conclusion is that bis-EDOT can electropolymerise in both H

2

O and MeCN.

Though, it seems that the electropolymerisation works best in acetonitrile due the much higher solubility of bis-EDOT in acetonitrile. The formed polymer also shows good stability in both MeCN and H

2

O.

4.6.2. Electropolymerisation and characterization of poly-PheDOT on glassy carbon electrode.

The precursor bis-PheDOT is insoluble in MeCN but soluble in dichloromethane (DCM).

However, DCM is not suitable for photoelectrochemical polymerization at dye-coated electrodes, due to dye desorption from the TiO

2

surface. For the latter purpose, a mixed solvent 1:1 v:v MeCN/DCM is a suitable medium so that electrochemical polymerization tests were performed in both the mixed solvent and pure DCM.

The cyclic voltammograms för the electropolymerisation of poly-PheDOT in MeCN/DCM mixed solvent are shown in Figure 44-46.

Figure 44: Cyclic voltammogram of 0.1 TBAHP and 1.0 mM bis-PheDOT in a 1:1 mixture of

MeCN and DCM. Scan rate 0.050 V/s.

.

47

The oxidation peak of bis-PheDOT in a mixed solvent, 1:1 v:v MeCN/DCM (the mixed solvent is prepared by mixing equal volumes of MeCN and DCM), appears at around 1.4 V, see Figure 44. The current density is much lower than for the cyclic voltammogram of bis-

EDOT, see Figure 37. One possible reason is that the added precursor concentration is lower in this case due to the lower solubility in the MeCN/DCM solvent.

Figure 45: Cyclic voltammograms of the electropolymerisation of bis-PheDOT in a 0.1 M

TBAHP and 1mM bis-PheDOT in a 1:1 mixture of MeCN and DCM 1 st

scan (black), 10 th scan (red), 20 th

scan (blue), 30 th

scan (green) and 40 th

scan (orange). Scan rate 0.050 V/s.

Figure 46: Cyclic voltammogram of polymer coated electrode in 0.1 M TBAHP in 1:1 mixture of MeCN and DCM . 1 st

scan (black) 5 th

scan (red), 10 th

scan (blue)

48

In the cyclic voltammogram for the electropolymerisation (Figure 45), the current density increases with number of scan cycles, which indicate a progressive increase of polymer thickness. In Figure 46 (polymer characterization in inert electrolyte, without monomer), we can see that the current density decreases somewhat with time, which is caused by slow dissolution of the polymer. The peak position does not change, which indicates that the formed polymer is insoluble and stable in the mixed solvent.

The cyclic voltamograms for the electropolymerisation of poly-PheDOT in DCM are shown in Figure 47-49.

Figure 47: Cyclic voltammogram of 0.1 TBAHP and 1 mM bis-PheDOT in dichloromethane.

Scan rate 0.050 V/s.

The oxidation peak of bis-PheDOT in DCM appears around 1.6 V, see Figure 47. The current density is much lower compared to the cyclic voltammogram of bis-EDOT (see Figure 37).

One possible reason is that the added precursor concentration is lower in this case due to the lower solubility in DCM. On the other hand, the peak current of bis-PheDOT in DCM (Figure

47) is higher than the peak current of bis-PheDOT in the MeCN/DCM mixed solvent (Figure

44). The reason is that the solubility of bis-PheDOT is higher in DCM than in the mixed solvent.

49

Figure 48: Cyclic voltammograms of the electropolymerisation of bis-PheDOT in 0.1 M

TBAHP and 1 mM bis-PheDOT in dichloromethane, 1 st

scan (black), 10 th

scan (red), 20 th scan (blue),

Figure 49: Cyclic voltammogram of polymer coated electrode in 0.1 M TBAHP in dichloromethane of dichloromethane and acetonitrile. 1 st

scan (black) 5 th

scan (red), 10 th

scan

(blue)

50

One can see the same behavior in Figure 48 and 49 as was seen in Figure 46 and 47. During the electropolymerisation (Figure 49), the current density increases with the number of cycles because the thickness of the polymer film increases. In Figure 49, we can see that the current density decreases with the number of scan cycles, probably caused by slow dissolution of the polymer in the electrolyte.

Our conclusion is that electropolymerisation is possible in both DCM and a 1:1 mixture of

DCM and MeCN and the formed polymer also show good stability in these solvents.

