ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2016

Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 1408

Organic Hole Transport Materialsfor Solid-State Dye-Sensitized andPerovskite Solar Cells

JINBAO ZHANG

ISSN 1651-6214ISBN 978-91-554-9659-3urn:nbn:se:uu:diva-300802

Dissertation presented at Uppsala University to be publicly examined in Häggsalen,Ångströmlaboratoriet, Lägerhyddsvägen 1, Uppsala, Friday, 7 October 2016 at 10:15 forthe degree of Doctor of Philosophy. The examination will be conducted in English. Facultyexaminer: Professor Ellen Moons (Karlstads Universitet).

AbstractZhang, J. 2016. Organic Hole Transport Materials for Solid-State Dye-Sensitized andPerovskite Solar Cells. Digital Comprehensive Summaries of Uppsala Dissertations fromthe Faculty of Science and Technology 1408. 83 pp. Uppsala: Acta Universitatis Upsaliensis.ISBN 978-91-554-9659-3.

Solid-state dye-sensitized solar cells (ssDSSCs) and recently developed perovskite solar cells(PSCs) have attracted a great attention in the scientific field of photovoltaics due to their lowcost, absence of solvent, simple fabrication and promising power conversion efficiency (PCE).In these types of solar cell, the dye molecule or the perovskite can harvest the light on the basisof electron excitation. Afterwards, the electron and hole are collected at the charge transportmaterials.

Photoelectrochemical polymerization (PEP) is employed in this thesis to synthesizeconducting polymer hole transport materials (HTMs) for ssDSSCs. We have for the first timedeveloped aqueous PEP in comparison with the conventional organic PEP with acetonitrileas solvent. This water-based PEP could potentially provide a low-cost, environmental-friendlymethod for efficient deposition of polymer HTM for ssDSSCs. In addition, new and simpleprecursors have been tested with PEP method. The effects of dye molecules on the PEP were alsosystematically studied, and we found that (a) the bulky structure of dye is of key importance forblocking the interfacial charge recombination; and (b) the matching of the energy levels betweenthe dye and the precursor plays a key role in determining the kinetics of the PEP process.

In PSCs, the HTM layer is crucial for efficient charge collection and its long term stability.We have studied different series of new molecular HTMs in order to understand fundamentallythe influence of alkyl chains, molecular energy levels, and molecular geometry of the HTM onthe photovoltaic performance. We have identified several important factors of the HTMs forefficient PSCs, including high uniformity of the HTM capping layer, perovskite-HTM energylevel matching, good HTM solubility, and high conductivity. These factors affect interfacialhole injection, hole transport, and charge recombination in PSCs. By systematical optimization,a promising PCE of 19.8% has been achieved by employing a new HTM H11. We believe thatthis work could provide important guidance for the future development of new and efficientHTMs for PSCs.

Keywords: photoelectrochemical polymerization, PEDOT, dye, hole transport material, smallmolecule, perovskite solar cell

Jinbao Zhang, Department of Chemistry - Ångström, Physical Chemistry, Box 523, UppsalaUniversity, SE-75120 Uppsala, Sweden.

© Jinbao Zhang 2016

ISSN 1651-6214ISBN 978-91-554-9659-3urn:nbn:se:uu:diva-300802 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-300802)

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals. Reprints were made with permission from the respective publishers.

I Poly (3, 4-ethylenedioxythiophene) Hole Transporting Ma-terial Generated by Photo-electrochemical Polymerization in Aqueous and Organic Medium for All Solid State Dye Sensitized Solar Cells Jinbao Zhang, Lei Yang, Yang Shen, Byung-Wook Park, Yan Hao, Erik M. J. Johansson, Gerrit Boschloo, Lars Kloo, Erik Gabrielsson, Licheng Sun, Adel Jarboui, Christian Perruchot, Mohamed Jouini, Nick Vlachopoulos, Anders Hagfeldt Journal of Physical Chemistry C, 2014, 118, 16591-16601

II Photoelectrochemical Polymerization of EDOT for Solid State Dye Sensitized Solar Cells: Role of Dye and Solvent Jinbao Zhang, Adel Jarboui, Nick Vlachopoulos, Mohamed Jouini, Gerrit Boschloo, Anders Hagfeldt Electrochimica Acta, 2015, 179, 220-227

III Solid-State Dye-Sensitized Solar Cells Based on Poly (3, 4-ethylenedioxypyrrole) and Metal-Free Organic Dyes Jinbao Zhang, Leif Häggman, Mohamed Jouini, Adel Jarboui, Gerrit Boschloo, Nick Vlachopoulos, Anders Hagfeldt Chemphyschem, 2014, 15, 1043-1047

IV Efficient solid-state dye sensitized solar cells: The influence of dye molecular structures for the in-situ photoelectro-chemically polymerized PEDOT as hole transporting mate-rial Jinbao Zhang, Nick Vlachopoulos, Mohamed Jouini, Malin B. Johansson, Xiaoliang Zhang, Mohammad Khaja Nazeeruddin, Gerrit Boschloo, Erik M. J. Johansson, Anders Hagfeldt Nano Energy, 2016, 19, 455-470

V Constructive Effects of Alkyl Chains: A Strategy to Design Simple and Non-Spiro Hole Transporting Materials for High-Efficiency Mixed-Ion Perovskite Solar Cells Jinbao Zhang, Bo Xu, Malin B. Johansson, Mahboubeh Hadad-ian, Juan Pablo Correa Baena, Peng Liu, Yong Hua, Nick Vla-chopoulos, Erik M. J. Johansson, Gerrit Boschloo, Licheng Sun, Anders Hagfeldt Advance Energy Materials, 2016, 1502536

VI Strategy to Boost the Efficiency of Mixed-Ion Perovskite So-lar Cells: Changing Geometry of the Hole Transporting Material Jinbao Zhang, Bo Xu, Malin B. Johansson, Nick Vlachopoulos, Gerrit Boschloo, Licheng Sun, Erik M. J. Johansson, and An-ders Hagfeldt ACS Nano, 2016, 10, 6816–6825

VII The Role of Three Dimensional Molecular Structural Con-trol in New Hole Transport Materials Outperforming Spiro-OMeTAD in Perovskite Solar Cells Jinbao Zhang, Yong Hua, Bo Xu, Li Yang, Peng Liu, Malin B. Johansson, Nick Vlachopoulos, Lars Kloo, Gerrit Boschloo, Erik M. J. Johansson, Licheng Sun, and Anders Hagfeldt Advanced Energy Materials, 2016, 1601062

Comments on my own contribution I was the main responsible person for all papers from I to VII. For those work I carried out most of the experiments, result analyses and writing of the manuscripts. For papers V, VI and VII, the new organic hole transport mate-rials were designed and synthesized by Dr. Bo Xu and Dr. Yong Hua. All the DFT calculations of the molecules were performed by Peng Liu. The SEM imaging was performed by Dr. Leif Häggman and Dr. Malin B. Johansson. The dye molecules were synthesized and provided by Dyenamo and EPFL collaborators.

I am a co-author of the following papers which are not included in this the-sis.

1. The effect of mesoporous TiO2 pore size on the performance

of solid-state dye sensitized solar cells based on photo-electrochemically polymerized Poly (3, 4-ethylenedioxythio-phene) hole conductor Jinbao Zhang, Meysam Pazoki, Justus Simiyu, Malin B. Johans-son, Ocean Cheung, Leif Häggman, Erik M. J. Johansson, Nick Vlachopoulos, Anders Hagfeldt and Gerrit Boschloo Electrochimica Acta, 2016, 210, 23-31

2. Facile Synthesis of Fluorene-based Hole Transport Materials for Highly Efficient Perovskite Solar Cells and Solid-State Dye-sensitized Solar Cells Yong Hua, Jinbao Zhang, Bo Xu, Peng Liu, Ming Cheng, Lars Kloo, Erik M. J. Johansson, Kári Sveinbjörnsson, Kerttu Aitola, Gerrit Boschloo, Licheng Sun Nano Energy, 2016, 26, 108-113

3. Matrix-Assisted Laser Desorption/Ionization Mass Spectro-metric Analysis of Poly (3, 4-ethylenedioxythiophene) in Sol-id-State Dye-Sensitized Solar Cells: Comparison of In Situ Photoelectrochemical Polymerization in Aqueous Micellar and Organic Media Jinbao Zhang, Hanna Ellis, Lei Yang, Erik. M. J. Johansson, Ger-rit Boschloo, Nick Vlachopoulos, Anders Hagfeldt, Jonas Bergquist, Denys Shevchenko Analytical Chemistry, 2015, 87, 3942-3948

4. Efficient Blue-Colored Solid-State Dye Sensitized Solar Cells: Enhanced Charge Collection By Using In-situ Photo-electrochemically Generated Conducting Polymer Hole Con-ductor

Jinbao Zhang, Nick Vlachopoulos, Yan Hao, Thomas W. Hol-combe, Gerrit Boschloo, Erik M. J. Johansson, Michael Grätzel, and Anders Hagfeldt Chemphyschem, 2016, DOI: 10.1002/cphc.201600064

5. Incorporation of a fluorophenylene spacer into a highly effi-cient organic dye for solid-state Dye-Sensitized Solar Cells Valentina Leandri, Jinbao Zhang, Edgar Mijangos, Gerrit Boschloo, Sascha Ott Journal of Photochemistry and Photobiology A: Chemistry, 2016, 328, 59-65

6. Carbon nanotube film replacing silver in high-efficiency sol-id-state dye solar cells employing polymer hole conductor Kerttu Aitola, Jinbao Zhang, Nick Vlachopoulos, Janne Halme, Antti Kaskela, Albert G. Nasibulin, Esko I. Kauppinen, Gerrit Boschloo, Anders Hagfeldt Journal of Solid State Electrochemistry, DOI 10.1007/s10008-015-2937-1

7. High-efficiency dye-sensitized solar cells with molecular cop-per phenanthroline as solid hole conductor Marina Freitag, Quentin Daniel, Meysam Pazoki, Kari Sveinb- jornsson, Jinbao Zhang, Licheng Sun, Anders Hagfeldt, Gerrit Boschloo Energy & Environmental Science, 2015, 8, 2634-2637

8. Dye-sensitized Solar Cells: New Approaches with Organic Solid-state Hole Conductors Nick Vlachopoulos, Jinbao Zhang, Anders Hagfeldt CHIMIA International Journal for Chemistry, 2015, 69, 41-50

9. The combination of a new organic D-π-A dye with different organic hole-transport materials for efficient solid-state dye-sensitized solar cells Peng Liu, Bo Xu, Martin Karlsson, Jinbao Zhang, Nick Vla-chopoulos, Gerrit Boschloo, Licheng Sun, Lars Kloo Journal of Material Chemistry A, 2015, 3, 4420-4427

10. Carbazole-Based Hole-Transport Materials for Efficient Sol-id-State Dye-Sensitized Solar Cells and Perovskite Solar Cells Bo Xu, Esmaeil Sheibani, Peng Liu, Jinbao Zhang, Haining Tian, Nick Vlachopoulos, Gerrit Boschloo, Lars Kloo, Anders Hag-feldt, Licheng Sun Advanced Materials, 2014, 26, 6629-6634

11. Integrated Design of Organic Hole Transport Materials for Efficient Solid-State Dye-Sensitized Solar Cells Bo Xu, Haining Tian, Lili Lin, Deping Qian, Hong Chen, Jinbao Zhang, Nick Vlachopoulos, Gerrit Boschloo, Yi Luo, Fengling Zhang, Anders Hagfeldt, Licheng Sun Advanced Energy Materials, 2014, 1401185

12. New Approach for Preparation of Efficient Solid-State Dye-Sensitized Solar Cells by Photoelectrochemical Polymeriza-tion in Aqueous Micellar Solution Lei Yang, Jinbao Zhang, Yang Shen, Byung-wook Park, Dongqin Bi, Leif Haggman, Erik. M. J. Johansson, Gerrit Boschloo, An-ders Hagfeldt, Nick Vlachopoulos, Alan Snedden, Lars Kloo, Ad-el Jarboui, Amani Chams, Christian Perruchot, Mohamed Jouini Journal of Physical Chemistry Letters, 2013, 4, 4026-4031

13. A Novel Blue Colored Organic Dye for Dye-Sensitized Solar Cells Achieving High Efficiency in Cobalt-based Electrolytes and by Co-sensitization Yan Hao, Yasemin Saygili, Jiayan Cong, Anna Eriksson, Wen-xing Yang, Jinbao Zhang, Kazuteru Nonomura, Lars Kloo, An-ders Hagfeldt, and Gerrit Boschloo (Manuscript)

14. Ambient air processed mixed-ion perovskite for high efficien-cy solar cells Kári Sveinbjörnsson, Kerttu Aitola, Jinbao Zhang, Malin B. Jo-hansson, Xiaoliang Zhang, Juan-Pablo Correa-Baena, Anders Hagfeldt, Gerrit Boschloo, and Erik M. J. Johansson (Manuscript)

15. Design, synthesis and application of π-conjugated, non-spiro molecular alternatives as hole-transport materials for highly efficient dye-sensitized solar cells and perovskite solar cells Peng Liu, Bo Xu, Yong Hua, Ming Cheng, Kerttu Aitola, Kári Sveinbjörnsson, Jinbao Zhang, Gerrit Boschloo, Licheng Sun, Lars Kloo

(Manuscript)

Contents

1. Introduction ............................................................................................... 13 1.1 Solar Energy ....................................................................................... 13 1.2 Overview of the Photovoltaics .......................................................... 14 

1.2.1 Organic Solar Cells ..................................................................... 15 1.2.2 Quantum Dot Solar Cells ............................................................ 17 1.2.3 Dye-Sensitized Solar Cells ......................................................... 18 1.2.4 Perovskite Solar Cells ................................................................. 22 

1.3 Aim of the Thesis ............................................................................... 24 

2. Solid-State Dye-Sensitized and Perovskite Solar Cells ............................ 25 2.1 Hole-Blocking Layer .......................................................................... 25 2.2 Mesoporous Layer .............................................................................. 26 2.3 Dyes .................................................................................................... 27 2.4 Perovskite ........................................................................................... 30 2.5 Hole Transport Material ..................................................................... 32 

2.5.1 Photoelectrochemical Polymerization ........................................ 32 2.5.2 Small Molecule HTMs ............................................................... 37 

3. Methods .................................................................................................... 39 3.1. Device Fabrication ............................................................................ 39 3.2. Device Characterization .................................................................... 42 

3.2.1 Current-Voltage Measurement ................................................... 42 3.2.2 Incident Photon-to-Current Conversion Efficiency .................... 44 3.2.3 Photocurrent and Photovoltage Decay Measurements ............... 46 3.2.4 Photoinduced Absorption Spectroscopy ..................................... 48 3.2.5 Hole Conductivity and Hole Mobility Measurements ................ 49 3.2.6 Photoluminesence ....................................................................... 50 

4. Results and Discussion ............................................................................. 51 4.1 Aqueous PEP versus Organic PEP (Paper I) ...................................... 51 4.2 New Precursors for PEP (Paper II and III) ......................................... 54 4.3 Effects of Dye Molecular Structures (Paper IV) ................................ 58 4.4 New Molecular HTMs for PSCs (Paper V, VI, VII) .......................... 60 

5. Conclusions and Outlook .......................................................................... 68 

Sammanfattning på svenska .......................................................................... 71 Solenergi .................................................................................................. 71 Solida färgämnessolceller och perovskitsolceller .................................... 71 Håltransportmaterial (HTM) .................................................................... 72 

Acknowledgement ........................................................................................ 75 

References ..................................................................................................... 77 

Abbreviations

A acceptor

∆A light absorption difference

light absorption coefficient

D+ oxidized dye

d film thickness

D donor

ETM electron transport material

e electron

e0 electron charge magnitude

Ecb conduction band of photoanode

Eredox-1 oxidation potential of the HTM

EF(TiO2) Fermi level of TiO2

EF(CE) Fermi level of counter electrode

EF(redox-1) Fermi level of the HTM

FA formamidinium

F Faraday constant

FTO fluorine-doped tin dioxide

FF fill factor

h hole

HTM hole transport material

HOMO highest occupied molecular orbital

HBL hole-blocking layer

ITO tin-doped indium oxide

IPCE incident photon-to-current conversion efficiency

I light intensity

∆I small modulation of light intensity

I0 baseline light intensity

J current density

Jsc short-circuit current density

)(5.1 AMphJ photon flux distribution

kinj rate constant of electron injection

kdeac rate constant of the deactivation

LUMO lowest unoccupied molecular orbital

LiTFSI lithium bistrifluoromethylsulfonyl imide

L length of Ag channel

LHE light harvesting efficiency

MA methylammonium

OPV organic solar cell

QDSC quantum dot solar cell

PSC perovskite solar cell

Pmax maximum power density

Pin incident solar energy

PEP photoelectrochemical polymerization

PCE power conversion efficiency

PEDOT poly (3, 4-ethylenedioxythiophene)

PEDOP poly (3, 4-ethylenedioxypyrrole)

PIA photoinduced absorption

PL photoluminescence

R resistance

Rcoll rate of charge collection

Rinj rate of electron injection

Rrec rate of charge recombination

SCLC space-charge limited current

Spiro-OMeTAD 2,2’,7,7’-tetrakis(N,N-dimethoxyphenylamine)-9,9′-

spirobifluorene

SHE standard hydrogen electrode

TCO transparent conductive oxide

V voltage

Voc open-circuit voltage

Voc* maximal open-circuit voltage

W distance between two Ag electrodes

η efficiency

ε extinction coefficient

σ conductivity

µ hole mobility

εr dielectric constant

ε0 vacuum permittivity

Φinj electron injection efficiency

Φcoll charge collection efficiency

λ wavelength

13

1. Introduction

1.1 Solar Energy Energy plays vital roles in the modern life to provide all the services and products that the human societies rely on. As regards the fast development of global economy and the large increase of world population, energy has un-doubtedly become one of the most challenging issues we face in the 21st century. This is because of the limited resources of the natural fossil fuels, the large energy need to meet the demands from people, the emerging cli-mate change, and the global energy security. Therefore, it is highly desirable to develop new technologies to provide sustainable energy as well as reduce the greenhouse gas emission, as required in the 2015 United Nation Climate Change Conference in Paris (COP21).

Renewable energy, which is obtained from the constantly replenished natural processes (such as sunlight, wind, rain, biomass, tides and waves), can be an excellent solution to compensate for the large world energy con-sumption. In recent years, the use of renewable energy presents an increasing trend according to Renewables Global Status Report. Figure 1 shows the detailed capacity of different types of renewable energy in recent years.

Figure 1. The power capacities of different types of renewable energy during the years from 2008 to 2014. GWe: gigawatts electric. (The data is obtained from https://en.wikipedia.org/wiki/Renewable_energy#cite_note-REN21-GSR-2012-76)

2008 2009 2010 2011 2012 2013 20140

200

400

600

800

1000

1200

1400

1600

1800

Re

ne

wa

ble

s c

ap

ac

ity

/ G

We

Year

Renewables power capaicity Hydropower Wind Solar photovoltaics

14

It can be noted that the solar energy demonstrates a fast increase in the supply of the electric power capacity from 16 GW in 2008 to 177 GW in 2014. The solar energy is a sustainable and inexhaustible source, and the large quantity of solar energy on earth (~1.7×105 TW of the electromagnetic radiation reaching the earth atmosphere) means that it can become a highly potential renewable source to provide the energy for human beings. The solar energy can be used to generate heat as well as the electricity. The ther-mal energy obtained from the solar irradiation has been widely used in the daily life, for applications such as heating, cooling, ventilation, and cooking. In addition, the solar energy can also be converted to electricity, and the corresponding devices are classified as photovoltaic cells (PVCs). In the last decade, significant increase of the global solar PV installations was achieved, although the high cost and long payback time still limit their wide-spread use in human life. In this context, the development of low-cost, high-efficiency, high-stability and environmental-friendly solar cells are demand-ed.

1.2 Overview of the Photovoltaics The photovoltaic effect was first found in 1839 by the French scientist Ed-mond Becquerel who observed the photogenerated voltage and current when connecting silver chloride in acidic solution and the platinum electrode. The photovoltaic effect is therefore known as the “Becquerel effect”. The device for the conversion of the light to electricity is called PVC. In a PVC there are different components which are very important for its efficient operation. For example, the light harvesting material is the key in the PVC which can absorb the photons; two electrical contacts adjacent with the light harvesting materials are necessary to extract the photogenerated free charge carriers (electrons and holes) to the external circuit. According to the different device structures and the operational principles, the PVCs can be divided into dif-ferent categories or generations. Specifically, silicon solar cells represent the 1st generation of PVCs, which includes single crystalline silicon solar cells (SCSSCs) and polycrystalline silicon solar cells (PCSSCs). The semiconduc-tor silicon has an indirect bandgap and, as a result, a relatively low light ab-sorption coefficient.1 Therefore, a thick layer of silicon is usually required in silicon-based solar cells. This is one of the reasons for its high production cost. In addition, the requirement of high-purity silicon materials and the complicated fabrication procedures in SCSSC technology makes it very ex-pensive and difficult to produce although it exhibits high power conversion efficiency (PCE) of ~25% (see Figure 2). In contrast, the PCSSCs demon-strate simpler and cheaper fabrication processes. However, their PCE achieved up-to-date is somewhat lower, at ~22%.2, 3 Silicon-based solar cells still dominate the PVCs in the commercial market (~90%). A key reason for

15

this is that silicon has been extensively used in the last fifty years for the fabrication of all types of electronic devices so that silicon-based PVCs are, in a sense, byproducts of the electronic industry. The 2nd generation PVCs, named thin-film solar cells, containing alternative to silicon active semicon-ductor materials, such as CdTe4 or copper indium gallium diselenide (CIGS),5 have been developing very fast. The low cost in production of thin-film PVCs is mainly attributed to a small quantity of active materials needed in the devices as well as the simple fabrication. The employed active materi-als have a direct bandgap and, therefore, a high light absorption coefficient. Besides, the low thickness of the materials in thin film solar cells makes it suitable for roll-to-roll fabrication of flexible products.6 However, the toxici-ty and rarity of several elements, such as In, Te, and Cd, incorporated into a number of efficient thin-film PVCs limit their potential for large scale appli-cation. To date, thin-film solar cells occupy ~8% of the PV market. In pursue of low cost and high sustainability, the newer generation of PVs have been extensively developed during the recent years. According to the working mechanism and the characteristics of the active materials, they are divided into organic solar cells, quantum dot solar cells, dye sensitized solar cells, and perovskite solar cells, which will be subsequently discussed in the pre-sent thesis work.

Figure 2. Efficiency charts for different types of photovoltaics available from NREL accessed on 2016-05-20, (http://www.nrel.gov/ncpv/images/efficiency_chart.jpg).

1.2.1 Organic Solar Cells In organic solar cells (OPVs), solution-processed small organic molecules or conducting polymers are used as light-harvesting materials. Based on the different architectures of the active layer, OPVs can be divided into single layer-, bilayer- and blend-based devices.7 For the single layer or bilayer

0

5

10

15

20

25

30

QDSCsOPVsPSCsDSSCsCIGSSCSSCs

11.3%11.5%

22.1%22.3%

11.9%

PC

E /

%

25.0%

Best Research-cell Certified Efficiency

16

based OPVs, the main issue is that the interfacial contact area between the active materials is insufficient to achieve fast charge separation of the photo-generated electron-hole pairs, or excitons, in order to collect them as effi-ciently as possible. In this context, in the organic solar cell variant with blend structure, a solution containing two molecules is deposited on the transparent conductive oxide (TCO) substrates, such as fluorine-doped tin oxide (FTO) or tin-doped indium oxide (ITO), in order to form a very thin solid-state film with blend materials. After that, a thin layer of a contact ma-terial, such as aluminum (Al) is deposited on top of the blend. In the blend with domain size in the order of nanometers, there exist a large number of random interfaces between two active materials. This allows the exciton to diffuse to the interface and dissociate into free charge carriers. The most representative blend materials consist of an electronically conducting poly-mer, poly (3-hexylthiophene) (P3HT) and a fullerene derivative, namely, [6, 6]-phenyl-C61-butyric acid methyl ester (PCBM).8 In this system, P3HT functions as both the light absorbing as well as the hole transport material (HTM), while PCBM serves as the electron transport material (ETM). The difference in the work function of the components in the devices gives the driving force for charge separation and collection. For example, after inci-dent photons absorption by P3HT the electrons in P3HT are excited from the highest occupied molecular orbital (HOMO) level to its lowest unoccupied molecular orbital (LUMO) level. Due to the fact that the LUMO energy level difference between the PCBM and P3HT is larger than the coulombic attrac-tion in the exciton (exciton binding energy), the photogenerated carriers are favoured to dissociate at the interfaces. Driven by the electric field formed between the two external contacts, for example the FTO and the Al layers, the electrons will migrate through PCBM towards the FTO side, while the holes travel through P3HT towards the Al side. However, due to the inter-face between the molecules in the blend, the free charge carriers, electrons and holes, can easily recombine before reaching the electrical contacts. The short lifetime of the electrons results into a short diffusion length of the free charge carriers, which is usually reported as <50 nm, much shorter than the light absorption depth (100-200 nm) needed for the optical absorption by the blend.9 The short diffusion length of the free charge carriers can in turn af-fect the active thickness of the blend material, and thus the light-harvesting ability of the thin film. In this respect, a great deal of research has been de-voted to the optimization of the blend structure,10 development of prepara-tion methods of the blend,11 design of molecular structures with appropriate-ly tuned energies, and architecture of the devices.12 To date, the highest re-ported PCE of organic solar cells is ~11% (see Figure 2).13 The simple fabri-cation process, low cost of active materials, molecular versatility, and flexibility of the organic molecular thin film render organic solar cells very attractive for large-scale flexible and plastic solar cell applications. Howev-

17

er, more work is needed for the improvement of both the efficiency and the stability of OPVs in order to fulfil the commercialization requirements.

1.2.2 Quantum Dot Solar Cells The working principle of the semiconductor quantum dot solar cells (QDSCs) can be dependent on the device architecture, for example, quan-tum-dot sensitized mesoporous oxides or thin-film quantum dots acting as the photoactive layer. In the former case, the QD sensitizer can absorb pho-tons of lower energy than the band gap of the supporting semiconductor, followed by the interfacial charge separation. Finally, photogenerated elec-trons and holes are collected at the external circuit. In contrast, in the thin-film QDSC, the charge separation and transport are performed in the bulk of the QD thin film. The size dependence of the physical and optical properties of the QD materials brings a lot of scientific interest on their potential for PVC applications. For example, the bandgap of the semiconductor QD can be tuned in a large range by controlling the size of the quantum dots so as to achieve a very broad light absorption spectrum. In addition, the often ob-served multiple exciton generation induced from one single photon holds high promise to enhance the PVC photocurrent. The maximum theoretical PCE of QDSCs can reach 44% on account of the high extinction coefficient and the possibility of multiple exciton generation.14 Besides, the simple device fabrication by solution-processed methods is favorable to large scale application. Various semiconductors have been studied in the QDSCs, such as PbS, CdS, CdSe, CuInS2, Ag2S as well as hybrid composites of such ma-terials.15 A certified high PCE of 11.6% in QDSCs based on ZnCuInSe has recently been reported by Wan and collaborators.16 One of the limitations in QDSCs is the fast electron-hole recombination, since the defect states re-maining at the surface of the QD materials provide recombination centers. Therefore, the surface passivation of the QD is crucial for the improvement of the PCE. For example, different organic compounds with amine, thiol or acid groups (thioglycolic acid or formic acid) have been investigated as pas-sivation agents in QDSCs.14 In addition, an inorganic insulating layer, e.g. consisting of ZnS and SiO2, coated on the QDs can also effectively retard the charge-carrier recombination. According to the nature of the charge-transport medium, the QDSCs can be classified as liquid QDSCs and solid-state QDSCs. Many kinds of electrolyte-dissolved redox couples have been studied as hole transport mediators, for example, I-/I3

-, Fe2+/Fe3+-based com-plexes, S2-/Sn

2-, and Co2+/Co3+-based complexes.14 The energy level of the redox couple relative to the valence band of the QDs was found to be im-portant for the hole injection and the open-circuit voltage (Voc) of the com-plete devices. For example, a CdS-based QDSCs with a redox shuttle [(CH3)4N)]2S/[(CH3)4N]2Sn exhibited a very high Voc of 1.2 V.17 Additional-ly, quasi-solid-state and all-solid-state QDSCs using a gel or solid-state elec-

18

trolyte attracted great attention due to the concern of the instability caused by the organic solvent evaporation in the conventional liquid electrolyte-based QDSCs.18, 19 However, their PCEs are relatively low (~9%) compared to that of the liquid QDSCs.20 Therefore, further optimization of the solid-state QDSCs toward high efficiency is needed.

1.2.3 Dye-Sensitized Solar Cells In order to build a well-defined charge transport network, dye-sensitized solar cells (DSSCs) were designed in the late 1980s,21 and efficiently fabri-cated by the Grätzel group in early 1990s.22 In these early studies, a high PCE of >7% was obtained by using I-/I3

- as the redox couple in the electro-lyte solution, a ruthenium coordination complex as light absorber and a mes-oporous oxide electrode, usually TiO2, as the ETM. The scheme of the DSSC is shown in Figure 3a. The promising performance of this system can be mainly attributed to the nanoporous (~20 nm) TiO2 layer (~10 μm) with a porosity of 50%-60% which provides a large surface area on which a mono-layer of dye molecules is adsorbed so as to efficiently harvest the incident light. As can be seen from Figure 3b, the DSSC exhibits complicated inter-faces between the components and also many charge transport pathways. The dye-adsorbing mesoporous oxide is deposited on a conductive sub-strate. Usually, the substrate is a transparent conducting oxide (TCO) glass, such as FTO. However, it is also possible to use a metallic conducting mate-rial, like Ti or stainless steel foil, as a mesoporous oxide support if the dye layer is irradiated from the electrolyte side of the oxide/dye/electrolyte inter-face. The dye molecules are attached on the surface of the mesoporous TiO2 particles via covalent bond formation between the dye and TiO2. The dark (i.e. not needing the light) cathode (counter electrode) is a conductive sub-strate (e.g. metal, carbon). A transparent counter electrode, usually TCO-based, is necessary in case of a photoelectrode support not transparent to light. Usually the conductive counter electrode substrate is coated with a thin layer of the catalyst, such as platinum or poly (3, 4-ethylenedioxythiophene) (PEDOT). The photoelectrode and dark counter electrode are coupled to-gether to build a sandwich-like architecture. Afterwards, an electrolyte solu-tion containing the redox shuttle, such as I-/I3

-, is injected into the space be-tween the two aforementioned electrodes. The working mechanism of the DSSCs can be described by different charge transfer (between different phases) and transport (in the same phase) processes in the devices, as depict-ed in Figure 3b.23 The essential DSSC operation processes usually are: (a) photoexcitation of the dye molecules, with an absorbed photon leading to an excitation of an electron from the HOMO level of the dye to its LUMO level, thereby, leaving an electron vacancy (hole) behind; (b) electron injection from the LUMO level of the dye to the conduction band of TiO2;

19

(c) electron transport through the mesoporous TiO2 layer; (d) hole injection from the HOMO level of the photooxidized dye to the HOMO level of the redox species, such as I-; this process is also called dye regeneration process; (e) hole transport through the redox species in the electrolytic solution; (f) electrochemical reduction of the oxidized redox species (I3

- in this case) at the counter electrode; (g) recombination between the electrons in TiO2 and the oxidized dye (D+); (h) deexcitation of the photoexcited electrons from the LUMO level to the HOMO level of the dye, in competition to electron injection into the semi-conductor conduction band, with conversion of the excitation energy to heat or luminescent light; (i) recombination between the electrons injected in TiO2 and the oxidized form of the redox shuttle, such as I3

-; (j) recombination between electrons from the TCO support of the oxides and the oxidized dye (D+); (k) recombination between electrons from the TCO and the oxidized form of the redox shuttle.

Figure 3. (a) Scheme of DSSCs with I-/I3

- as the redox couple. (b) Charge transport pathways and the relative energy levels of different components in DSSCs.

EF

TiO2 Dye I-/I3-

hν

HOMO

HOMO

LUMO

Voc

Pt

FTO

FTO

PlatinumTiO2

I3-/I-

DyePt

(a)

energy level / eV vs. vacuum-4.0

-4.5

-5.0

-5.5

(b)

a

bc

d

e fg

h

i

j

k

20

Processes (a)-(f) are favorable for the functional operation of DSSCs while processes (g)-(k) are unfavorable ones.

With respect to (j) and (k) it should be mentioned that the charge-transport medium can penetrate into the porous oxide layer and come into contact with the TCO conducting glass support. In the case of I3

- as the oxi-dized form of the redox mediator reaction the process (k) is rather slow at the TCO conducting glass/electrolyte interface. This is a complex electro-chemical reaction requiring an electrocatalyst in order to proceed efficiently. For this reason, if a TCO counter electrode is used, it should be coated by an electrocatalyst, such as platinum or conducting polymer PEDOT. However, for other redox mediators, especially several one-electron redox couples, like coordination complexes of Co, reaction (k) will be pretty fast even at the TCO glass.

The operation of an efficient solar cell with this redox mediator will be possible by interposing between the TCO layer and the porous oxide layer a thin-compact-blocking semiconductor layer, or underlayer, of thickness in the order 10-100 nm, depending of the preparation method. For the usual TiO2 mesoporous electrodes a TiO2 underlayer is appropriate. The thickness of the underlayer should be sufficient to prevent the dark current according to (k) but, contrarily, not prevent the transfer of the photogenerated electrons from the porous dye-coated oxide into the TCO contact on the way to the external circuit. Various methods are possible: spray pyrolysis of a titanium coordination compound dissolved in an organic solvent, spin coating using the same type of solution, dip-coating in an aqueous TiCl4 solution, atomic layer deposition, or sputtering.

The effects of the recombination reaction (k) in the DSSC performance are particularly severe for device with organic solid-state HTM; in this case the requirements for a good-quality underlayer are particularly stringent.

The conversion efficiency (η) of the DSSCs depends on the different pho-tovoltaic parameters including the Voc, short-circuit current density (Jsc) and fill factor (FF). The Voc of the DSSC is determined by the Fermi-level of the TiO2 and the Fermi level (or electrochemical potential) of the redox shuttle in the electrolyte solution. The Jsc is mainly affected by the light-harvesting capacity of the dye molecules in conjunction with efficient electron injection and dye regeneration. The FF may provide information on the series and shunt resistances of the devices as well as on the electrocatalytic properties of the counter electrode toward the oxidized redox mediator. By comparing, on the one hand, the part of the incident photon energy harvested by the dye, corresponding to the HOMO-LUMO gap, and, on the other hand, the cell voltage output, there is a clear voltage loss in the DSSC. The observed volt-age loss in the DSSC is largely due to the driving forces which are needed for the charge-transfer processes at the interfaces, such as (b) and (d). Usual-ly, it is assumed that the process (b) is kinetically very efficient as compared to (g), although the kinetics (fs~ps) of the electron injection process is very

21

dependent on the properties of the TiO2 as well as the properties of the dye layer, e.g. dye aggregation.24 For DSSCs with I-/I3

- as redox mediator, which was the only efficient redox mediator until 2010, the main voltage loss in the device is attributed to the dye regeneration step occurring at the interface between the dye and the redox shuttle in the electrolyte. From the kinetics point of view, the dye regeneration reaction proceeds at the ns to μs time-scale, which depends on the electrolyte composition, the types of cations adsorbed on the TiO2 surface, and the dye properties; the dye regeneration needs to effectively compete with the recombination process (g) between TiO2 electrons and the oxidized dye. As concluded from many systematic studies, a large voltage of 500-600 mV is required for the Ru-based dye re-generation by the I-/I3

-.24 This is due to the fact that the I-/I3- reaction is a

complex one, involving several electrochemical and chemical steps. There-fore, the 500-600 mV energy loss includes the free energy losses of chemical reactions associated with the complex overall I-/I3

- reaction. As a result, DSSCs based on I-/I3

- usually show a relatively low Voc. In order to reduce this large voltage loss, different redox couples have been explored and inves-tigated. For example, Co2+/Co3+-based coordination complexes have success-fully shown efficient dye regeneration with a low voltage loss in the device and often exhibited high Voc.