4.7. Results from solar simulator measurements

4.7.1. Solar simulator measurements with PEDOT.

After the photoelectrochemical polymerization, a black surface coating was obtained on both the D35 and LEG-4 electrodes, which shows that the photoelectrochemical polymerization works in acetonitrile. The D35 electrodes showed more formed polymer on the surface than the LEG-4 electrodes. The polymer layers were not removed when the electrodes were washed with EtOH which indicates the formation of a stable polymer film. The polymer film became darker when the post treatment solution was placed on the surface. The results from the solar cell measurements are shown in Table 4, and the corresponding J-V curves in

Figures 50-52:

Table 4: Results from the solar cell measurements of solar cells with PEDOT as holeconductor material. The polymerization was made in 0.1 M LiTFSI and 5 mM bis-EDOT in acetonitrile

I

sc

(mA*cm

-2

)

FF

η (%)

Dye + dye bath solvent

V

oc

(V)

LEG-4 in 1:1 mixture of

MeCn and tert-

BuOH

D35 in EtOH

0.35

0.37

D35 in MeCN 0.73

7.4

7.7

7.6

0.28

0.45

0.46

0.72

1.28

2.56

51

-0,1

-5

-7

-9

1

-1

0

-3

0,1 0,2 0,3 0,4

-11

Voltage (V)

Figure 50: J-V curve for LEG-4 in MeCN/BuOH with PEDOT as hole-conducting material.

The photoelectrochemical polymerization was made in 0.1 M LITFSI and 5 mM bis-EDOT in acetonitrile.

-0,1

2,5

0,5

-1,5

0 0,1 0,2 0,3 0,4

-3,5

-5,5

-7,5

-9,5

Voltage (V)

Figure 51: J-V curve D35 in EtOH with PEDOT as hole-conducting material. The photoelectrochemical polymerization was made in 0.1 M LITFSI and 5 mM bis-EDOT in acetonitrile.

52

1

-0,1

-1

-3

-5

-7

0,1 0,3 0,5 0,7

-9

Voltage (V)

Figure 52: J-V curve for D35 in MeCN with PEDOT as hole-conducting material. The polymerization was made in 0.1 M LITFSI and 5 mM bis-EDOT in acetonitrile.

The PEDOT solar cells with D35 dye show better performance than the PEDOT solar cells with LEG-4 dye. Inspection of the electrodes showed that more PEDOT polymer had been formed on the D35 electrodes than on the LEG-4 electrodes. This is a possible reason why the

D35 solar cells showed better performance. If more polymer molecules are deposited on the electrode, we can transfer more electrons from the polymer to the dye and therefore also obtain a more efficient solar cell. According to Figure 31, LEG-4 has a broader absorption spectrum than D35, but it seems that this does not have any effect on the solar cell performance because the D35 solar cells show better performance even if D35 has a less broadened absorption spectrum than LEG-4. It is, however, still unclear exactly why more polymer is formed on D35 electrodes compared to LEG-4 electrodes.

We can also see that the dye bath solvent also had some effect on the solar cell performance.

D35 in MeCN gives higher efficiency than D35 in EtOH. One possible explanation is that when the dye molecules bind to the TiO

2

, they will maybe have different molecular configurations on the TiO

2

surface depending on the solvent. The dye configuration obtained in MeCN probably gives rise to a more effective electron transfer from PEDOT to D35 than the dye configuration obtained in EtOH.

4.7.2. Solar simulator measurements with poly-PheDOT.

When the polymerization was made in the organic electrolyte (1:1 mixture of MeCN and

DCM), very little polymer was formed on the electrodes, and the performance of the solar cells was very poor. It seems that the photoelectrochemical polymerization of poly-PheDOT does not work so well in organic electrolytes. One possible reason is that the bis-PheDOT radical is very stable due to the phenyl group which will lead to resonance stabilization and

53

low reactivity of the radical. The low reactivity will lead to formation of short polymer chains and short polymer chains are more soluble in the electrolyte and, therefore, do not precipitate out on the electrode surface to the same extent

When the polymerization was made in water, more polymer was formed on the surface. Due to the lower solubility in water, it is easier for the polymer molecules to participate on the electrode surface when we perform the polymerization in water. The polymer films became darker when the post treatment solution was added. More polymer was formed on the D35 electrode than om the LEG-4 electrode which is the same behavior that was observed in the polymerization of PEDOT. Though, the reason for this behavior is unclear

The best solar cells were thus obtained when the polymerization was made in water. The best performing solar cells with D35 and LEG-4 are shown in Table 5 and the corresponding J-V curves are shown in Figure 53-55.