25, 26 Other alternatives, such as Fe2+/Fe3+, Cu+/Cu2+-based organometallic or coordination complexes, heterocyclic species, like TEMPO and its derivatives, and organic disulfide/thiolate, also present interesting properties.27, 28 In addition, in order to further enhance the working efficiency of DSSCs, a large number of research have been done in designing novel dye molecules. The design principle of dye molecules is varied according to the specific purpose, e.g. for preventing the interfacial charge recombination, for increasing the light absorption coefficient, for cost-effective synthesis, for specified colour. The detailed dye molecular engineering can be found in the respective reviews.29, 30 To date, the highest reported PCEs of DSSCs include 14.3% by using co-sensitization with two dye molecules adsorbed,31 13% by using a single transition metal coordina-tion complex SM315 as sensitizer,32 and 12% by employing a single organic dye.33, 34

In spite of the great achievements in liquid DSSCs, the concern about their long-term stability arises due to the use of a volatile solvent, e.g. ace-tonitrile, the dye desorption in the electrolytic solution,35 the degradation of the cobalt-based complexes,36 and the electrode corrosion and the visible light absorption by I-/I3

-.27 In this respect, solid-state DSSCs (ssDSSCs) (Figure 4a) were developed in 1998 by the Grätzel group, by using a mo-lecular HTM 2,2’,7,7’-tetrakis(N,N-dimethoxyphenylamine)-9,9′-Spirobifluorene (Spiro-OMeTAD),37 (see Figure 4b) as solid-state HTM. In this report, the PCE of ssDSSCs was quite low compared to that of the con-ventional liquid-based DSSCs. In order to improve the photovoltaic perfor-mance of ssDSSCs, substantial efforts have been devoted into the following

22

directions: (a) fundamental understanding of the operating limitations in ssDSSCs in comparison with the liquid-based counterparts; (b) developing new dye molecules and HTMs; (c) interfacial engineering for efficient charge collection.38 The detailed development of ssDSSCs will be discussed later.

Figure 4. (a) Scheme of the ssDSSCs devices. (b) Chemical structures of HTMs Spiro-OMeTAD and PEDOT.

1.2.4 Perovskite Solar Cells Semiconductor perovskites have received a great attention in the scientific community of the PVCs after the first report of a perovskite sensitized solar cells, of the quantum-dot type, based on CH3NH3PbI3 in 2009 by Miyasaka and co-workers.39 In that work the fabricated perovskite solar cells (PSCs) were following the device architecture of DSSCs with I-/I3

--based liquid electrolyte and semiconductor CH3NH3PbI3 as a sensitizer. The as-obtained PSCs demonstrated a PCE of 3.6% but exhibited a very short lifetime due to the instability of the perovskite CH3NH3PbI3 which can dissolve in acetoni-trile. Inspired by this pioneer work, in 2012 the Grätzel/Park and Snaith/Miyasaka groups, in Lausanne and Oxford, respectively, successfully applied CH3NH3PbI3 as light absorber in solid-state perovskite sensitized solar cells with Spiro-OMeTAD as HTM so as to avoid the dissociation of CH3NH3PbI3.40, 41 A promising PCE of >10% was achieved. After that en-couraging start, a large number of publications about PSCs have been pro-duced during the past few years.42, 43In general, the organic-inorganic hybrid perovskite material has a chemical formula of ABX3 (A: cations, B: Pb, and C: anions), as shown in Figure 5a. The perovskite materials exhibit unique properties, such as direct band gap excitation leading to high light absorption

23

coefficient (see Figure 5b), long charge carrier diffusion length and high charge transport conductivity.44, 45 These properties are favorable for a light harvesting material in solar cells.

Figure 5. (a) Crystal structure of the perovskite CH3NH3PbI3. (b) The absorption and emission properties of (FAPbI3)0.85(MAPbBr3)0.15.

It can be seen from the efficiency chart (see Figure 2) that the currently certified PCE is 22.1% by KRICT/UNIST. This significant improvement in the PCE of PSCs can be mainly attributed to the modification of the perov-skite formation methods, the device architecture optimization, the interface engineering, the wide compositional engineering of the perovskite, and the deeper understanding of the physical properties of the devices. For example: (a) different methods were developed for the perovskite formation, such as two-step and three-step sequential deposition,46, 47 spray-assisted formation,48 vapor deposition,49 hot casting,50 and anti-solvent method;51 (b) different device structures, e.g. standard PSCs,52 inverted PSCs,53 or pla-nar PSCs,54 have been investigated; (c) great efforts were made in designing efficient charge selective layers, including the ETMs and HTMs; the details can be found in the reviews;55, 56 (d) the wide combination of cations and anions in the perovskite composition promote different perovskite materials for the solar cells, such as, CH3NH3PbCl3,57 CH3NH3PbBr3,58 CsPbI3,59 CsPbBr3,60 CH(NH2)2PbI3,61 (CH(NH2)2PbI3)x(CH3NH3PbBr3)1-x.62

In addition, due to the toxicity of Pb, several new lead-free perovskites are being explored, for example CH3NH3SnI3, Cs3Bi2I9, but they presented low PCEs;63

The fast development of PSCs has also drawn the attention to possibility for the large scale application. However, the known poor stability of the perovskite materials due to their high sensitivity to moisture is a big chal-lenge for the perspective of long term use of devices based on them.64, 65 Different encapsulation layers, such as polymer poly-(methyl-methacrylate), graphite and carbon nanotubes, have been successfully used in PSCs to im-prove the stability.66, 67 Recently, Han and co-workers developed inorganic charge-selective layers for PSCs with the result that stable devices are ob-

400 500 600 700 800 9000.0

2.0x106

4.0x106

6.0x106

8.0x106

1.0x107

1.2x107

Wavelength / nm

Extin

ctio

n co

effic

ient

/ m

-1

0

1x106

2x106

3x106

4x106

5x106

PL

inte

nsity

(a. u

.)

(FAPbI3)0.85(MAPbBr3)0.15

(a) (b)

24

tained with only ~10% loss in PCE during 1000-hour light soaking.68 In ad-dition, the state-of-the-art HTM Spiro-OMeTAD HTM is very expensive to synthesize so that alternative inexpensive HTMs are desired for the large scale application of PSCs.

1.3 Aim of the Thesis The general target of the thesis is to develop solid-state dye sensitized and perovskite solar cells with efficient organic HTMs in order to obtain a fun-damental understanding of the requirements of the HTMs and to further im-prove the photovoltaic performance of the solar cell devices. Two types of HTMs, conducting polymers and small molecules, are studied in this thesis. Firstly, the poor pore filling (pore infiltration) of the mesoporous TiO2 by spin-coated solution-processed small molecular or polymeric HTMs in ssDSSCs significantly limits the thickness of the dye/TiO2 layer and reduces the dye regeneration efficiency.69 In this respect, we have extensively studied an alternative method, called in-situ photoelectrochemical polymerization (PEP), to in-situ synthesize an electronically conducting polymer, serving as HTM of the ssDSSC, into the TiO2 pores.70 By employing PEP, it is feasible to use mesoporous TiO2 layer with a high thickness of up to 6 µm, compared to an optimal thickness of 2 µm for Spiro-OMeTAD-based ssDSSCs.71 The PEP method was first reported by the Yanagida group in Japan,70 and the electrolytic solution for the PEP was always based on the organic solvent acetonitrile because it can easily solubilize the organic precursors used dur-ing PEP. The first aim of the thesis is to further develop a variant of the PEP approach by instead using water as solvent because water is low-cost and environmentally friendly. In order to achieve this goal it was necessary to further understand the PEP process and the effect of different dye molecules, different precursors, and to relate this to the final solar cell efficiencies.

One of the most important and also expensive parts in efficient perovskite solar cells is the hole transport material (HTM). The second main aim of the thesis is therefore to develop new alternative molecular HTMs to replace the high-cost Spiro-OMeTAD for efficient solution-processed PSCs. Earlier work has shown the importance of the HTM layer, and in this thesis we try to systematically study the effect of the structure of the HTM on the different photovoltaic properties of the solar cell and the overall energy conversion efficiency.

25

2. Solid-State Dye-Sensitized and Perovskite Solar Cells

In this thesis, the focus is on the solid-state dye sensitized and perovskite solar cells. This chapter is devoted to a brief description of the different components in ssDSSCs and PSCs, including the hole-blocking layer, the photoanode (which is also electron-transport material, ETM), the light ab-sorber (dye or perovskite), and the hole-transport material (HTM).

2.1 Hole-Blocking Layer As can be seen from Figure 4a, a hole-blocking layer (HBL) or underlayer is needed between the TCO substrate and the mesoporous oxide layer in ssDSSCs. The HBL is usually a compact high-bandgap semiconductor. Up to the present point TiO2 has been the mostly used HBL for ssDSSCs and PSCs; this HBL is also used in this work. The purpose of the HBL is to pre-vent the photoinjected electrons in the conductive substrate from directly recombining with the holes in the HTM. The ideal HBL should cover the rough surface of the TCO substrate, such as FTO, completely, be free of pinholes and cracks, and have a thickness as low as possible in order to re-duce the charge-transport resistance. There are different methods to prepare the TiO2 blocking layer, including atomic layer deposition, electrodeposi-tion, spray pyrolysis, sol-gel transformation, dip coating or spin coating of the titanium precursor solution.72-77 In addition, there are some alternative blocking layers, e.g. copolymer, hybrid polymer/TiO2,

78, 79 and other insulat-ing oxides.80 However, the most common method is the spray pyrolysis of a titanium precursor solution (0.2 M titanium isopropoxide and 2M acety-lacetone in isopropanol) on top of the conductive glass under a high tem-perature of ~500oC in order to facilitate the fast TiO2 formation. This spray pyrolysis method for preparing the HBL is used in this thesis. The optimal thickness of the TiO2 compact layer is 100 nm-200 nm for ssDSSCs, de-pending on the nature of the HTMs.81

In spite of the fact that a rather high PCE of >14% has been achieved for HBL-free PSCs,82 the most efficient PSCs do always need the HBL, for ex-ample of an optimal thickness of ~135 nm for CH3NH3PbI3-xClx-based PSCs.83 It is possible to study its blocking effect by electrochemical meth-

26

ods, usually by measuring the current flowing through the HBL and corre-sponding to the electron reaction of a solution-based redox-active species, for example K4Fe(CN)6, as presented in the published literature.84, 85

2.2 Mesoporous Layer The mesoporous TiO2 substrate (Figure 4a) has different functionalities in DSSCs: (a) provides a large surface area for the loading of a monolayer of dye molecules in ssDSSCs; (b) builds the electron transport network through the interconnected TiO2 particles; (c) possesses a high bandgap (3.2 eV for rutile TiO2) for the transmission of the visible light; (d) facilitates the ions diffusion in the electrolyte permeating its mesoporous structure; (e) accepts the photoinduced electrons from the sensitizer so as to reduce the possibility of the excited state deactivation competing with dye-to-TiO2 electron injec-tion. Due to all the aforementioned properties, TiO2 has been established as the most popularly used photoanode. The electron transport in the pho-toanode is one of the important factors which significantly affect the charge collection in the solar cells devices. Therefore, detailed studies towards ob-taining a fundamental understanding of the electron-transport mechanism in mesoporous TiO2 had been undertaken. In conclusion, there are two widely accepted approaches: the diffusion model,86 and the multiple trapping mod-el.87 It has also been found that the attached dye molecules and the surface treatment by TiCl4 play important roles in the electron transport in TiO2 ma-trix.88 It is known that in ssDSSCs the injected electrons located in the TiO2 particles can be easily recombined with the HTMs before moving to the ex-ternal circuit because of the direct contact between the TiO2 and the solid HTMs. Therefore, a lot of work has been done to modify the TiO2 morphol-ogy or the dimensionality, as well as to develop alternative photoanode ma-terials, in order to efficiently transport the charges and retard the interfacial electron recombination.89 For example, different TiO2 morphologies, such as nanowires,90 nanorods,91 nanotubes,92 nanofibers,93 nanocrystals by laser pyrolysis,94 hollow spheres,95 microbeads,96 honeycomb-like layers,97 and hierarchical fibers,98 have been synthesized and used for ssDSSCs. The de-tailed critical parameters in TiO2 electrodes have been discussed in a cited article.99 In addition, ZnO has been another widely studied photoanode mate-rial due to its good electron-transport properties, with remarkable electron mobility in the bulk, which is 2-3 orders of magnitude higher than that for TiO2.

100 Different ZnO nanostructures, such as nanowires,101 nanorods,102 and nanocrystals,103 have been tested in ssDSSCs. However, up to the pre-sent, ZnO-based ssDSSCs have exhibited relatively low Voc and FF com-pared to TiO2-based devices due to the faster electron recombination.104 In this respect, it should be noted that the surface functionalization of the pho-

27

toanode by molecules or additives could be an effective approach to decrease the interfacial recombination.

2.3 Dyes In ssDSSCs, dye molecules which are adsorbed on the surface of the TiO2 layer are key components for the functional operation of the device. Firstly, the dye molecule determines the light harvesting ability of the ssDSSCs and thus the photocurrent obtained from the device. Besides, the monolayer of the dye on TiO2, acting as an isolating layer between the ETM TiO2 and the HTM, plays a crucial role in blocking the electron recombination process between the electrons in TiO2 and the holes in HTMs so as to enhance the Voc of the device. Moreover, the energy levels of the dye molecules relative to that for TiO2 and the HTMs affect the kinetics of the charge-transfer pro-cesses at these interfaces, including the electron injection and the dye regen-eration (see Figure 3b). Therefore, a minor change in the molecular structure of the dye could contribute to a significant change in the photovoltaic per-formance of the ssDSSCs. The most widely used dyes are Ru-based coordi-nation dyes and metal-free organic dyes. In the first report of ssDSSCs, a Ru-dye was used and a low PCE was observed. From the following studies by comparing Ru-dyes and metal-free organic dyes it was established that the latter usually show higher extinction coefficients but narrower absorption spectra, as well as better interfacial recombination blocking effects com-pared to the former in ssDSSCs.105, 106 In this respect, different types of dye structures have been proposed in order to improve the performance of ssDSSCs. For example: Grätzel and co-workers developed an indoline dye D102 in combination with the HTM Spiro-OMeTAD, and a high PCE of 4.0% was achieved in 2003.107 Palomares and co-workers developed a cya-nine dye in order to compare the charge-transfer kinetics in liquid and solid-state DSSCs, and they found that the ssDSSCs based on HTM Spiro-OMeTAD demonstrated much faster dye regeneration kinetics compared to I-/I3

--based liquid DSSCs.108 Hagfeldt and co-workers proposed a perylene dye ID176 with a high absorption coefficient (25000 M-1 cm-1 at 590 nm) as well as a broad absorption spectrum. A high photocurrent (~9 mAcm-2) at-tributed to the fast dye regeneration process and a PCE of 3.2% were ob-tained.109 Nazeeruddin and co-workers developed a series of donor-π-acceptor dyes to study the effects of the alkyl chains in the donor part for ssDSSCs.110 Grätzel and collaborators designed a Ru-dye, C101, with a high extinction coefficient and a high PCE of 4.5% was obtained for ssDSSCs.111 Then Hagfeldt, Sun and co-workers synthesized a series of efficient tri-phenylamine-based dyes and applied them in ssDSSCs (PCE=4.5%).112 Grätzel and co-workers studied the donor effects in the organic dyes and found that the donor structures play important roles in the dye surface cover-

28

age as well as in the shift of the conduction band of TiO2.113 Alan Sellinger

and co-workers systematically investigated the relationship between organic dye structure and recombination lifetime.114 Snaith and collaborators applied a donor-free dye in ssDSSCs and obtained a high PCE of 4.4%.115 Therefore, further understanding of the role of the different units in the dye structure towards developing new efficient dyes particularly for ssDSSCs is desired for the improvement of their performance. The dye molecules used in this thesis include Ru-based dyes together with metal-free organic dyes of the donor-π-acceptor (D-π-A) structure, such as LEG4 dye, as shown in Figure 6. In D-π-A-based dyes, the donor part acts as the electron-donating unit where the electron density of the HOMO of the dye molecules is mainly located. In contrast, the density of the LUMO of the dye molecules is located on the acceptor part. The photoinduced electron transfer from the HOMO to the LUMO is achieved by transiting through the π-bridge, as shown in Fig-ure 6. By this design, the electron density on the D-π-A dye molecules is well divided and separated in order to reduce the recombination within the dye molecules; this design is usually called “push-pull” structure.

Figure 6. Chemical structure of the organic D-π-A dye LEG4.

The donor group is an electron-rich unit and it is very important for the light absorption ability of the dye. Therefore, the extinction coefficient of the dye molecules can be tuned by the modification of the donor group. In addition, differences in the electron density on the donor part also affect the HOMO level of the dye. The triphenylamine-based donor group is one of the most used units in efficient dyes due to its excellent electron-donating ability.116 The conjugation system on the linker provides a pathway for the fast elec-tron transfer from HOMO (donor) to LUMO level (acceptor). The acceptor part usually has a good electron-withdrawing ability as well as the good af-finity to bind on the surface of TiO2. The cyanoacetic acid is one of the typi-cal representative acceptors in the organic dyes. The dyes under considera-

29

tion can present different optoelectronic properties by combining the donor, linker and acceptor in different ways, as shown in Table 1.

Table 1. The optoelectronic properties of the dye molecules.

Dye EHOMO V/SHE

Abs./max nm

ε 103 M-1cm-1

C218117 0.97 517 59D35118 1.16 445 35

LEG4116 1.08 541 49D21 L6118 0.98 458 37MK253119 0.90 534 34.7Z907120 1.01 375, 515 12.6 (515 nm) B11121 1.04 305, 388, 554 24.2 (554 nm)

Figure 7. Chemical structures of dye molecules used in the thesis.

Ru

N

NCS

N

NNCS

N

O

HO

OOH

C9H19C9H19

N

C4H

9OOC4H9 OC4H

9

C4H

9O

S

NCCOOH

N

C6 H

13 O OC 6H 13

S

S

COOHNC

N

S

S

COOHNC

C6H13

C6H13

OC6H13

N

S

S

COOHNC

C6H13

C6H13

C4H9O OC4H9OC4H9 C4H9O

C6H

13O

N

S

S

COOHNC

C6H13

C6H13

C4H9OOC4H9

OC4H9 C4H9O

Ru

N

NCS

N

NNCS

N

O

HO

OOH

S

SS

SS

S

C6H13

C6H13

D21L6 D35 C218

LEG4 MK253

Ru-Z907 Ru-B11

30

The dye molecules employed in this thesis are listed in Figure 7. They in-clude metal-free organic dyes C218, MK253, D35, LEG4, D21L6, and Ru-based coordination compound dyes Z907, and B11. It can be noted that there is a gradual evolution of the donor group from C218, D35, LEG4, to MK253. For D35, C218 and D21L6 different π-linkers are present. Ru-Z907 and Ru-B11 have different side groups attached to the bipyridine ligands. In order to illustrate the effect of the molecular structure on the dye properties, the detailed optoelectronic parameters of the dyes are listed in Table 1.

2.4 Perovskite Since the first report on the perovskite sensitized solar cells, the perovskite material has been widely used as light absorber in PSCs. In order to improve the efficiency and stability of the PSCs, a systematic perovskite composi-tional optimization has been conducted. For example, MAPbI3 (MA=CH3NH3, methylammonium), MAPbBr3, FAPbI3 (FA=HC(NH2)2, formamidinium), MASnI3 and CsPbI3 have been tested and investigated in PSCs.122 It has been established that the optoelectronic properties, i.e. bandgap, light absorption spectrum, and photochemical stability, are highly dependent on the cation-anion combination in the composition. In particular, the size (ionic radius) of cations can affect the perovskite structure (dimen-sionality), and the halide anions can tune the bandgap of the perovskite ma-terials. For instance, MAPbI3, with a bandgap of 1.5-1.6 eV and a broad light absorption spectrum till 800 nm, has been widely studied in PSCs. In con-trast, FAPbI3 exhibits a bandgap of 1.48 eV and an extended absorption spectrum up to 840 nm. Recently, Seok and co-workers investigated the composition effects of (MAPbBr3)x(FAPbI3)1-x by combining different ratios between MAPbBr3 and FAPbI3.

62 They found that (a) the PCE is improved from 13.5% for FAPbI3-based PSCs to 17.3% for the optimal combination (MAPbBr3)0.15(FAPbI3)0.85; (b) the hexagonal non-perovskite phase (poly-morph) of FAPbI3 (sintered at 100oC) is substantially changed to the trigonal perovskite phase after adding a 5% mol of MAPbBr3. Both phase stability and crystallinity is improved by the optimal combination of cations and ani-ons; (c) the surface uniformity of the formed perovskite layer is significantly improved for (MAPbBr3)0.15(FAPbI3)0.85 compared to the rough surface of FAPbI3. As regards the fabrication of the perovskite film, different formation methods have been reported, such as dual source evaporation,123 sequential (two-step) solution deposition,46 one-step spin coating,51 and vapor-assisted solution process.49 The one-step solution process is believed to be an effi-cient and reproducible method to prepare the perovskite film. Seok and co-workers studied the solvent effects in the one-step anti-solvent process and they concluded that the solvent dimethyl sulfoxide (DMSO) is critical for the formation of the PbI2-DMSO complex and retards the perovskite crystalliza-

31

tion during spin coating in order to form smooth and uniform perovskite layer.51

Figure 8 shows the preparation steps for the one-step spin coating method. Briefly, the perovskite precursor solution is spin coated on the TiO2 substrate followed by the addition of a droplet of anti-solvent, such as chlorobenzene or toluene. The obtained perovskite film is then sintered at 100oC for one hour.

Figure 8. The scheme of the one-step solution process for making the perovskite films.

In this thesis the employed perovskite composition is the mixed-ion per-ovskite (MAPbBr3)0.15(FAPbI3)0.85. This optimized perovskite material is used to investigate the properties of different HTMs. Figure 9 depicts a top view of a perovskite layer which is prepared by the one-step solution process as illustrated in Figure 8. The perovskite layer presents a uniform grain size distribution between 200-300nm. A dense perovskite film is obtained with-out large pinholes; such a quality is important for efficient charge transport and collection in the devices.

Figure 9. SEM image, top-view, of the formed mixed-ion perovskite materials by the one-step solution process.

perovskite

Perovskite solutionloading

Perovskite solutionSpin coating

Chlorobenzenedripping

Intermediate phase Perovskite film

100oC, 1hhotplate

32

2.5 Hole Transport Material The HTM is very important in the solar cell devices for efficient charge transport and extraction. A good HTM for high-performance solar cells should accept the holes from the light absorber efficiently, transport them quickly, suppress hole (HTM)-electron (TiO2) recombination as much as possible. There are two types of HTMs used in the present thesis. One is electronically conducting polymer-based HTM for ssDSSCs, and the other is small molecular HTMs for PSCs. In this session, these two kinds of HTMs will be discussed separately.

2.5.1 Photoelectrochemical Polymerization Firstly, in ssDSSCs, the mesoporous TiO2 layer provides a large surface area for adsorbing a sufficient amount of dyes to harvest the light. However, in the case of the solution-processed molecular HTMs, such as Spiro-OMeTAD, the random distribution of TiO2 pores poses a challenge as re-gards complete penetration of the HTM into the pores so as to ensure good contact with the dye. This problem, often called pore-filling limitation, can influence the dye regeneration process, i.e. the transfer of the photogenerated holes from the dye to the HTM. A slow or inefficient dye regeneration pro-cess in ssDSSCs could provide more possibilities for the deleterious elec-tron-hole recombination in the devices. In order to improve the pore filling of the HTM, the increase of the pore size has been considered.124 However, in this case the surface area of TiO2 for dye loading is reduced, resulting in a low photocurrent. Grätzel and co-workers developed different oligomer HTMs in order to study the pore-filling effects.125 In consideration of the fact that isolated air cavities exist in the TiO2 pores, a vacuum-assisted method was applied to inject the HTM solution into the pores. By using this tech-nique, the pore filling and, as a consequence, the device performance were improved compared to those for the drop-casting method.126 Boschloo, Hag-feldt and co-workers successfully applied photoinduced absorption spectros-copy and spectroelectrochemistry methods to study the dye regeneration process.157 Later, McGehee and collaborators used XPS depth profiling and UV-vis absorption to determine the optimal pore-filling volume, which amounted to ~60%-65% for 2.5 μm TiO2. They also showed that the pore filling can be improved by lowering the speed of the spin coating and in-creasing the concentration of Spiro-OMeTAD.127 Lenzmann and co-workers used photoinduced absorption spectroscopy (PIA) to study the relationship between the pore-filling fraction and the concentration of Spiro-OMeTAD.128 They concluded that the optimal thickness of TiO2 is ~3 μm and a higher concentration of HTM in the spin-coating solution is favorable for a high filling fraction. McGehee and co-workers systematically studied the effect of variations in the pore-filling fraction of the HTM on the dye

33

regeneration process and the electron (TiO2)-hole (HTM) recombination by transient absorption spectroscopy.129 An alternative combined sequential filling method consisting of soaking of dye/TiO2 electrode in an HTM solu-tion followed by spin-coating of the HTM solution on top of the electrode. A high pore filling of 84% was obtained by using this method compared to 67% for solution casting.130 Later, Snaith and co-workers concluded on the basis of optical reflectometry measurements that a pore filling of 60% is required for the optimized device.131

Another limitation is the short electron diffusion length which is affected by the fast interfacial charge recombination and the slow charge transport rate. The small molecular HTMs usually exhibit a low hole conductivity. Snaith and co-workers have determined an electron diffusion length of ~2 µm under applied bias conditions. Therefore, the poor pore filling of the HTM in mesoporous TiO2 film and the short charge diffusion length limit the thickness of TiO2 to 2-3 μm. However, the incident light cannot be fully absorbed by a 2 μm thick TiO2 layer if using the common Ru-dyes. There-fore, it is necessary to develop either new dye molecules with a high extinc-tion coefficient or to an alternative approach for the efficient penetration of the HTM into the pores of the thick TiO2 layer.

Figure 10. (left) Scheme of in-situ photoelectrochemical polymerization process with EDOT as precursor in the electrolytic solution; (right) the image of TiO2/LEG4 films before and after PEP.

In this context, the in-situ photoelectrochemical polymerization (PEP) meth-od was developed in order to generate conducting polymer HTMs into the mesoporous TiO2 by a successive polymerization of small precursors in the solution. The scheme of the PEP process is shown in Figure 10. The process can be summarized as follows: (a) the incident light excites the dye molecule which is adsorbed on the TiO2 surface, and the photo-induced electron-hole pair separate at the interfaces between dye-coated TiO2 and precursor-containing solution (such as EDOT in Figure 10); (b) the electron will be

photoelectrochemical polymerization (PEP) Before PEP

After PEP

34

injected from the LUMO level of the excited dye into the TiO2 and the hole will be injected from the HOMO of dye to the precursor in the solution so as to effect the oxidative conversion of the precursor to a radical; (c) the reac-tive radicals will couple together to form oligomers and polymers. There-fore, there are many components (TiO2, dye, precursor, solvent, electrolyte, light source) in this system which can affect the PEP process. The PEP method was for the first time presented by Yanagida in 1997.132 In that work, they used the PEP in order to prepare the polypyrrole HTM and a PCE of ~0.1% was obtained. The doping density of the polypyrrole by applying a bias potential on the dye/TiO2/polypyrrole film was discussed and the con-clusion was that a high oxidation state of the polypyrrole is favorable for the charge transport in the devices. Later, Yanagida and co-workers used PEP to synthesize PEDOT as HTM in ssDSSCs, and investigated the importance of the additives (lithium salt and tert-butyl-pyridine) for the post-treatment after PEP.133 The same group compared the polymerized HTMs polypyrrole and PEDOT in ssDSSCs and achieved a PCE of 0.5% by using PEDOT with high transparency and high conductivity.134 It is noted that they all used Ru-dyes in the PEP process.135 In a later publication, a PCE of 1.3% was report-ed by using an amphiphilic Ru-dye Z907and the polymerized PEDOT as HTMs. It was found that the dye Z907 with long alkyl chains could suppress the electron back transfer to the PEDOT.136 Lithium salts with different ani-ons (BF4

-, ClO4-, CF3SO3

-, TFSI-) were used as the electrolyte in the PEP and the conductivity of the formed PEDOT was compared.137 A high PCE of 2.8% was obtained by using LiTFSI as the doping agent in the PEP; the PCE increase is due to the high conductivity of the PEDOT with TFSI- as doping anions. Metal-free organic dye D149 was used in the PEP to synthesize PE-DOT by Liu and co-workers and a high PCE of 6.1% was achieved. The large improvement of the photovoltaic performance was mainly attributed to high light absorption of the dye, the diminished interfacial electron recombi-nation rate and the good penetration of PEDOT.138 A description of the de-tailed development of the PEP method can be found in the review papers by Vlachopoulos et al. and Liu et al.139, 140

Figure 11. Detailed charge-transfer pathways during the PEP process.

35

Figure 11 illustrates the detailed charge-transfer processes during the PEP with bis-EDOT as an example. After irradiation (~5% of sun light, intensity of 5 mWcm-2) on the dye-sensitized TiO2 film (working electrode), the elec-trons are extracted under an applied bias potential of 0.2V (vs. reference electrode) at the photoanode. In contrast, the photogenerated holes are trans-ferred to the PEP precursor in the electrolytic solution, driven by the electron energy (oxidation potential) difference between the dye and the precursor. Therefore, the energy level matching between the TiO2, dye, and the precur-sor is very important to operate the PEP efficiently.

Figure 12. The transient current curve in the PEP process by using chronoamperom-etry (applied potential is 0.2V vs. reference) method.

Figure 12 shows the typical transient current curve during the PEP process obtained by the chronoamperometry method. Since a fixed electrochemical potential was applied to the dye/TiO2 electrode, the recorded current can pro-vide clear information on the kinetics of PEP. As shown in Figure 12, a very low dark current is recorded when light is off because (a) the TiO2 electron concentration in the dark is very low; (b) the applied potential of 0.2V vs. the reference electrode (0.4V vs. SHE) is not sufficient for the oxidation reaction of the precursor in the electrolyte; for example, the oxidation potential of bis-EDOT in acetonitrile is 0.9V/SHE; (c) the compact TiO2 blocking layer is very resistive and blocking. In contrast, when the light is on, a high transient current is observed due to the sudden generation, by photoinjection from the dye, of a large amount of free electrons in TiO2. This transient current behavior demon-strates a typical deposition according to a nucleation and growth mechanism. The increase of the photocurrent between 120 s to 300 s could indicate a grad-ually faster polymerization occurring once the formation of an oligomer layer is initiated in the dye/TiO2 pores. Afterwards, the photocurrent gradually de-creases (300s to 600s) to a virtually plateau. This decrease of the photocurrent

0 200 400 600 800 10000

1x10-5

2x10-5

3x10-5

4x10-5

5x10-5

light

on

Curr

ent / A

time / s

PEP

light

off

36

could be caused by (a) the diffusion limitation of the precursors through the porous structure after the formation of solid PEDOT; (b) faster electron re-combination with the formed PEDOT which is in close contact with the dye molecules. The PEP process can be also conducted by using the method of chronopotentiometry by applying a fixed current (~20 µA/cm2) flowing in the system. An advantage of the constant current vs. constant potential PEP is that the rate of polymer formation is uniform and can be easily controlled by ad-justing the applied current density. The electrode potential vs. the reference electrode is monitored in order to follow the kinetics of the PEP as well as to ensure that it does not shift to such positive values that the polymer can be irreversibly over-oxidized, with irretrievable loss of electronic conductivity. In that case, the rate of the polymerization reaction is controlled so as to obtain a uniform distribution of PEDOT in the TiO2 pores. Typically a PEP charge of 30-40 mC/cm2 was applied to a 6 μm dye/TiO2 electrode, with a current densi-ty of 10-20 μA/cm2, in order to get a good pore filling.

The uniformity of the formed PEDOT in the TiO2 pores is very important for efficient dye regeneration vs. electron recombination as well as for fast hole transport. Figure 13 shows a top view of surface morphology of the elec-trode before and after PEP. It can be seen that after PEP a layer of PEDOT covers the dye/TiO2 surface and that the PEDOT layer has a quite porous structure.

Figure 13. The top view of the surface morphology of the TiO2/dye before PEP (a, c) and TiO2/dye/PEDOT after PEP (b, d).

(a) (b)

(d) (c)

37

2.5.2 Small Molecule HTMs Solution-processed organic molecular HTMs have been attracting great at-tention due to their simple deposition by spin-coating, possibly for low cost and structural versatility. These properties are favorable for large-scale ap-plication. State-of-the-art molecular HTM Spiro-OMeTAD (see Figure 14), of great importance for organic electronics applications such as organic light-emitting diodes and organic field-emitting transistors, is the mostly used HTM for ssDSSCs and recently for PSCs as well.

Figure 14. Synthetic routes of Spiro-OMeTAD.

However, the synthetic procedure for producing Spiro-OMeTAD is tedi-ous, as shown in Figure 14. It includes five steps with a total yield of <50% and therefore the final product is expensive and difficult to obtain with high purity. In addition, Spiro-OMeTAD has a relatively low hole mobility and low conductivity, so that its applicability as fast and efficient hole trans-porter in solar cells is limited for both ssDSSCs and PSCs. In this respect, many kinds of new molecular HTMs have been synthesized and applied in PSCs. The detailed published molecular structures and their performance can be found in the reviews by Ameen et al. and by Yu and Sun.56, 141 Besides, in order to improve the conductivity of the molecular HTMs, chemical doping of the HTMs by different compound has been reported, such as a cobalt complex Tris (2-(1H-pyrazol-1-yl)pyridine) cobalt (III) tri (hexafluorophos-phate),142 2,3,5,6-Tetrafluoro-7,7,8,8-tetracyanoquinodimethane (TCNQ),143 Lewis acid SnCl4,

144 LiTFSI,145 AgTFSI,146 and 1,1,2,2-Tetrachloroethane (TeCA).147

In this thesis, different organic molecules with simple structures were de-signed and synthesized by our close collaborators in KTH, Dr. Bo Xu and

38

Dr. Yong Hua. We tested these HTMs for PSCs and investigated them by different characterization methods. The influences of the molecular struc-tures on their photovoltaic performance were studied in details.

Figure 15. Molecular structures of the HTMs X2, X21, X22, X23, X25, H11, and H12.

N

O

NN

O

O O

O

N

O

NN

O

O O

O

N

O

N N

O O

OO

N

O

N

N

O

O

O

O

N

O

N

N

O

O

O

O

NN OCH3

OCH3H3CO

H3CO

NN OCH3

OCH3H3CO

H3CO

N NH3CO OCH3

H3CO OCH3

N NH3CO OCH3

H3CO OCH3

X2 X25

X23

X21

X22

H12

H11

39

3. Methods

3.1. Device Fabrication The detailed preparation of ssDSSCs and PSCs can be found in the individu-al papers. In this chapter the general procedures of ssDSSC and PSC fabrica-tion are introduced.

The fabrication of ssDSSCs includes several steps: (1) the conducting glass substrate, FTO glass, was cut into certain patterns as desired in solar cells. The substrates are then etched by using zinc powder and HCl solution (2.0 M in water), followed by careful cleaning using water, ethanol, and acetone separately; (2) a thin and compact TiO2 layer is coated on the FTO glass by spray pyrolysis of a titanium precursor solution (0.2M titanium isopropoxide, 2M acetylacetone in isopropanol); (3) a mesoporous TiO2 lay-er is loaded on top of the TiO2 underlayer by doctor blading or spin coating of the diluted TiO2 paste so as to obtain the desired thickness; the obtained films are sintered in the oven under 500oC for 30 minutes (mins); (4) the sintered TiO2 film is dipped into an aqueous TiCl4 solution (40 mM) and kept in the oven (70oC) for 30 mins, followed by another sintering process (500oC for 30 mins); (5) the substrates are then dipped in the dye bath, i.e. 0.2 mM LEG4 in a mixed solution of acetonitrile and tert-butanol (v:v=1:1) for 18h; (6) dye sensitized TiO2 films is covered by spin coating of a Spiro-OMeTAD solution or the polymer HTMs prepared by PEP process, as shown in Figure 16; (7) a thin counter electrode layer of Ag (~200 nm) in ssDSSCs or Au (80 nm) in PSCs, also serving as back contact, is deposited on top of the electrodes.

In the PEP process, a three-electrode electrochemical setup was used to conduct the synthesis of the polymer, with a dye/TiO2/FTO as working pho-toelectrode, an Ag/AgCl/Cl- reference electrode, and a stainless steel plate as counter electrode. The electrolyte solution containing the precursor was ei-ther aqueous micellar or nonaqueous (acetonitrile-based). The supporting electrolyte is 0.1 M LiN(CF3SO2)2 (LiTFSI). In aqueous solution, 50 mM Trition X-100 as surfactant was added in the water solution in order to ena-ble the dissolution of the sparingly soluble or insoluble PEP precursor. The light for PEP (~5 mWcm-2) was provided by a LED (ELFA Distrelec, 3 W, 4.5 V, 700 mA, providing cold white light).

40

Figure 16. Device fabrication procedures for ssDSSCs with spin-coated Spiro-OMeTAD (a) and the polymer HTM prepared from PEP (b).

PEP was conducted by applying to the photoelectrode either a constant po-tential (chronoamperometry) or a constant current (chronopotentiometry) for the time necessary to pass the required quantity of charge so as to obtain a polymer layer of the desired thickness. After the PEP process the electrode was carefully rinsed with water (aqueous PEP) or acetonitrile (organic PEP) to remove the residual precursors.