Table 5: Results from the solar cell measurements of solar cells with poly-PheDOT as HTM.

The polymerization was made in 0.1 M LiTFSI, 0.05 M Triton-X100 and 0.5 mM bis-PheDOT in H

2

O

I sc

(mA*cm

-2

)

FF

η (%)

Dye + dye bath solvent

LEG-4 in 1:1 mixture of

MeCN and tert-

BuOH

V

oc

(V)

0.66

D35 in acetonitrile

0.68

D35 in ethanol 0.52

0.036

0.273

0.075

0.693

0.410

0.67

0.016

0.075

0.026

0,04

0,02

-0,2

0

0

-0,02

-0,04

0,2 0,4 0,6

-0,06

Voltage (V)

Figure 53: J-V curve for LEG-4 in MeCN/BuOH with poly-PheDOT as hole-conducting material. The photoelectrochemical polymerization was made in 0.1 M LITFSI, 0.05 M

Triton-X100 and 0.5 mM bis-PheDOT in H

2

O.

54

0

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8

-0,1

-0,2

-0,3

Voltage (V)

Figure 54: J-V curve for D35 in MeCN with poly-PheDOT as hole-conducting material. The polymerization was made in 0.1 M LITFSI and 0.5 mM bis-PheDOT in H2O

0,25

0,15

-0,2

0,05

-0,05

0 0,2 0,4 0,6 0,8

-0,15

Voltage (V)

Figure 55: J-V curve for D35 in EtOH with poly-PheDOT as hole-conducting material. The polymerization was made in 0.1 M LITFSI and 0.5 mM bis-PheDOT in H

2

O

We can see the same behavior for the solar cells with Phe-DOT that we saw for the solar cells with PEDOT. Solar cells with D35 dye show a higher efficiency compared to solar cells with

LEG-4 dye probably because more polymer is formed on the surface when the D35 dye is used. We can also see that we obtain a higher efficiency when we use a D35 in MeCN dye bath than the efficiency obtained when we used D35 in EtOH dye bath probably due to that the D35 in MeCN has a more suitable molecular configuration which gives rise to more effective electron transfer from the polymer molecules to the oxidized D35 molecules.

55

By comparing the data in Table 4 and 5, we can conclude that the solar cells with poly-

PheDOT show much lower efficiency than the solar cells with PEDOT. The reason for this behavior will be further discussed in Section 5.

4.8. PIA analysis of dyed TiO

2

electrodes coated with poly-PheDOT.

The results from the PIA measurements are shown in Figure 57 and 58:

0,00015

0,0001

0,00005

0

500

-0,00005

-0,0001

600 700 800 900 1000

D35

D35/poly-PheDOT

-0,00015

Wavelength (nm)

Figure 56: PIA spectrum of TiO

2

electrode coated with LEG-4 and TiO

2

electrode coated with

LEG-4 and poly-PheDOT

0,0003

0,0002

0,0001

0

-0,0001

550

-0,0002

-0,0003

-0,0004

750 950

LEG-4/poly-PheDOT

LEG-4

-0,0005

Wavelength (nm)

Figure 57: PIA spectrum of TiO

2

electrode coated with LEG-4 and TiO

2

electrode coated with

LEG-4 and poly-PheDOT.

56

As can be seen in Figure 56 and 57, we obtain a negative peak around 600 nm in the PIA spectra. This peak is caused by the Stark effect

[23]

. If we look at the curves which corresponds to LEG-4 and D35 without polymer, we can see that we have positive absorption (Delta A>0) above 600 nm. This corresponds to oxidized LEG-4 and D35 respectively

[23]

. If we look at the curves which correspond to the dyes together with poly-PheDOT, we can see that we still have light absorption from the oxidized dye. This is an interesting behavior. Poly-PheDOT is supposed to regenerate the dye and, if the regeneration works efficiently, we should not see any positive signal from oxidized dye at λ > 650 nm. This is an indication that the regeneration of the dye does not work very well and this is also a possible explanation why the efficiency of the solar cells with poly-PheDOT was very low.