Figure 17. Cross-sectional image of the ssDSSCs with the photopolymerized PE-DOT as HTMs.

TiO2/dye Spiro-OMeTADSpin coating

TiO2/dye/Spiro-OMeTAD TiO2/dye/Spiro-OMeTAD/Ag

TiO2/dye

V

PEP

TiO2/dye/polymer TiO2/dye/polymer/Ag

(a)

(b)

41

This post-treatment step was followed by spin coating a nonaqueous solu-tion of LiTFSI and tert-butyl-pyridine on top of the dye electrodes. Finally, a 200 nm Ag layer was evaporated on top of the electrode to complete the solar cell. Figure 17 shows the cross-sectional image of the complete ssDSSCs with the PEDOT HTM. It can be seen that the PEDOT is uniform-ly distributed in the whole TiO2 layer.

As regards the part of the present thesis devoted to the PCSs, the em-ployed perovskite material is the mixed-ion perovskite (MAP-bBr3)0.15(FAPbI3)0.85. The detailed fabrication procedures of the PSCs can be found in the published references.148 The steps of cutting and cleaning as well as that of the compact, blocking layer deposition were the same as for the case of the previously mentioned ssDSSCs. Afterwards, a thin layer (~150 nm) of mesoporous TiO2 (particle size=30 nm) was deposited on the substrate. Subsequently, a sintering heat treatment was applied under 500oC. The perovskite precursor solution was prepared by mixing MABr, PbBr2, FAI, and PbI2 with a molar ratio of 1:1:5.6:5.6. The perovskite solution was deposited on top of the TiO2 films by spin-coating, followed after 15s by the addition of some droplets of chlorobenzene. It should be noted that in the PSC the TiO2 layer is often much thinner than that for traditional ssDSSCs. The perovskite substrate was then sintered on a hotplate with a temperature of 100oC for 1 hour. The HTM solution (70 mM in chlorobenzene) was then deposited on the perovskite film by spin-coating. Finally, a layer (80 nm) of Au was evaporated on the substrate to complete the solar cells. Figure 18 depicts the cross-sectional image of the PSCs with the device configuration glass/FTO/TiO2 blocking layer/mesoporous TiO2/perovskite/HTM/Au.

Figure 18. Cross-sectional image of PSCs with Spiro-OMeTAD as HTMs.

42

3.2. Device Characterization 3.2.1 Current-Voltage Measurement The current density vs. voltage (J-V) measurement is the basic characteriza-tion step of a solar cell in order to obtain the essential photovoltaic parame-ters, such as the short-circuit current density (Jsc), the open circuit voltage (Voc), the fill factor (FF), and the light-to-electricity power conversion effi-ciency (η). The solar cell is measured at a controlled light intensity by using a solar simulator lamp which can provide a spectrum (light intensity=1000 Wm-2) similar to the full-sun under AM1.5G illumination. The J-V curve (see Figure 19) is recorded by applying a linearly-varying voltage scan be-tween photoelectrode and counter electrode, and monitoring the photocurrent change; the voltage range encompasses both 0 V (short-circuit condition) and Voc (open-circuit condition); the upper limit is usually slightly larger than Voc, the lower limit is 0 V or somewhat beyond 0 V in the reverse direc-tion, i.e. with the photoelectrode more negative than the counter electrode; such an extension is necessary if the current under illumination is not poten-tial-independent at the vicinity of 0 V.

Figure 19. J-V curves of a device under light and in dark.

For PSCs, the J-V curves are often highly dependent on the scan direction, which is called hysteresis behavior. The reasons for the hysteresis are still unclear. Therefore, the scan direction and the scan rate are important for obtaining reliable photovoltaic parameters so that the application of a trian-gular scan, as in electrochemical 3-electrode cyclic voltammetry, is pre-

0.0 0.2 0.4 0.6 0.8 1.0 1.2-0.005

0.000

0.005

0.010

0.015

0.020

0.025

0.030

Jsc

Voc

Jsc

Voc

J /

Acm

-2

Voltage / V

Pmax

43

ferred. J-V curves can also be measured under the dark condition, which can provide information on the electron-transfer reaction from the FTO side to the HTM. However, it should be noted that in several cases the recombina-tion is faster under light than in the dark due to an increased electron concen-tration in TiO2 as a result of photoexcitation. In efficient solar cells, the dark current should be minimized by increasing the shunt resistance in the devic-es, in order to avoid the electron loss at the operation condition under light. In particular, the dark current at low bias potentials around 0 V should be virtually zero, otherwise such a dark-condition shunt may be deleterious to the cell performance, especially under low light intensity operation.

The overall solar cell efficiency η is determined by the ratio of maximum power density (power per unit area, Pmax) of the device and the incident solar energy (Pin). The maximum power Pmax of the device is expressed in terms of Jsc, Voc, and FF according to the equation

FFVJP ocscmax

so that

in

ocsc

in P

FFVJ

P

P max

The open-circuit voltage Voc is determined by the difference of the Fermi levels at photoelectrode EF(TiO2) and counter electrode EF(CE) as

0

)()( 2

e

EEV CEFTiOF

oc

where e0 is the elementary charge. The maximal Voc, namely Voc

*, may be estimated by the difference be-tween the conduction band edge (Ecb) of TiO2 and the Fermi level, EF,redox-1, of the solid-state HTM (at the open-circuit condition, EF(CE)=EF,redox-1).

0

1,*e

EEV redoxFcb

oc

This may be useful for a first guess of the possibility for a HTM to be used in the solar cell. However, this is only a rough estimation of the possi-ble voltage.

Jsc is the maximum photocurrent extracted from the device, which is con-trolled by the light harvesting efficiency of the light absorber and the charge-transfer yield at the interfaces; normally the charge recombination rate is negligible under short circuit. Therefore, the optimization of the energy level matching between different components and the improvement of the absorp-tion coefficient of the light absorber are critical for enhancing the Jsc. The FF of a device can be influenced by (a) the series resistance (resistance of the

44

component, such as the HTM, TiO2, Au counter electrode and especially, FTO); (b) the overpotential at the HTM/Au interface (corresponding to the deviation from equilibrium under current); (c) the shunt resistance, inter-posed between photoelectrode and counter electrode, of the device; for ex-ample a shunt is created if the gold layer penetrates the TiO2 pores and con-tacts the FTO glass if the underlayer is defective; similarly, if the polymer hole conductor generated by PEP can contact the FTO through defects in the underlayer and create a shunt contact. On the other hand the resistance at the underlayer TiO2 and FTO contact is usually very low so that this contact is considered as ohmic. The shunt resistance may also result from recombina-tion of the photogenerated charges, which can occur by many different reac-tions for example directly in the dye molecule or in the perovskite layer.

3.2.2 Incident Photon-to-Current Conversion Efficiency The incident photon-to-current conversion efficiency (IPCE), also called external quantum efficiency, is very useful for determining the fraction of the incident photons converted to electrons flowing in the external circuit. The IPCE spectrum of the solar cell is usually measured by recording the generated short-circuit photocurrent by illuminating with tunable mono-chromatic light; such light is usually generated by passing the output of a white-light source through a monochromator. However, the IPCE spectra can also be measured under a certain applied cell voltage, for example at the voltage where the solar cell generates highest power. It is also possible to superimpose a constant bias light to the tunable monochromatic light. The IPCE value depends on different parameters including the light harvesting efficiency (LHE), electron injection efficiency from the dye to TiO2 (φinj), and the charge collection efficiency (φcoll), as shown in the equation below:

collinjLHEIPCE

LHE is defined as the fraction of the incident light absorbed by the de-vice. For monochromatic light, LHE(λ) is derived from the absorbance (A(λ)) of a dye sensitized TiO2 film,

)(101)( ALHE

and φinj is defined as

deactinj

injinj kk

k

where kinj is the rate constant of injection reaction D*→e-+D+ and, kdeact is the rate constant of deactivation D*→D. This deactivation can be radiative (fluorescence) or non-radiative (heat evolution).

45

If Rcoll is the rate of electron collection to the FTO contact, i.e. the rate of electron transport from TiO2 into the FTO contact, Rrec is the rate of electron recombination with the oxidized hole conductor and/or the oxidized dye and Rinj is the rate of electron injection from the excited dye into semiconductor,

reccollinj RRR

The collection efficiency is defined as,

inj

rec

inj

coll

collrec

collcoll R

R

R

R

RR

R

1

For ssDSSCs, in order to model the recombination kinetics it is normally assumed that the recombination reaction is first-order. In contrast, the re-combination kinetics is presented with a different model in PSCs.149 In ssDSSCs, Rrec can be expressed in terms of the electron concentration (e-), the oxidized dye concentration (D+) and the oxidized HTM concentration (HTM+) as

))(())(( ,,

DekHTMekR DrecHTMrecrec

It should be noted that the dye regeneration efficiency has been included in φcoll. In case of fast dye regeneration (D+) will be very low so that

))((

))(())((

,

,,

HTMekR

HTMekDek

HTMrecrec

HTMrecDrec

Since the rate of electron injection is equal to the sum of the rates of elec-tron collection and electron recombination, the collection efficiency is given as

)(

))(())((1

))(())((

)(

1

*,,

,,

*

Dk

DekHTMek

DekHTMekR

DkR

R

R

inj

DrecHTMreccoll

DrecHTMrecrec

injinj

inj

reccoll

46

On the basis of the IPCE spectrum, the photocurrent expected from the

standard tabulated AM1.5 solar spectrum can be calculated. If )(JAM1.5ph is

the photon flux distribution, i. e. the number of photons with wavelengths between and d divided by d , the corresponding solar cell maxi-mum photocurrent, e.g. at short circuit, will be

dFdJ sc )(IPCE)(JAM1.5ph

AM1.5

where F is Faradaic constant, and the calculated integrated photocurrent AM1.5scJ between the limits min and max of the solar spectrum

max

min

)(IPCE)(JAM1.5ph

AM1.5

dFJsc

The calculated integrated short-circuit solar cell AM1.5 current, within good approximation, should in principle be the same current as that meas-ured by exposing the cell to the output of a solar simulator. A prerequisite for this is that there is no diffusion limitation, i.e. that in a laboratory exper-iment the photocurrent will be proportional to the light intensity in an exper-iment where the latter is varied by interposing neutral density filters (e.g. 5%, 10%, 50%) to the beam so that the reduced-intensity beam has a spec-trum identical to that of the full beam. Figure 20 shows a typical IPCE spec-trum of PSCs with Spiro-OMeTAD as HTM.

Figure 20. The IPCE spectrum of PSC.

3.2.3 Photocurrent and Photovoltage Decay Measurements Transient photocurrent and photovoltage decay measurements are commonly used to characterize the charge-transport and transfer processes occurring in

400 500 600 700 8000

20

40

60

80

100

IPC

E /

%

Wavelength / nm

47

the complete solar cell devices under the operational conditions. Generally, a small modulation of the light intensity (∆I) is applied on top of a baseline light intensity I0 to the solar cell (see Figure 21a),

III 0

The I0 is controlled by applying different voltages to a white LED (Luxeon Star 1W). For example, a baseline light intensity of 100 mWcm-2 is attained by applying a voltage of 3.6 V to which a small light modulation is applied by introducing an additional voltage of 10-50 mV. Since both photo-voltage and photocurrent are dependent on the light intensity, applying a light modulation is a straightforward way to induce a transient change of Jsc and Voc. The decay kinetics of the transient photocurrent or photovoltage caused by this small modulation of illumination can be monitored in order to study the charge recombination and charge transport properties in the devic-es. The time constant can be derived by fitting of the transient decay by sin-gle exponential function, as follows:

resp

n

tsc

toc

JeJtJ

VeVtV

/

/

)(

)(

where V(t) and J(t) are the transient photovoltage and photocurrent. The Voc and Jsc are measured when a certain light intensity I0 is introduced. ∆V and ∆J are the changes of voltage and current induced by the ∆I. The increased photovoltage and photocurrent will relax to a steady state after a while. This decay process is measured in order to investigate the kinetics of different processes.

48

Figure 21. (a) Photocurrent rise and decay by applying a small modulation on top of a baseline light intensity. (b) The normalized decay of the photocurrent.

These measurements are instructive for the qualitative study of similar sys-tems involving the comparison of different components, such as HTMs, alt-hough the derived time constants may not provide sufficiently accurate and quantitative information for individual processes. In the example shown in Figure 21b the photocurrent decay measurements can be used to investigate the charge transport rate in the ETMs or HTMs in PSCs.

3.2.4 Photoinduced Absorption Spectroscopy Photoinduced absorption (PIA) spectroscopy can be used to characterize the dye regeneration process in ssDSSCs. Usually a thin dye-sensitized TiO2 electrode, which can allow the light to pass through, is used in a PIA meas-urement. There is a pump-probe setup in the PIA system. A square-wave modulated monochromatic pump light beam (460 nm used in this thesis) is superimposed on a white probe beam supplied by a white light (20 W tung-

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

2.78

2.80

2.82

2.84

2.86

2.88

2.90

2.92

photocurrent decay

perturbation light off

I sc /

mA

Time / s

perturbation light on

photocurrent rise

0.000 0.001 0.002 0.003 0.004 0.005

0.0

0.2

0.4

0.6

0.8

1.0

No

rmal

ized

cu

rre

nt

Time / s

(a)

(b)

49

sten-halogen lamp). The probe light can be partially absorbed by the species in the electrode, and the transmitted probe light can be collected on a mono-chromator and then monitored by a silicon photodiode connected to a current amplifier and lock-in amplifier. The PIA spectrum consists of the difference ∆A of the absorbance of the measured electrode before and after the illumi-nation plotted vs. wavelength.

Figure 22. PIA signals of LEG4-sensitized TiO2 films with and without HTM PE-DOT.

The PIA is often used to compare the dye/TiO2 electrodes with and without the charge-transport medium, such as a redox couple (liquid DSSCs) or a solid-state HTM (ssDSSCs). In the case of a bare dye/TiO2 electrode without HTM, the light absorption by the oxidized dye will be observed. As shown in Figure 22, a positive ∆A in the range of 700-850 nm for TiO2/LEG4 elec-trode (black squares) is caused by the light absorption of oxidized LEG4. After the deposition of the HTM PEDOT, a decreased ∆A (black circles) is obtained for the LEG4/TiO2/PEDOT electrode. This lower ∆A results from the dye regeneration reaction: LEG4++PEDOT→LEG4+PEDOT+.

Therefore, the PIA is a reliable method to qualitatively study the dye re-generation process in ssDSSCs.

3.2.5 Hole Conductivity and Hole Mobility Measurements The electrical conductivity of the HTM is measured by using a two-probe setup. The device configuration glass/TiO2/HTM/Ag was used for this meas-urement. The detailed preparation of TiO2 film and HTM layer can be found in the previous chapter. The measurement is carried out by applying a volt-

600 700 800 900 1000 1100 1200 1300 1400 1500 1600

-2.0x10-4

-1.0x10-4

0.0

1.0x10-4

2.0x10-4

A

Wavelength / nm

TiO2/LEG4

TiO2/LEG4/PEDOT

50

age (-2 V to 2 V) between two Ag contacts and the flowing current is rec-orded. The conductivity is calculated by using equation:

dLR

W

where σ is conductivity of the HTM, W is the distance between two Ag elec-trodes (W=2 mm in this thesis), L is the length of the Ag channel (L=10 mm in this thesis), d is the film thickness of the HTM, and R is the resistance which can be calculated from the recorded I-V curves.

A high charge carrier mobility of the charge-transport medium is of key importance for obtaining a long charge diffusion length in the solar cell. Therefore, the hole mobility is one of the most important parameters for the HTM. The hole mobility is obtained by measuring the space-charge limited current (SCLC). The device configuration of ITO/PEDOT:PSS/HTM/Ag was employed during the SCLC measurement. The SCLC is recorded by applying a voltage on the device and measuring the current. Under the ap-plied voltage, the injected charge carriers form the space charge layer which has an electric field of opposite direction from that of the applied bias volt-age. This can induce a limited current at a high voltage, which is described as follows:

3

2

08

9

d

VJ r

where J, µ, ε0, εr, V, d are the current density, hole mobility, vacuum permit-tivity (8.85×10-12 F/M), dielectric constant (εr=3 in this thesis), applied volt-age and film thickness. The hole mobility can be derived from the above equation and the measured I-V current.

3.2.6 Photoluminesence In order to analyze the light emission of the perovskite film as well as to study the electron quenching at the interface between the perovskite and the HTM, photoluminescence (PL) was used in this thesis. Light of excitation wavelength of 500 nm was applied and the device configuration was glass/ZrO2/perovskite/HTM. A continuous incident light was directed on the substrate with an angle of 60o during the measurement.

In the case of bare perovskite film without HTM, a high PL signal result-ing from the radiative recombination is usually observed. This PL intensity of the bare perovskite film is dependent on the quality of the perovskite film and, therefore, perovskite films of the same-quality are required in order to obtain a reliable comparison. When the HTM is deposited on the perovskite film, the PL signal is decreased due to the hole injection from the perovskite layer to the HTM. Therefore, the difference in the PL intensity for different HTMs could provide information on the yield of the interfacial hole transfer.

51

4. Results and Discussion

4.1 Aqueous PEP versus Organic PEP (Paper I) For the PEP approach, the solvent acetonitrile was always used in the previ-ously published work because the precursor, such as bis-EDOT, has a high solubility in it. However, acetonitrile is expensive and volatile. So we tried for the first time to develop aqueous PEP by using water as the solvent.150 In order to dissolve the employed precursor bis-EDOT in water, a supporting micellar medium, Triton X-100, was added in the water to improve its solu-bility. We will name the PEP based on acetonitrile and water as “Organic PEP” and “Aqueous PEP” respectively, abbreviated as A-PEP and O-PEP respectively. The devices based on organic PEP and aqueous PEP were fab-ricated by following the same procedures. The two types of dyes principally used for this project were organic metal-free and Ru-based dyes.

Table 2. J-V parameters of ssDSSCs with organic PEP and aqueous PEP

Medium η / % Voc / V Jsc / mAcm-2 FF

LEG4 Organic 5.6 0.91 10.8 0.57

Aqueous 5.2 0.84 10.9 0.56

D35 Organic 4.6 0.83 8.2 0.68

Aqueous 3.8 0.78 8.6 0.56

Z907 Organic 1.5 0.51 4.5 0.66

Aqueous 0.07 0.42 0.4 0.39

Table 2 presents the photovoltaic parameters of the ssDSSCs based on or-ganic PEP and aqueous PEP. The device based on aqueous PEP (5.2% for LEG4 dye) showed comparable PCE as the ones with organic PEP (5.6% for LEG4 dye). By comparing their photovoltaic parameters, we found that the devices based on organic PEP had a much higher Voc and a slightly lower photocurrent Jsc for both D35 and LEG4 dye. Therefore, it is very interesting to understand the difference between the organic PEP and aqueous PEP.

52

In order to study the reasons for the difference in Voc and Jsc observed for two types of PEP, transient photovoltage decay (electron lifetime) and PIA spectra were carried out, as shown in Figure 23. Electron lifetime measure-ment can provide information on the rate of the photoinduced electron re-combination in the devices. Electron lifetime significantly affects the elec-tron concentration in the TiO2 and thus the Voc of the devices if it is assumed that the conduction band position of TiO2 is the same for the two types of devices. It is noted in Figure 23a that the device based on aqueous PEP and LEG4 (A-LEG4) showed a much shorter electron lifetime than that with organic PEP and LEG4 (O-LEG4). This result can indicate that the low Voc obtained in A-LEG4 is caused by fast electron recombination with holes in the formed PEDOT. For the difference in Jsc, we could exclude the influence of some factors such as light harvesting efficiency and electron injection efficiency because the same dye LEG4 is used and the electron injection is believed to be efficient. Therefore, the different Jsc could be caused by dif-ferences in the dye regeneration process because the PEDOT could have different properties in the two cases. PIA measurements were carried out to study the dye regeneration process, as shown in Figure 23b. We observed that the A-LEG4 based device showed a more efficient dye regeneration than the O-LEG4 one, as evidenced by the decreased absorption of the oxidized LEG4 in the range of 700-800 nm. This could explain the slightly improved Jsc in A-LEG4 than O-LEG4.

Figure 23. (a) Electron lifetime measurements, (b) PIA signals for A-LEG4 and O-LEG4.

In these two devices, the only difference lies in the details of the PEP pro-cess for the formation of PEDOT, so we believe that the differences in the electron recombination rate and dye regeneration efficiency are related to the properties of PEDOT. In order to compare the PEDOT in the two cases, NIR absorption spectral measurements of D35/TiO2/PEDOT were performed, which are shown in Figure 24. The absorption spectra were recorded in the

600 700 800 900 1000-2.0x10-4

-1.0x10-4

0.0

1.0x10-4

A

Wavelength / nm

TiO2/LEG4

TiO2/LEG4/O-PEDOT

TiO2/LEG4/A-PEDOT

0.70 0.75 0.80 0.85 0.90 0.95

1E-3

0.01

O-LEG4

A-LEG4

Life

time

/ s

Voc

/ V

53

range from 500 nm to 2000 nm. It should be noted that there is no contribu-tion in light absorption from the dye molecules from 800 nm to 2000 nm (see the curve with triangles). However, for both substrates with PEDOT from organic and aqueous PEP considerable light absorption occurs in that range. The A-PEDOT presents relatively low absorption intensity in the whole wavelength range but with a broad absorption peak at 1200 nm. In contrast, the O-PEDOT exhibited a broad and gradually increasing absorp-tion from 800 nm to 2000 nm. We assign this behavior to the different elec-tronic properties of the formed PEDOT. A-PEDOT has a feature with main contribution from the polaron state, corresponding to a less delocalized elec-tronic structure with shorter polymer chains. In contrast, O-PEDOT possess-es bipolaron states, attributed to a highly delocalized electronic structure and longer polymer chains. This difference in the properties of obtained PEDOT could be due to the different polymerization mechanism. In the O-PEP case the high concentration of the precursor radicals and their fast diffusion through the TiO2 pores could favor fast polymerization and formation of longer chains.

Figure 24. NIR absorption spectra of the substrates TiO2/D35 with and without the HTM PEDOT.

If we correlate the properties of the PEDOT and the obtained electron life-time and PIA results, we could make a hypothesis: a relatively shorter poly-mer obtained in A-PEDOT compared to that in O-PEDOT could result in a larger contact area between the PEDOT and the dye molecules, and thereby to more efficient dye regeneration by the HTM PEDOT. This agrees with the faster dye regeneration for A-PEDOT than O-PEDOT as evidenced by the PIA experiments (see Figure 23b). In addition, the larger contact area be-tween PEDOT and dyes in A-PEDOT could also result in faster electron

600 800 1000 1200 1400 1600 1800 20000

1

2

3

4

Abs

orb

an

ce

Wavelength / nm

TiO2/D35/O-PEDOT

TiO2/D35/A-PEDOT

TiO2/D35

54

recombination in the devices compared to that for O-PEDOT, which is con-sistent with electron lifetime measurements (see Figure 23a).

Therefore, the environmentally friendly aqueous PEP approach exhibited the advantages of fast dye regeneration due to the shorter polymer chains. This could be interesting for further optimization to enhance the PCE of ssDSSCs.

4.2 New Precursors for PEP (Paper II and III) Bis-EDOT is the mostly used precursor in the PEP because the precursor bis-EDOT has an appropriately low oxidation potential to conduct the PEP pro-cess. However, it is difficult to synthesize, especially with high purity. There-fore, it is highly desirable to develop simple and cheap precursors for the PEP process in ssDSSCs.

EDOT is the monomer of PEDOT; it is cheap and widely commercially available due to its relatively simple synthesis. However, EDOT has not been successfully explored as the precursor in PEP due to its high oxidation poten-tial (~1.6 V/SHE) in an acetonitrile-based solution, as shown in Figure 25. However, it is noteworthy (in paper I) that the oxidation potential of the pre-cursor bis-EDOT is decreased from 0.9 V/SHE in organic solvent to 0.6 V/SHE in water-based solution. Generally, the lower oxidation potential of the precursor facilitates the PEP to proceed because the driving force for the PEP process is defined by the energy level difference between the dye and the pre-cursor in the electrolyte solution. Inspired by this, we measured the onset oxi-dation potential of the EDOT in organic and aqueous solution, as shown in Figure 25.

Figure 25. Cyclic voltammetry curves of EDOT in acetonitrile and water-based solution.

0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.00.0

5.0x10-2

1.0x10-1

1.5x10-1

2.0x10-1

2.5x10-1

3.0x10-1

3.5x10-1

4.0x10-1

I / m

A

Voltage / V/SHE

EDOT / organic

EDOT / aqueous

55

It is noted that the onset oxidation potential of EDOT is decreased to 0.9 V/SHE in aqueous solution from 1.6 V/SHE in acetonitrile.151 Therefore, in the case of aqueous PEP, it is possible from the energy level point of view to polymerize the EDOT by photoexcitation of the metal-free organic dyes.

Table 3. Photovoltaic parameters of the devices based on the PEDOT and D35.

Devices Light intensity (mWcm-2) ŋ (%) Voc (V) Jsc

(mA cm-2) FF

D35 O-PEDOT 100 0.04 0.62 0.3 0.25

D35 A-PEDOT

100 3.02 0.82 6.2 0.60

46 3.41 0.73 3.3 0.65

11.4 3.04 0.66 0.8 0.61

The devices were fabricated by using EDOT as the precursor in both organic PEP and aqueous PEP, and the photovoltaic parameters are shown in Table 3. It is noted that the PCE is significantly improved from 0.04% (organic PEP of EDOT) to 3.0% (aqueous PEP of EDOT) in the devices with D35 dye. This further illustrates that the oxidation potential matching between the dye and the precursor is of key importance for performing PEP efficiently. Therefore, the aqueous PEP paves the way for the potential use of low-cost EDOT as the precursor in PEP for ssDSSCs.

An EDOT analogue, 3, 4-ethylenedioxpyrole (EDOP) is another precur-sor which is commercially available. EDOP has combined two advantages: low cost due to the simple chemical structure, as that of EDOT, as well as the low onset oxidation potential in acetonitrile like bis-EDOT. The PEP of EDOP was previously tested at a Ru-dye electrode; the corresponding ssDSSCs with the thus generated polymer, PEDOP, as HTM had a low effi-ciency of 0.01% due to the fast electron recombination of the formed PE-DOP and the employed Ru-dye.152 In this work we tried to perform the PEP of EDOP with the organic metal-free dyes D35 and D21L6.153

Table 4. Photovoltaic parameters of ssDSSCs with dyes D35, D21L6 and Ru-Z907.

Devices Light intensity (mWcm-2) ŋ (%) Voc (mV) Jsc

(mA cm-2) FF

D35 100 4.34 825 8.0 0.66

D21L6 46 4.53 785 3.9 0.67

100 3.05 645 7.9 0.59 46 3.38 630 4.1 0.61

Ru-Z907 100 0.46 440 2.0 0.53 46 0.46 395 1.0 0.53

56

It can be seen from Table 4 that a high PCE of 4.34% was achieved for D35-based ssDSSCs with PEDOP as HTM, compared to a low PCE of 0.46% for the devices with Ru-Z907 dye. The device with D21L6 showed a PCE of 3.05%. By comparing their photovoltaic parameters, we can conclude that (a) the devices with metal-free organic dyes presented much higher Jsc than that for Ru-Z907; (b) the D35-based device showed a much higher Voc (825 mV) than these based on D21L6 (645 mV) and Ru-Z907 (440 mV); (c) Organic dye D35 sensitized devices exhibited a higher FF (66%) than these sensitized by D21L6 (59%). The respective J-V curves are shown in Figure 26.

Figure 26. J-V curves of ssDSSCs with dyes D35 (circle), D21L6 (square) and Ru-Z907 (up-triangle).

From the J-V curves, we can see that the lower FF in the D21L6-based de-vices compared to the D35-based ssDSSCs was mainly caused by the lower shunt resistance which can be observed by the fast decrease of the photocur-rent when applying low bias voltages (0 V-0.5 V). This could be because of the fast electron recombination with the oxidized D21L6 or the PEDOT HTM. For D35-based devices, the obtained higher Voc and FF could be mainly due to the bulky molecular structure of the dye which acts as a block-ing layer to retard the electron recombination with the formed PEDOP. Simi-larly to the previous report, the Ru-Z907-based ssDSSCs showed a poor photovoltaic performance with low Voc, Jsc and PCE. The low photovoltaic parameters could be due to (a) the low extinction coefficient of Ru-Z907; (b) the poor blocking effect between the e- (TiO2) and h+ (PEDOP) and thus the fast charge recombination. The electron loss due to the recombination be-tween e- (FTO) and h+ (PEDOP) can be investigated by recording the dark

0.0 0.2 0.4 0.6 0.8

0

2

4

6

8

J / m

Acm

-2

Voltage / V

57

current (see Figure 26). It can be seen that D35-based devices showed much lower dark current compared to these based on D21L6 and Ru-Z907, demonstrating that D35 has a much better recombination blocking effects than the others. Therefore, the bulky chemical structure of the dye is critical indeed for retarding the charge recombination in ssDSSCs.

Figure 27. (a) UV-Vis absorption of TiO2/dye films. (b) Electron lifetime meas-urements of the ssDSSCs with PEDOP and dyes D35, D21L6 and Ru-Z907.

In order to further justify the origins of the differences in Jsc and Voc values for various dyes, measurements of the UV-Vis spectra of dye/TiO2 and the electron lifetime were performed, as shown in Figure 27.

From the UV-Vis spectra of three dyes on TiO2 films (see Figure 27a), it is evident that the metal-free organic dyes D35 and D21L6 possess a much higher light harvesting ability than Ru-Z907, and this could be the main rea-son for the lower Jsc of Ru-Z907-based devices. Figure 27b illustrated the electron lifetime values which are derived from transient photovoltage decay

0.4 0.5 0.6 0.7 0.8 0.9

1E-3

0.01

Life

time

/ s

Voc / V

D35

D21L6

Ru-Z907

400 500 600 700 8000.0

0.5

1.0

1.5

2.0

Ab

sorb

an

ce

Wavelength / nm

D21L6

D35

Ru-Z907

(a)

(b)

58

measurements. We can see that the devices with D35 showed a much longer electron lifetime than that for D21L6 and Ru-Z907. This longer electron lifetime in D35-based ssDSSCs contributed to a higher Fermi level of TiO2 and thus a higher Voc value than those for D21L6 and Ru-Z907. This can well explain the Voc difference in the three types of devices. Therefore, the dye molecular structures play important roles for ssDSSCs in terms of both light harvesting and interfacial electron recombination. So it would be very important to fundamentally investigate the effects of dye structures on the PEP process and the photovoltaic performance of the ssDSSCs.

4.3 Effects of Dye Molecular Structures (Paper IV) In this paper, the influences of the dye molecule structures on the PEP and photovoltaic performance of the ssDSSCs were systematically studied.154 A series of metal-free organic dyes (C218, MK253, and LEG4) and two Ru-dyes (Z907 and B11) were chosen for PEP. It should be noted that (a) three metal-free organic dyes have different donor or linker parts, which affect their energy levels and light harvesting ability; (b) the difference in the bulky structure for three metal-free organic dyes could influence the electron re-combination time; (c) the two Ru-dyes showed significantly different extinc-tion coefficients. The detailed optoelectronic parameters can be found in Table 1.

Firstly, the effect of the dyes on the PEP process was studied by monitor-ing the transient current during the PEP process by the chronoamperometry (0.4 V/SHE) method, as shown in Figure 28. The PEP current can provide information on the rate of the PEP.

Figure 28. Photocurrent variation with time during the PEP process by applying a potential of 0.4 V/SHE for different dyes under the same light intensity.

0 200 400 6000.00000

0.00001

0.00002

0.00003

0.00004

light off

Cu

rren

t / A

Time / s

LEG4

MK253

C218

B11

Z907

light on

59

It can be seen that a very low current was obtained when the light is off. In contrast, after irradiation a high transient photocurrent was observed for all dyes due to the sudden photoexcitation of the dyes followed by the charge separation. After a while, the photocurrent is different for various dyes. Spe-cifically, the Ru-Z907 electrode showed a relatively low photocurrent com-pared to B11 due to its low light absorption coefficient and possibly the fast interfacial electron recombination for the former. For metal-free organic dyes, the LEG4-based electrode showed a much higher photocurrent than the C218- and MK253-based ones. This can be attributed to (a) the higher ex-tinction coefficient of LEG4 than C218; (b) more positive oxidation poten-tial of LEG4 compared to MK253, facilitating an efficient PEP process; (c) the more bulky structure of LEG4 than C218 retards the interfacial recombi-nation more effectively. Overall, in order to conduct the PEP efficiently the dye molecule should have a high extinction coefficient, a bulky molecular structure and an appropriate energy level in relative to the precursor of PEP.

Table 5. Photovoltaic parameters of the ssDSSCs obtained at light intensity of 100 mWcm-2 with different dyes.

Parameter/Dye LEG4 C218 MK253 Ru-B11 Ru-Z907

Jsc / mAcm-2 13.4 9.2 10.9 11.9 4.4 Voc/ mV 830 710 680 605 515

FF 0.64 0.63 0.50 0.56 0.58 η / % 7.11 4.12 3.75 4.05 1.01

The photovoltaic parameters of ssDSSCs based on various dyes and PEDOT are shown in Table 5. The device with LEG4 dye gave a high PCE of 7.11% with a Jsc=13.4 mAcm-2, Voc=830 mV and FF=0.64. In contract, metal-free organic dyes C218 and MK253 showed a lower performance with PCE of 4.12% and 3.75%, respectively. In addition, the device with Ru-B11 exhibit-ed a much higher PCE of 4.05% than 1.01% for Ru-Z907. The main obser-vation in the photovoltaic parameters for different dyes is the variation of the Voc although the same HTM PEDOT was employed. As discussed in the previous chapter, the Voc of the ssDSSC is primarily determined by the ener-gy difference between the energy levels of TiO2 and the HTM. Additionally, the electron recombination at the TiO2/dye/PEDOT interfaces is a critical factor affecting the Voc of the device. In order to study the interfacial charge recombination, the transient photovoltage decay was recorded; the derived electron lifetime is shown in Figure 29. Among the three organic dye-sensitized solar cells considered, the LEG4-based device showed a much longer electron lifetime than these based on C218 and MK253. This agrees well with the Voc difference presented in Table 5. Similarly Ru-B11-based device presented a longer electron lifetime than that with Ru-Z907. The dif-

60

ference in the electron lifetime can be attributed to the dye molecular struc-ture and PEDOT properties.

Figure 29. Electron lifetime measurement for the ssDSSCs with different metal-free organic dyes (a) and metal-complex dyes (b).

4.4 New Molecular HTMs for PSCs (Paper V, VI, VII) The HTM layer is an essential part in PSCs which has different functionali-ties, such as (a) extracting the holes from the perovskite layer; (b) conduct-ing the holes through the HTM layer; (3) preventing the recombination caused by the direct contact between the Au electrode and the perovskite layer; (4) isolating the perovskite layer from the moisture in the atmosphere. So far the most commonly used HTM is Spiro-OMeTAD. However, as men-tioned above Spiro-OMeTAD has disadvantages of high cost and relatively low conductivity. Therefore, designing new alternative HTMs for PSCs, and understanding the relationship between molecular structure and performance are required for the future application of the HTMs in PSCs. In this chapter, we will discuss how the molecular structures of the HTMs affect the optoe-lectronic properties in the PSCs devices. The molecular structures of the HTMs to be discussed are shown in Figure 15.

0.4 0.5 0.6 0.7 0.8 0.9

1E-3

0.01

C218 MK253 LEG4

t / s

UCELL(OC)

/ V

(a)

0.3 0.4 0.5 0.6 0.71E-3

0.01

0.1

B11 Z907

t / s

UCELL(OC)

/ V

(b)

61

Firstly, we investigated the influence of the alkyl chains on the triphenyl-amine-based HTMs by grafting the alkyl chains with different length and positions on the reference HTM X2.148

Table 6. Optoelectronic parameters of the HTMs.

The redox potentials and the maximum absorption peaks for different HTMs are listed in Table 6. It is noted that X2 and X23 exhibited higher Eredox-1 than X21 and X22. The less positive Eredox-1 for the HTMs X21 and X22 could be attributed to the more planar bifluorene groups due to the fluorene structure. In addition, X21 and X22 showed a ~40 nm red shift compared to those of X2 and X23 in the maximum absorption peak. This can illustrate again that a more delocalized electron configuration in X21 and X22 is due to the fluo-rene structure compared with X2 and X23. The less positive Eredox-1 level for X21 and X21 could be favorable for the hole transfer from the perovskite to the HTMs. Besides, the planar structure and delocalized electron distribution in X21 and X22 could provide good charge transport properties.

Figure 30. (a) Device structure, (b) charge transport pathways, (c) I-V curves and (d) IPCE of the PSCs with different HTMs.