The reason for the poor regeneration of the dye is probably due to the large size of the precursor for the photoelectrochemically generated polymer. If the monomer is large, it gives rise to steric hindrance, and the resulting polymer cannot penetrate deeply into the dyed TiO

2 film. Another reason is the very low solubility of the monomer in the water electrolyte. If the solubility is low, shorter polymer chains will be formed, which have a less positive oxidation potential than long polymer chains; therefore, the driving force for dye regeneration will be lower for the short than for the long polymer chains. Moreover, when the bis-PheDOT monomer is oxidized, it will be stabilized by resonance caused by the phenyl group in the molecule. The resonance stabilization will give low reactivity of the radical, and it will be harder to form longer polymer chains.

5. Conclusions

In this project, I have investigated two different hole-conducting polymers and their ability to be used as hole-conductive materials in solid-state dye sensitized solar cells. By using

PEDOT as hole-conductive material, I obtained 2.56% efficiency, which is still lower than the efficiency of 6% which has been obtained in earlier studies

[10]

. The reason is probably that the reproducibility of these kinds of solar cells is bad. Many factors, such as the quality of the blocking layer and the quality of the deposited polymer film on the dyed TiO

2

electrode, at present, is difficult to control and will strongly affect the performance of the solar cell. The quality of the blocking layer or polymer film may vary a lot even if one uses the same experimental procedure, which automatically results in solar cells with very varying performance. The reported values from the solar cells measurements in Table 4 only show the data from the overall best performing solar cells, but the performances for the solar cells varied, some cells had a much lower efficiency even if they were made in the exactly the same way as the high-performing solar cells.

It may be possible to obtain a higher efficiency than 6% by using PEDOT as hole-conductive material. By changing some parameters, such as the composition of the solvent for the photoelectrochemical polymerization, or the dye molecule, we may obtain higher efficiencies.

The performance of the PEDOT solar cells is, however, much better than the performance of the poly-PheDOT solar cells which was studied for the first time in this thesis. By using poly-

57

PheDOT as hole-conductive material, together with the D35 dye, I only reached a maximum of 0,075% power conversion efficiency. The solar cells with poly-PheDOT showed very low

J sc

values, which is an indication on that one or several of the electron transfer processes in the solar cell do not work efficiently. The PIA spectrum of the poly-PheDOT coated electrodes looked very similar to the PIA spectrum for the dyed coated electrodes, which is an indication that the regeneration of the dye is not so effective. One possible reason for the poor dye regeneration is that the bis-PheDOT monomer is larger than the bis-EDOT monomer, which gives rise to more steric hindrance. Therefore, it will be hard for the poly-PheDOT molecule to penetrate the dyed TiO

2

film, and dye regeneration will be more difficult.

Another problem may be the very low solubility of the monomer in water. The low solubility will result in very short polymer chains precipitating on the electrode surface. Short polymer chains have a less positive oxidation potential than long polymer chains; therefore, the driving force for dye regeneration will be lower. Resonance stabilization of the monomer radical by the phenyl group is another possible reason why we did not obtain longer polymer chains.

If one wants to improve the performance of solar cells based on poly-PheDOT as holeconductive material, one needs to find methods to improve the regeneration of the dye. Some suggestions for future work are to try to do the photoelectrochemical polymerization in another solvent. If we choose a solvent where the monomer has a little higher solubility, we may obtain longer polymer chains which will penetrate the TiO

2

film more deeply and, consequently, can increase the regeneration rate. Another suggestion is to use another dye molecule. In this project, we saw that the MK2 dye had both a broader absorption spectrum

(Figure 31) and more positive redox potential (Figure 65) which gives a higher driving force for dye regeneration. This dye, may thus improve the performance of the solar cells. Since the monomer is very large, another suggestion is to use a smaller dye molecule so that to obtain a lower steric resistance between the dye and the poly-PheDOT molecules.

By replacing the solvent in the photoelectrochemical polymerization and also replacing D35 and LEG-4 with other dyes, we may obtain more efficient solar cells with poly-PheDOt than we did in this project.

6. Summary in non-scientific form

It is a well-known fact that there is an urgent need to replace the fossil fuels by other renewable energy sources in order to stop the emission of greenhouse gases. One possible alternative is to use solar cells for production of electrical energy. By using solar cells we can convert sunlight directly into electricity. The best-performing solar cells, including siliconbased ones, have an efficiency over 20%

[5]

. The production cost of these solar cells is, however, very high, so that a lot of research is devoted to develop cheaper alternatives, one which is the dye- sensitized solar cell.