300 400 500 600 700 8000

20

40

60

80

100

IPC

E / %

Wavelength / nm

X2 X21 X22 X23

+

FTO

TiO2

perovskite

hv

X2

X21X22

X23

Au

(a) (b)

(c) (d)

HTM λabs, max nm

Eredox-1 [V vs SHE]

EHOMO [eV]

Dihedral Angle (degree)

X2 371 0.725 -5.04 146.03 X21 403 0.576 -4.90 179.46 X22 404 0.577 -4.90 179.11X23 365 0.724 -4.99 145.41

62

In order to study the effects of the alkyl chains on the performance of the HTMs in PSCs, the perovskite devices were fabricated by following the de-vice configuration shown in Figure 30a. The details of the device fabrication can also be found in chapter 3. Figure 30c shows the J-V curves and the pho-tovoltaic parameters of the PSCs with different HTMs. It shows that for X21 a highest PCE of 17.33% is obtained compared to 14.6%, 15.45% and 12.31% for X22, X2 and X23. By comparison of the photovoltaic parameters it is found that relatively lower Voc values were obtained for X21 and X22 compared to X2 and X23. This could be caused by the less positive redox potential Eredox-1 (higher HOMO) of the X21 and X22 in relative to that for X2 and X23. In addition, X21-based devices showed a higher Jsc and thus the higher PCE than that for X22. In order to investigate the causes of the different photovoltaic performance for these HTMs, we studied different charge transport pathways occurred in PSCs, as exhibited in Figure 30b. We performed photoluminescene (PL) measurements in order to investigate the hole-transfer process between the perovskite and various HTMs and, addi-tionally, transient photocurrent decay measurements in order to study the charge transport in HTMs.

Figure 31. (a) PL signal of the device glass/ZrO2/perovskite with and without HTMs. (b) Transient photocurrent decay of the PSCs with different HTMs.

650 700 750 800 850 900 950 1000 1050 1100

0

1x106

2x106

3x106

4x106

5x106

PL

in

ten

sity

Wavelength / nm

No HTM X2 X21 X22 X23

650 700 750 800 850 900

0

1x105

2x105

3x105

4x105

5x105

PL

in

ten

sit

y

Wavelength / nm

0.00 0.02 0.04 0.06 0.08 0.100.0

0.1

0.2

0.3

0.4

No

rmal

ized

ph

oto

cu

rren

t

Time / s

X2 X21 X22 X23

0.0 0.5 1.0 1.5 2.00.0

0.2

0.4

0.6

0.8

1.0

Time / ms

(a)

(b)

( )

63

On the basis of the PL measurements in Figure 31a it can be seen that for X21 and X23 the hole transfer from the perovskite to the HTMs is more efficient than for X2 and X22. This can explain the lower Jsc obtained for X22-based PSCs than that for X21-based ones. According to the data of Figure 31b HTM X21 and X22 have much more efficient hole transport than that for X2 and X23. Therefore, the fluorene structure with a good planarity is favorable for the hole transport in the HTMs. This could explain the supe-rior photovoltaic performance obtained for X21-based PSCs.

In a following development, X2 was used as the reference HTM once more; a similar HTM X25 was developed by simply linking the phenyl groups by a carbon-carbon single bond. The chemical structures of X2 and X25 can be found in chapter 2.155

From the electrochemical and optoelectronic characterization, it can be seen that X25 showed a more positive Eredox-1 and a higher conductivity than X2, as shown in Table 7. These two parameters are very important to achieve high Voc and Jsc in PSCs.

Table 7. Optoelectronic parameters of the HTMs

Figure 32 depicts the photovoltaic properties of the PSCs with X2 and X25; according to Figures 32a and 32b , X25-based devices showed a much high-er PCE (17.4%) with a higher Voc (1100 mV) and Jsc (22.64 mAcm-2), com-pared to those based on X2 (PCE=14.6%, Voc=1055 mV, Jsc=21.73 mAcm-2). Figures 32c and Figure 32d show the IPCE and the PCE histogram of the devices with X2 and X25.

To further understand the influence of the molecular structures on their photovoltaic performance, PL and transient photovoltage/photocurrent decay measurements were performed as shown in Figure 33. The PL signals of the devices with X2 and X25 shown in Figure 33a demonstrate that for X2 the hole transfer from the perovskite to the HTM is slightly more efficient. This is probably due to the less positive Eredox-1 for X2 than X25, corresponding to a larger driving force for the interfacial charge transfer.

HTM λabs,

max

nm

λemission,

max

nm

Eredox-1

[V vs SHE]

E0-0

[eV]

EHOMOa

[eV]

ELUMOb

[eV]

Conductivity

S cm-1

X2 370 430 0.69 3.05 -5.09 -2.04 1.3×10-4

X25 335 420 0.76 3.20 -5.16 -1.96 2.8×10-4

a EHOMO is calculated by EHOMO=-4.4 eV-Eredox-135

b ELUMO value calculated by ELUMO=EHOMO+E0-0

64

Figure 32. (a) and (b) J-V curves of the PSCs with HTM X2 and X25. (c) and (d) IPCE curves of the PSCs with X2 and X25. Inset diagram: efficiency statistics.

Figure 33. (a) PL signal of the perovskite film with and without HTMs. (b) Transi-ent photovoltage decay. (c) Transient photocurrent decay. (d) Electron lifetime of the PSCs with HTM X2 and X25.

400 500 600 700 8000

20

40

60

80

100

IPC

E /

%

Wavelength /nm

X2

400 500 600 700 8000

20

40

60

80

100

IPC

E /

%

Wavelength / nm

X25

0.0 0.2 0.4 0.6 0.8 1.0

-0.005

0.000

0.005

0.010

0.015

0.020

0.025

Cu

rre

nt

de

ns

ity

/ A

cm

-2

Voltage / V

Efficiency: 14.65%V

oc: 1055 mV

Jsc

: 21.73 mAcm-2

FF: 0.64

5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 200

1

2

3

4

5

Co

un

ts

PCE / %

X2

0.0 0.2 0.4 0.6 0.8 1.0 1.2-0.005

0.000

0.005

0.010

0.015

0.020

0.025

Efficiency: 17.4%V

oc: 1100 mV

Jsc

: 22.54 mAcm-2

FF: 0.70

Cu

rren

t d

ens

ity

/ Ac

m-2

Voltage / V

6 8 10 12 14 16 18 200

1

2

3

4

5

Co

un

ts

PCE / %

X25

600 650 700 750 800 850

0.0

5.0x106

1.0x107

1.5x107

2.0x107

2.5x107

3.0x107

3.5x107

PL

inte

ns

ity

Wavelength / nm

perovskite perovskite/X25 perovskite/X2

0.00 0.02 0.04 0.06 0.08 0.10

0.0

0.2

0.4

0.6

0.8

1.0

No

rmal

ize

d c

urr

en

t

Time / s

X25 X2

0.000 0.001 0.002 0.003 0.004 0.005

0.0

0.2

0.4

0.6

0.8

1.0

No

rmal

ize

d c

urr

ent

Time / s

1E-4 1E-3 0.011E-5

1E-4

1E-3

0.01

0.1

1

Ph

oto

cu

rre

nt

de

ca

y t

ime

/ s

Jsc / A

X25 X2

0.00 0.02 0.04 0.06 0.08 0.100.0

0.2

0.4

0.6

0.8

1.0

No

rma

lized

ph

oto

volt

age

Time / s

X2 X25

0.000 0.001 0.002 0.003 0.004 0.0050.0

0.2

0.4

0.6

0.8

1.0

No

rmal

ized

ph

oto

volt

age

Time / s

X2 X25

65

From the photovoltage decay measurements as shown in Figure 33b it is deduced that the PSCs with X2 and X25 showed very similar decay kinetics. PSCs with X2 and X25 can be demonstrated to have similar electron recom-bination kinetics in the devices. Therefore, the higher Voc for X25 than X2 is mainly attributed to the lower Eredox-1 of X25 compared to that of X2. The transient photocurrent decay measurement at 100 mWcm-2 (see Figure 33c) and the derived photocurrent decay time at different light intensities (see Figure 33d) showed a clear difference in the charge transport behavior for X2 and X25. Specifically, X25-based PSCs had a much faster charge transport ability compared to that for X2-based PSCs. This could be mainly due to the higher conductivity of X25 than that of X2, as shown in Table 7. The faster hole transport in X25 leads to a higher Jsc compared to that for X2.

In addition, two Spiro-OMeTAD analogues, H11 and H12, were devel-oped. Their chemical structures and the synthetic procedures can be found below. 156

Figure 34. Synthetic process of H11 and H12.

Figure 34 shows two-step procedures for preparing H11 and H12 with a higher yield (>75%). The simple synthetic steps could make them easily available for the future large scale application. An additional advantage is that the solu-bility of H11 and H12 is higher than Spiro-OMeTAD in chlorobenzene, which is a common solvent used for preparation of the HTM solution.

The PSCs with H11 and H12 were tested in comparison with Spiro-OMeTAD. The photovoltaic parameters of the PSCs with H11, H12 and Spiro-OMeTAD are shown in Table 8. Compared with Spiro-OMeTAD-based devices (PCE=18.9%), the H11-based devices showed a slightly im-proved performance with a PCE of 19.8%, but the PSCs with H11 exhibited a lower PCE of 16.6%. The main difference for these three HTMs lies in the different Jsc although a slightly higher Voc is obtained for H11-based PSCs compared the other two HTMs. The difference in Jsc could be attributed to the hole conductivity variation as well as to the difference in the interfacial hole-transfer kinetics between the perovskite and the HTMs, as discussed in

O

Br Br

BrBr

NH

H3CO OCH3

Pd(dba)2, P(t-Bu)3, NaOt-Bu

O

N N

NN

OCH3

H3CO

H3CO

OCH3

OCH3

OCH3H3CO

H3CO

1

2

Lawesson’s reagent

NN OCH3

OCH3H3CO

H3CO

NN OCH3

OCH3H3CO

H3CO

BuLi, CoCl2

N NH3CO OCH3

H3CO OCH3

N NH3CO OCH3

H3CO OCH3

H11

H12

66

the previous case. Therefore, transient photocurrent decay and PL experi-ments were carried out to study these two processes, as shown in Figure 35.

Table 8. Photovoltaic parameters of the PSCs with different HTMs.

Transient photocurrent decay measurements for H11 and H12 are depicted in Figure 35a. For H11 the photocurrent decay is somehow faster than Spiro-OMeTAD and much faster than H12, demonstrating faster charge transport ability in the order H12 < Spiro-OMeTAD < H11. Besides, from the PL measurements, we see that H11- and H12-based PSCs showed more efficient hole transfer at the interface, which could be attributed to the good electronic contact between the HTM and the perovskite layer. Therefore, the fast charge transport in the HTM as well as the efficient hole transfer at the HTM-perovskite interface for H11 contribute to a higher Jsc compared to Spiro-OMeTAD and H12.

Figure 35. (a) Transient photocurrent decay and (b) PL measurement of the PSCs with different HTMs.

0.000 0.002 0.004 0.006 0.008 0.0100.0

0.2

0.4

0.6

0.8

1.0

No

rma

lize

d J

sc

Time / s

spiro-OMeTAD H11 H12

0.0 0.10.0

0.2

0.4

No

rma

lize

d J

sc

Time / s

spiro-OMeTAD H11 H12

550 600 650 700 750 800 850

0.0

5.0x106

1.0x107

1.5x107

2.0x107

2.5x107

3.0x107

3.5x107

PL

in

ten

sity

Wavelength / nm

No HTM H11 H12 Spiro-OMeTAD

600 650 700 750 800 8500.0

5.0x104

1.0x105

1.5x105

2.0x105

PL

in

ten

sit

y

Wavelength / nm

H11 H12 Spiro-OMeTAD

(a)

(b) (c)

PCE / % Voc / mV Jsc / mAcm-2 FF Spiro 18.9 1120 23.3 0.72 H11 19.8 1150 24.2 0.71 H12 16.6 1068 22.7 0.69

67

Therefore, we conclude that, in relation to good photovoltaic perfor-mance, several important parameters for HTMs need to be taken into ac-count when designing new and efficient HTMs for PSCs, such as (a) the Eredox-1 level of the HTM, appropriately positioned with respect to the va-lence band of perovskite, for optimizing the interface hole transfer and for obtaining higher Voc; (b) the solubility of the HTM, crucial for forming a uniform HTM capping layer; and (c) the hole conductivity, important for efficient charge extraction from the photoexcited perovskite to the counter electrode.

68

5. Conclusions and Outlook

Solid-state dye-sensitized solar cells (ssDSSCs) and perovskite solar cells (PSCs) have been developed as a newer generation of photovoltaics (PVs) due to their advantages of low cost and simple fabrication. However, one of the main limitations for the relatively low power conversion efficiency (PCE) of ssDSSCs is the poor pore filling caused by the spin-coating method which is usually employed for the deposition of the molecular hole transport material (HTM). The low pore infiltration by the HTM can affect the dye regeneration, the conductivity of the HTM and thus the photocurrent in the ssDSSCs. In contrast, the thin-film type perovskite solar cells do not have the pore-filling issue for the HTM but it is necessary to deposit the HTM layer on top of the perovskite capping layer in order to efficiently extract the charges at the interfaces and also to protect the perovskite film from the hu-midity in the atmosphere.

In ssDSSCs, in order to improve the pore filling, we have employed in-situ photoelectrochemical polymerization (PEP) to synthesize the conducting polymer HTM into the mesoporous TiO2. Compared to small molecule HTM, the prepared polymer HTMs have high conductivities resulting from the electrochemically doping during the PEP process. There are several im-portant factors involved to the efficient operation of the PEP process, such as the light source, dyes, precursors, TiO2 underlayer and electrolytic solution. In the previous publications, the organic solvent acetonitrile was always used in PEP because the precursors have good solubility in it. Firstly, we have for the first time explored the water-based electrolyte solution for the PEP in consideration of the fact that water-based solution is low-cost and environ-mentally friendly. By performing the aqueous PEP of PEDOT with the or-ganic dye LEG4, a power conversion efficiency (PCE) of 5.2% was achieved for aqueous PEP comparable to a PCE of 5.6% for organic PEP with the same dye. As regards the essential photovoltaic parameters, a slightly higher short-circuit current (Jsc) in conjunction to a low open-circuit voltage (Voc) was observed for the ssDSSCs in the case of aqueous PEP compared to that of organic PEP. We have systematically investigated the origins of the dif-ferences in Jsc and Voc. By measuring the electron lifetime and the dye regen-eration yield, we found that faster dye regeneration as well as faster interfa-cial charge recombination occurred in the device prepared with aqueous PEP compared to that prepared with organic PEP. These results could consistent-ly explain the differences in Jsc and Voc as mentioned above. From the differ-

69

ent NIR absorption spectra of the PEDOT obtained from organic and aque-ous PEP, we suggest that shorter polymer chains with less delocalized elec-tronic structure were formed in aqueous PEP compared to these in organic PEP. The shorter polymer chains from aqueous PEP could provide a larger contact area between the dyes and PEDOT while on the one hand facilitating an efficient dye regeneration process; on the other hand it will accelerate interfacial charge recombination. Therefore, we believe that further optimi-zation of aqueous PEP is interesting to tune the polymer structures and thus enhance the PCE of ssDSSCs.

The precursor bis-EDOT is mostly used and studied in the PEP because it has a relatively low oxidation potential of 0.9V/SHE in organic solution so that PEP can be initiated by hole injection from oxidized dye into the precur-sor. But it is expensive and difficult to synthesize. In contrast, monomer EDOT, less costly and of simpler structure has not been successfully used in PEP because EDOT has a very high oxidation potential of 1.6V/SHE in or-ganic solution. The higher oxidation potential of the precursor than that of the dye molecule makes the PEP process inefficient. Inspired by the finding that the oxidation potential of the precursor bis-EDOT is significantly lower in aqueous solution, we have for the first time employed EDOT as precursor in the PEP. The oxidation potential of EDOT was 0.9V/SHE in aqueous solution, which makes it possible to conduct efficient PEP. By effective combination of the aqueous PEP of EDOT and organic dye D35, a promising PCE of 3.0% was achieved in ssDSSCs compared to a low PCE of <0.1 for the case of organic PEP. Therefore, we concluded that the matching of the oxidation potentials of the dye and the precursor plays indeed an important role in the operation of PEP. Moreover, one of EDOT analogues, EDOP, was used as alternative precursor for PEP towards generating PEDOP, a polymer analogous to PEDOT with –NH replacing –S in the polymer chain. A PCE of 4.3% was obtained for organic dye D35-based ssDSSCs compared to a low PCE of 0.5% for the device with Ru-Z907 dye. The detailed charac-terization demonstrated that the device with metal-free organic dyes signifi-cantly retard the interfacial charge recombination and promote a high Voc. The high light harvesting ability of organic dye also contributed to a high Jsc. Following this work, we have systematically studied the effects of the dye molecules on the photovoltaic performance by comparing a series of dye molecules. The highest PCE of 7.1% was obtained by using the organic dye LEG4 with PEDOT. From the detailed characterization of the devices based on different dyes, we concluded that (a) the bulky structure of the dye mole-cule is crucial for retarding the electron recombination with the HTM, and achieving high Voc value; (b) the redox potential of the dye should be more positive than that of the precursor in order to conduct the PEP efficiently; (c) with respect to the light absorption the extinction coefficient of the dye mol-ecule can affect the kinetics of the PEP and also the Jsc of the devices.

70

For the future work, the interfacial functionalization of the TiO2/dye/HTM interfaces is interesting in order to enhance the charge col-lection efficiency of ssDSSCs by retarding the electron recombination. Be-sides, developing new dyes with high extinction coefficient and efficient recombination-blocking effect are needed to further improve the PCE of ssDSSCs.