The dye-sensitized solar cell is considered as one of the most promising candidates due to its low cost and relatively high efficiency. The main component of a dye sensitized solar cell (the photoelectrode) is a thin film of titanium dioxide deposited on a conductive glass electrode.

58

The titanium dioxide is covered with a thin dye layer. Natural dyes (from fruit and berries) can be used, but artificially-made dyes are most commonly preferred, since their structure is designed so that the performance in solar cells is optimized. The photoelectrode is connected to another electrode (the counter electrode). The counter electrode is not necessarily a glass electrode. In some types of solar cells, evaporated silver metal can serve as a counter electrode. Between the electrodes, a thin film of a charge-transport medium is inserted. The charge-medium is usually a liquid containing a redox couple. The charge-transporting medium can also be a solid material. A typical example of a dye-sensitized solar cell is to use a dye coated TiO

2

as photoelectrode, a platinum counter electrode and acetonitrile (MeCN) with a dissolved I

3

-

/I

-

or Co

2+

/Co

3+

redox couple as electrolyte. When the solar cell is irradiated, the dye molecules absorb the sunlight and inject an electron into the TiO

2

layer.

The electron will move through the TiO

2

layer and towards the counter electrode via an external circuit. The dye will compensate for the lost electron by taking up a new electron from a species in the electrolyte, which in turn will receive a new electron from the counter electrode so that no net change occurs in the charge-transport medium. Overall, these processes will generate a continuous electric current and this is how we produce electricity with a solar cell.

One problem with dye sensitized solar cells with a liquid electrolyte is the low boiling point of the electrolyte, which makes it hard to use these solar cells, especially at higher temperatures. Another problem is the toxicity of MeCN. One way to overcome this problem is to replace the liquid electrolyte with a solid conductive polymer.

In this project, I have investigated two different hole-conductive polymers and their possible application as hole-conductive materials in a solid dye sensitized solar cell. The first polymer,

PEDOT (poly 3,4-ethylenedioxythiophene), has already been used in solar cell applications giving about 6% power conversion efficiency

[10]

. In this project, I investigated if it was possible to obtain even higher efficiency with this polymer. The second polymer that was investigated, poly-PheDOT (poly-phenylenedioxythiophene) has still not been investigated in solar cell applications, so this was the first test. By using PEDOT together with D35 dye in a solar cell, I managed to obtain 2.56% efficiency, which is lower than which has been reported in earlier studies, but it still shows that PEDOT works quite well in solar cells applications.

By using poly-PheDOT together with D35 dye, however, I only obtained 0.075% efficiency.

One possible reason for the low efficiency of the poly-PheDOT solar cell is that the electron transfer reaction between the poly-PheDOT molecules and the dye molecule does not work very efficiently. By optimizing the experimental conditions during the deposition of the polymer film on the dye electrode, we may overcome this problem and obtain more efficient solar cells.

59

7. Acknowledgments

Firstly, I would like to thank my main supervisor Nick Vlachopoulos. Thank you so much for giving me the opportunity to do my master thesis together with you and also for your useful comments

I would also like to thank Jinbao Zhang. Thank you very much for all your help in the lab during this semester, I really appreciated it. Thanks to all other people in Anders Hagfeldt group, it has been so inspiring and informative to do my thesis in your group. You have all been so helpful and friendly.

To all my friends, thank you so much for all funny parties and crazy activities during this semester. You are all awesome. To my family, thank you so much for all support and delicious food. You have saved me from starving to death many times.

8. References

[1]: Solar Energy Perspectives: Executive Summary". - International Energy Agency, 2011

Archived from the original on 2011-12-03.

[2]: Solar energy: A new day dawning? Silicon Valley sunrise - Access Nature 2006, 443,

19-22

[3]: http://en.wikipedia.org/wiki/File:Breakdown_of_the_incoming_solar_energy.svg 2014-

01-31

[4]. D.M. Chapin, C.S. Fuller, G.L. Pearson. - J. Appl. Phys., 1954, 25, p. 676

[5]: C.D. Grant, A.M. Schwartzberg, G.P. Smestad, J. Kowalik, L.M. Tolbert, J.Z. Zhang, J. -

Electroanal. Chem. 2002, 522, 40

[6]: B. O’Regan , M. Grätzel , Nature 1991 , 353 , 737 .

[7]: H. J. Snaith and L. Schmidt-Mende. - Adv. Mater., 2007, 19, 3187.

[8]: http://www.nature.com/am/journal/v1/n1/full/am200914a.html - 2014-02-13

[9]: Y. Cao, Y. Bai, Q. Yu, Y. Cheng, S. Liu, D. Shi, F. Gao and P. Wang, - J. Phys. Chem.

2009, 113, 629

[10]: L. Yang, M. Jouini, J. Zhang, Y. Shen, B. Wook Park, D. Bi, L. Häggman, E. M. J.

Johansson, G. Boschloo, A. Hagfeldt, N. Vlachopoulos, A. Snedden, L. Kloo, A. Jarboui, A.

Chams, C. Perruchot – J Phys. Chem. Letters 2013, 4, 4026-4031

[11]: Photoelectrodeposited polymers as hole-conductors for solid state dye-sensitized solar cells. Yang Shen, Master thesis at Department of Chemistry, Ångström laboratory 2012

60

[12]: I. F. Perepicka, S. Roquet, P. Leriche, J. M. Raimundo, P. Frère and J. Roncali – Chem.

Eur. J. 2006, 12, 2960-2966.

[13]: S. Mathew, A. Yella, P. Gao, R. Humphry-Baker, B. F. E. Curchod, N. Ashari-Astani,

I. Tavernelli, U. Rothlisberger, M. K. Nazeeruddin & M. Grätzel. - Nature Chemistry, 2014,

6, 242-247

[14]: http://www.dyenamo.se/dyenamo/images/L0.jpg 2014-04-14

[15]: K. Oum, P. W. Lohse , J. R. Klein , O. Flender , M. Scholz , A. Hagfeldt , G. Boschloo and T. Lenzer. - Phys. Chem. Chem. Phys., 2013, 15, 3906-3916

[16]: http://www.sigmaaldrich.com/catalog/product/aldrich/728705?lang=en&region=SE

2014-05-11. 2014-05-19

[17]: S. Lu, T. Wu, B. Ren, R. Geng - J Mater Sci: Mater Electron, 2013, 24, 2346–2350

[18]: J. Zhang, L. Häggman, M. Jouini, A. Jarboui, G. Boschloo, N. Vlachopoulos and A.

Hagfeldt. - Chem. Phys. Chem, 2014, 15, 1043–1047.

[19]: http://www.pvmaterials.umicore.com/productsServices/OrganoMetallicComplexes/

2014-03-16

[20]: K. Murakoshi, R. Kogure, Y. Wada, S. Yanagida - Solar Energy Materials and Solar

Cells, 1998, 55, 113–125

[21]: Gordon Aylward, Tristan Findlay – SI Chemical Data 6 th

edition, p. 142, ISBN: 0-470-

81638-4,

[22]: S. Roquet , P. Leriche, I. Perepichka, B. Jousselme , E. Levillain , P. Frère and Jean

Roncali - J. Mater. Chem., 2004, 14, 1396-1400.

[23]: S. M. Feldt, E. A. Gibson, E. Gabrielsson, L. Sun, G. Boschloo, A. Hagfeldt – J. Am

Chem. Soc. 2010, 132, 16714-16724

[24]: C. H. Hamann, Andrew Hamett, Wolf Vielstich – Electrochemistry, 2 nd

edition IBSN:

978-527-31069-2

[25]: https://ww2.chemistry.gatech.edu/reynolds/research/electrochemistry -2014-05-26

[26]: R. Kawano, H. Matsui, C. Matsuyama, A. Sato, Md. Abu Bin Hasan Susan, N. Tanabe,

M. Watanabe. - Journal of Photochemistry and Photobiology, Chemistry, 2004, 06

[2 ]: ay, A. Grätzel, M. Sol. Energy Mater. Sol. Cells 1996, 44, 99−117.

[28]: Wu, M. X.; Lin, X.; Wang, T. H.; Qiu, J. S.; Ma, T. L. Energy Environ. Sci. 2011, 4,

2308−2315.

61

[29]: Roy-Mayhew, J. D.; Bozym, D. J.; Punckt, C.; Aksay, I. A. ACS. Nano 2010, 4,

6203−6211.

[30]: Lee, K. M.; Hsu, C. Y.; Chen, P. Y.; Ikegami, M.; Miyasaka, T.; Ho, K. C. Phys. Chem.

Chem. Phys. 2009, 11, 3375−3379.

[31] Tai, Q.; Chen, B.; Guo, F.; Xu, S.; Hu, H.; Sebo, B.; Zhao, X.-Z.. ACS Nano 2011, 5,

3795−3799.

[32]: C. H. Law, S. C. Pathirana, X. Li, A. Y. Anderson, P. R. Barnes, A. Listorti, T. H.

Ghaddar and B. C. O’Regan, Adv. Mater. 2010, 22, 4505.

[33]: http://elchem.kaist.ac.kr/vt/chem-ed/spec/uv-vis/uv-vis.htm 2014-03-18

62

9. Appendix 1

Results from cyclic voltammetry experiments on dye coated electrodes and dye

solutions.

Figure 58: Cyclic voltammogram of LEG-4 dye coated electrode using 0.05 V/s (black) and

0.5 V/s (blue). The electrode area was 0.55 cm

2

. The scan range was 0 V to 1.3 V.

Figure 59: Cyclic voltammogram of 0.2 mM LEG-4 solution using different scan rates. 0.05

V/s (black line), 0.2 V/s (red line) and 0.5 V/s (blue line).

63

Figure 60: Cyclic voltammogram of D35 for the using a dye coated TiO

2

electrode with area of 0.50 cm

2

. The scan rates were 0.05 V/s (black), 0.2 V/s (red) and 0.5 V/s (blue). The scan range was 0 V to 1.3 V.

Figure 61: Cyclic voltammogram of 0.2 mM D35 solution using different scan rates. 0.05 V/s

(black line), 0.2 V/s (red line) and 0.5 V/s (blue line).

64

Figure 62: Cyclic voltammogram of K77 dye coated electrode using 0.05 V/s (black) and 0.5

V/s (blue). The electrode area was 0.50 cm

2

.

Figure 63: Cyclic voltammogram of 0.2 mM K77 solution using different scan rates. 0.05 V/s

(black line), 0.2 V/s (red line) and 0.5 V/s (blue line).

65

Figure 64: Cyclic voltammogram of MK2 dye coated electrode using 0.05 V/s (black), 0.2 V/s

(red) and 0.5 V/s (blue). The electrode area was 0.45 cm

2

. The scan range was 0 V to 1.5 V.

Figure 65: Cyclic voltammogram of 0.2 mM MK2 solution using different scan rates. 0.05 V/s

(black line), 0.2 V/s (red line) and 0.5 V/s (blue line).

66

Figure 66: Cyclic voltammogram of L0 using a 0.5 cm

2

electrode. 0.05 V/s (black), 0.2 V/s

(red) and 0.5 V/s (blue). The scan range was from 0 V to 1.5 V.

Figure 67: Cyclic voltammogram of 0.2 mM L0 solution using different scan rates. 0.05 V/s

(black line), 0.2 V/s (red line) and 0.5 V/s (blue line).

67

Figure 68: Cyclic voltammogram of D21 dye coated electrode using 0.05 V/s (black) and 0.5

V/s (blue. The electrode area was 0.50 cm

2

. The scan range was from -0.8 V to 1.5 V.

Figure 69: Cyclic voltammogram of 0.2 mM D21 solution using different scan rates. 0.05 V/s

(black line), 0.2 V/s (red line) and 0.5 V/s (blue line)

68

Figure 70: Cyclic voltammogram of Z-907 using a 0.45 cm

2

electrode. The scan rates were

0.05 V/s (black), 0.2 V/s (red) and 0.5 V/s (blue) and the scan range was -0.8 V to 1.5 V.

Figure 71: Cyclic voltammogram of 0.2 mM Z-907 solution using different scan rates. 0.05

V/s (black line), 0.2 V/s (red line) and 0.5 V/s (blue line).

69

Figure 72: Cyclic voltammogram of N3 using a 0.33 cm

2

electrode. The scan rate was 0.05

V/s (black), 0.2 V/s (red) and 0.5 V/s (blue).

Figure 73: Cyclic voltammogram of 0.2 mM N3 solution using different scan rates. 0.05 V/s

(black line), 0.2 V/s (red line) and 0.5 V/s (blue line).

70

Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Related manuals

Download PDF

advertisement