For perovskite solar cells (PSCs), the operation of which is similar in sev-eral aspects to that of ssDSSCs, developing low-cost and efficient small molecule HTM to replace the state-of-the-art Spiro-OMeTAD is highly re-quired. In this respect, we have characterized a series of recently designed triphenylamine-based HTMs for PSCs in order to fundamentally investigate the effects of the alkyl chains, energy levels, and molecular morphology on their photovoltaic performance. Firstly, by grafting alkyl chains with differ-ent length at different positions on the triphenylamine core, we found that (a) the solubility of the HTM is enhanced by grafting alkyl chains, which pro-moted a high uniformity of the HTM capping layer; (b) the alkyl chains showed a large effect on the electronic contact with the perovskite layer, demonstrated by the different photoluminescence properties. In addition, we studied two newly developed similar molecules, X2 and X25, with minor structure difference but different energy levels. It was found that the relative lower HOMO level in X25 than X2 could provide a higher Voc in PSCs. Fur-thermore, we studied the influences of the molecular geometry of the HTM by changing the central Spiro connection in Spiro-OMeTAD with carbon-carbon single bond (H11) and carbon-carbon double bond (H12). By com-bining the mixed-ion perovskite with the HTM H11, a high PCE of 19.8% was obtained compared to a PCE of 18.9% for the standard device with Spi-ro-OMeTAD. Besides, the synthesis of H11 is simpler as compared to that of Spiro-OMeTAD. Therefore, we believe that H11 holds promise to be an alternative HTM for the future application of the PSCs. The high perfor-mance for H11 is mainly attributed to a relatively low HOMO level, high uniformity of the capping layer, and high conductivity. For the HTMs we could therefore conclude that it is important for the HTM to have good solu-bility, high uniformity of the capping layer, high conductivity, and appropri-ate energy levels in order to ensure efficient operation in PSCs.

71

Sammanfattning på svenska

Solenergi Tillgången på energi är en av de viktigaste frågorna för framtiden. De si-nande tillgångarna på olja och ökade föroreningar från fossila bränslen har bidragit till problem såsom klimatförändringar, energikris och medföljande säkerhetsproblem. Därför är förnyelsebara energikällor viktiga för att möta energibehovet och stävja miljöproblemen. Bland förnyelsebara energikällor är solenergi den som betraktas som den med störst kapacitet, om man ser till mängden solljus som når jorden. Apparater som kan omvandla solljuset till elektricitet kallas solceller. Olika typer av solceller har utvecklats och den så kallade första generationens solceller är baserade på kisel som det aktiva materialet som absorberar solljuset. Kiselsolceller har en verkningsgrad på över 25% för omvandlingen av solljus till elektrisk effekt, och är stabila för en lång tid. Kiselsolceller dominerar idag solcellsmarknaden och ca 90% av solcellerna på marknaden är gjorda av kisel. Tyvärr är kostnaden för att till-verka kiselsolcellerna ganska hög, vilket gör det svårt för kiselsolceller att konkurera med andra energikällor. Därför behövs utveckling av nya typer av solceller med lägre material- och tillverkningskostnader. I den här avhand-lingen fokuseras på två typer av solceller med lägre kostnader, solida färg-ämnessolceller och perovskit solceller.

Solida färgämnessolceller och perovskitsolceller Solida färgämnessolceller (ssDSSC) och perovskitsolceller (PSC) har lik-nande uppbyggnad och funktion. Kortfattat så absorberas ljus i ett lager med färgämnesmolekyler i ena fallet och i ett kristallint perovskitlager i andra fallet. Då ljuset absorberas exciteras en elektron och ett exciterat tillstånd uppstår. Från detta exciterade tillstånd dissocierar en negativ laddning (elektron) och en positiv laddning (hål). Materialen på båda sidor om färg-ämnesmolekylerna eller perovskiten transporterar antingen elektroner eller hål från färgämnet och perovskiten. Dessa material på de olika sidorna kallas därför elektrontransportmaterial (ETM) eller håltransportmaterial (HTM) beroende på vilken typ av laddning de transporterar. Kontakten vid gränsy-torna mellan de olika materialen är väsentlig för laddningssepareringen av elektroner och hål och för insamlandet av dessa laddningar till kontakterna

72

och därför också för verkningsgraden hos solcellen. Den mest effektiva HTM för ssDSSC och PSC är en molekyl som kallas Spiro-OMeTAD, se Figur 1 nedan. De mest effektiva ssDSSC och PSC med Spiro-OMeTAD har en verkningsgrad på >7% respektive >22%. Tyvärr är Spiro-OMeTAD svår att tillverka och kostnaden för molekylen är ungefär samma som för guld. Den höga kostnaden för Spiro-OMeTAD är därför en stor begränsning för storskalig produktion av ssDSSC och PSC.

Figur 1. Kemisk struktur för Spiro-OMeTAD.

Håltransportmaterial (HTM) I den här avhandlingen är huvudfokus att utveckla nya HTM för att ersätta Spiro-OMeTAD. Dessutom försöker vi hitta de fundamentala kraven för HTM i solceller.

I ssDSSC används vanligtvis spin-coating för att deponera HTM på sol-cellerna, vilket ger en ganska dålig fyllning av porerna i solcellen. Den då-liga fyllningen av porerna resulterar i (a) dålig laddningsöverföring från fär-gämnet till HTM, (b) en begränsning av tjockleken av det aktiva färg-ämne/TiO2 lagret, och därmed en begränsning av ljusabsorptionen i solcell-en. På grund av detta har vi utvecklat in-situ fotoelektropolymerisering (PEP) för att syntetisera polymera HTM. Mekanismen för PEP är beskriven i Figur 2. Då PEP används kan den ledande polymeren som används som HTM genereras som ett jämt lager inuti en tjock färgämne/TiO2 elektrod.

N N

NN

O O

O

O

OO

O

O

Spiro-OMeTAD

73

Figur 2. Fotoelektropolymerisering (PEP) med bis-EDOT som monomer.

I PEP finns flera komponenter som är kritiska för bildandet av den ledande polymeren, såsom: monomeren, elektrolytlösningen och färgämnesmoleky-lerna. Därför har vi undersökt hur dessa olika delar påverkar PEP och solcel-lerna gjorda med denna metod. Till exempel har vi för första gången använt en vattenbaserad elektrolyt istället för det vanliga lösningsmedlet acetonitril. Med den vattenbaserade metoden fick vi solceller med en verkningsgrad på 5,2% för omvandling av solljus till elektricitet, och med metoden baserad på acetonitril fick vi 5,6% verkningsgrad. Vi kunde också se att lösningsmedlet påverkade egenskaperna hos den polymer (PEDOT) som bildades. För sol-cellen gjord med den vattenbaserade metoden kunde vi se att hålen snabbt transporterades till polymeren, men också att hålen snabbt rekombinerade med elektronerna igen. Därför behövs vidare optimering av PEP i vatten för att öka verkningsgraden hos solcellerna.

Utöver detta gjordes också experiment med en annan monomer (EDOT) som startmaterial. För EDOT som startmaterial och med en vattenbaserad elektrolyt erhölls en verkningsgrad på 3% för solcellerna, vilket jämfördes med EDOT i acetonitril, vilket gav en verkningsgrad på <0,1%. Detta förkla-rade vi med att EDOT har en annan oxidationspotential i vatten jämfört med i acetonitril så att PEP är möjlig i vatten men inte i acetonitril. En annan monomer, EDOP, användes också med PEP vilket gav solceller med 4% verkningsgrad tillsammans med organiska färgämnen men <0,5% tillsam-mans med ruteniumbaserade färgämnen. Den stora skillnaden i effektivitet för solcellerna med organiska- respektive ruteniumfärgämnen förklarades av skillnader i rekombinationshastighet mellan de fotogenererade elektronerna och hålen. För de organiska färgämnena var rekombinationshastigheten be-tydligt lägre, varvid en högre effektivitet kunde erhållas.

Eftersom en så stor effekt kunde ses för olika färgämnen studerade vi närmare hur en serie av olika organiska och ruteniumfärgämnen påverkade PEP och solcellernas effektivitet. Från karakteriseringen och jämförelse mel-lan de olika systemen kunde vi fastslå att (a) strukturen hos färgämnet är viktig för att blockera rekombination av de fotogenererade laddningarna, (b)

74

energiskillnaden (skillnaden i oxidationspotential) mellan färgämnet och monomeren bestämmer hastigheten på polymeriseringen i PEP, (c) ljus-absorptionen hos färgämnena har stor inverkan på PEP processen och verk-ningsgraden hos solcellerna. Dessa faktorer måste därför beaktas när nya färgämnen eller monomerer utvecklas.

Perovskit baserade solceller (PSC) är oftast tunnare och porerna är fyllda med perovskit och behöver inte fyllas av HTM, vilket medför mindre pro-blem med molekylära HTM. Men pga. den höga kostnaden för Spiro-OMeTAD, är det ändå väsentligt att försöka hitta andra molekyler som HTM som är effektiva, men kan tillverkas till låg kostnad. Vi har designat en serie med triarylamin-baserade HTM för PSC för att förstå hur strukturen för en HTM påverkar effektiviteten för PSC. Genom att jämföra olika molekylära HTM fann vi att (a) lösligheten för HTM är viktig vid bildandet av ett jämnt HTM lager på solcellen, vilket ger en mer effektiv och stabil solcell, (b) passande energinivåer hos HTM är viktigt för att nå en hög spänning hos solcellen och samtidigt en effektiv laddningsöverföring.

0.0 0.2 0.4 0.6 0.8 1.0 1.2-0.005

0.000

0.005

0.010

0.015

0.020

0.025

HTM H11 PCE / % 19.8Voc / mV 1150Jsc / mAcm-2 24.2 FF 0.71

J / A

cm-2

Voltage / V

H11

Figur 3. Ström och spännings-mätning för en perovskitsolcell med HTM H11.

Ett av HTM (H11) kunde i en perovskitsolcell ge en hög verkningsgrad 19,8%, vilket var högre än en likadan perovskitsolcell med Spiro-OMeTAD (18,9%). H11 var dessutom ganska enkel att syntetisera och är därför ett lovande alternativ för en HTM med låg kostnad som samtidigt ger en hög effektivitet i en perovskitsolcell.

75

Acknowledgement

I would like to thank all my supervisors: Professor Anders Hagfeldt, thank you very much for giving me the great opportunity to be a PhD student in Uppsala University, guiding me the direction of the research projects, always providing kind motivation on the research projects I like to explore. Dr. Nick Vlachopoulos, thank you very much for your kind support for being a very good advisor during my PhD study. You introduced me into professional research, led me to conduct interesting experiments, provided fruitful scien-tific discussions and suggestions, supervised me on the scientific writing, and gave me constructive suggestions on my personal development. I am very thankful to you for sharing your broad experience and always answer-ing my questions with great patience. Dr. Erik M. J. Johansson, thank you very much for the kind supervision on the projects, nice discussion on the scientific questions, and great help on the revision and submission of the manuscripts. Thank you for the great kindness and helpfulness. Dr. Gerrit Boschloo, thank you for sharing your knowledge and ideas in my projects, helping me to solve the scientific questions, and organizing the meaningful scientific discussions in regular group meeting.

I would also like to thank all my collaborators for doing wonderful re-search together: Professor Mohamed Jouini and his University of Paris 7 (Diderot) group members, Associate professor Christian Perruchot, Dr. Adel Jarboui and Annette Delices, for the nice collaboration on the PEP project where you all provided a great contribution; Dr. Lei Yang, Dr. Byung-wook Park, Yang Shen and Niklas Wahlström of our laboratory for the initial con-tribution on the PEP project; Dr. Bo Xu, Dr. Yong Hua, Professor Licheng Sun of KTH, Organic Chemistry division, for providing me new molecules to investigate in my projects; Dr. Peng Liu and Professor Lars Kloo of KTH, Applied Physical Chemistry division, for nice discussion on the DFT calcu-lations and careful revision on manuscript; Dr. Malin B. Johansson and Dr. Xiaoliang Zhang of our laboratory for the help on SEM measurements; Dr. Palas Baran Pati and Professor Haining Tian from the Department of Chem-istry for providing new dyes for study in my project; Dr. Meysam Pazoki of our laboratory, Dr. Justus Simiyu, visiting scientist from the University of Nairobi, and Dr. Ocean Cheung from the Department of Engineering Sci-ences of Uppsala University for developing and measuring the new TiO2 nanoparticles; Mahboubeh Hadadian and Dr. Juan Pablo Correa Baena in the Swiss Federal Institute of Technology (EPFL) as well as Dr. Kerttu Aitola,

76

Kári Sveinbjörnsson from our laboratory for the collaboration in the perov-skite project; Dr. Valentina Leandri and Professor Sascha Ott from the Chemistry Department of Uppsala University for nice collaboration on new dye molecule; Dr. Luca D' Amario and Lei Zhang from physical chemistry division of Uppsala University for the help on measuring laser spectroscopy. Li Yang and Changqing Ruan from Nanotechnology division of Uppsala University for TGA measurements; Dr. Kerttu Aitola of our laboratory for the collaboration on testing carbon nanotubes; Dr. Marina Freitag of our laboratory for exploring interesting ideas; Dr. Denys Shevchenko from the Chemistry Department of Uppsala University and Dr. Hanna Ellis of our laboratory for the collaboration on the mass spectrometric characterization; Dr. Yan Hao of our laboratory and Dr. Thomas W. Holcombe and Professor Michael Grätzel of EPFL for collaboration on blue dye application; Profes-sor Mohammad Khaja Nazeeruddin of EPFL for the nice collaboration on new dyes. Dr. Esmaeil Sheibani of KTH, Organic Chemistry division, for doing synthesis of new precursors; Miss Alba Mingorance and Professor Monica Lira-Cantu of the Catalan Institute of Nanoscience and Nanotech-nology for your kind hosting and nice research collaboration during my re-search visit at your laboratory; Dr. Adam Hultqvist and Professor Marika Edoff from solid-state electronics division of Uppsala University for the research collaboration on making ALD layers for solar cell application.

I would like to thank Dr. Leif Häggman, our “boss” in the laboratory. You helped to create an organized lab environment for us. You are always helpful to me on the lab issues. I also thank all former and current group members for having a great pleasure, joy and fun together, sharing great experience and knowledge, helping me in all respects.

I want to thank the China Scholarship Council for financial support for my PhD study. I acknowledge the following sources for research travel grants: Anna Maria Lundin foundation, French-Swedish Research Mobility Program “Gustaf Dalén”, Wallenberg Foundation, Liljewalchs Scholarship Fund, Graduate SAM21 Travel Grant, ÅFORSK Foundation, STSM-COST Ac-tion.

Finally I would like to appreciate my parents, my sister, my wife Li Yang and all other family members for their love, support, advice, encouragement and understanding.

77

References

1. J. M. Essick and R. T. Mather, Am. J. Phys., 1993, 61, 646-649. 2. F. J. Haug and C. Ballif, Energ. Environ. Sci., 2015, 8, 824-837. 3. Y. Lee, C. Park, N. Balaji, Y. J. Lee and V. A. Dao, Isr. J. Chem., 2015, 55,

1050-1063. 4. B. M. Basola and B. McCandless, J. Photon. Energy, 2014, 4, 040996. 5. T. Wada, Y. Hashimoto, S. Nishiwaki, T. Satoh, S. Hayashi, T. Negami and H.

Miyake, Sol . Energ. Mat. Sol. C, 2001, 67, 305-310. 6. M. A. Green, J. Mater. Sci-Mater. El., 2007, 18, S15-S19. 7. O. A. Abdulrazzaq, V. Saini, S. Bourdo, E. Dervishi and A. S. Biris, Particul.

Sci. Technol., 2013, 31, 427-442. 8. X. Lin, J. Seok, S. Yoon, T. Kim, B. Kim and K. Kim, Synthetic Met., 2014,

196, 145-150. 9. Y. Tamai, H. Ohkita, H. Benten and S. Ito, J. Phys. Chem. Lett., 2015, 6, 3417-

3428. 10. G. L. Schulz, M. Lobert, I. Ata, M. Urdanpilleta, M. Linden, A. Mishra and P.

Bauerle, J. Mater. Chem. A, 2015, 3, 13738-13748. 11. N. Chaturvedi, S. K. Swami and V. Dutta, Thin Solid Films, 2016, 598, 82-87. 12. I. Etxebarria, J. Ajuria and R. Pacios, J. Photon Energy, 2015, 5. 057214 13. W. Zhao, D. Qian, S. Zhang, S. Li, O. Inganas, F. Gao and J. Hou, Adv. Mater.,

2016, DOI: 10.1002/adma.201600281. 14. J. L. Duan, H. H. Zhang, Q. W. Tang, B. L. He and L. M. Yu, J. Mater. Chem.

A, 2015, 3, 17497-17510. 15. M. R. Kim and D. L. Ma, J. Phys. Chem. Lett., 2015, 6, 85-99. 16. J. Du, Z. Du, J. S. Hu, Z. Pan, Q. Shen, J. Sun, D. Long, H. Dong, L. Sun, X.

Zhong and L. J. Wan, J. Am. Chem. Soc., 2016, DOI: 10.1021/jacs.6b00615. 17. L. Li, X. C. Yang, J. J. Gao, H. N. Tian, J. Z. Zhao, A. Hagfeldt and L. C. Sun,

J. Am. Chem. Soc., 2011, 133, 8458-8460. 18. J. Qian, Q. S. Liu, G. Li, K. J. Jiang, L. M. Yang and Y. L. Song, Chem. Com-

mun., 2011, 47, 6461-6463. 19. W. L. Feng, Y. Li, J. Du, W. Wang and X. H. Zhong, J. Mater. Chem. A, 2016,

4, 1461-1468. 20. G. H. Carey, L. Levina, R. Comin, O. Voznyy and E. H. Sargent, Adv. Mater.,

2015, 27, 3325-3330. 21. N. Vlachopoulos, P. Liska, J. Augustynski and M. Gratzel, J. Am. Chem. Soc.,

1988, 110, 1216-1220. 22. B. Oregan and M. Gratzel, Nature, 1991, 353, 737-740. 23. B. E. Hardin, H. J. Snaith and M. D. McGehee, Nat. Photonics, 2012, 6, 162-

169. 24. A. Hagfeldt, G. Boschloo, L. C. Sun, L. Kloo and H. Pettersson, Chem. Rev.,

2010, 110, 6595-6663. 25. S. M. Feldt, E. A. Gibson, E. Gabrielsson, L. Sun, G. Boschloo and A. Hag-

feldt, J. Am. Chem. Soc., 2010, 132, 16714-16724.

78

26. S. M. Feldt, G. Wang, G. Boschloo and A. Hagfeldt, J. Phys. Chem. C, 2011, 115, 21500-21507.

27. M. K. Wang, N. Chamberland, L. Breau, J. E. Moser, R. Humphry-Baker, B. Marsan, S. M. Zakeeruddin and M. Gratzel, Nat. Chem., 2010, 2, 385-389.

28. W. L. Hoffeditz, M. J. Katz, P. Deria, G. E. Cutsail Iii, M. J. Pellin, O. K. Far-ha and J. T. Hupp, J. Phys. Chem. C, 2016, 120, 3731-3740.

29. P. Chawla and M. Tripathi, International Journal of Energy Research, 2015, DOI: 10.1002/er.3366.

30. V. Sugathan, E. John and K. Sudhakar, Renewable and Sustainable Energy Reviews, 2015, 52, 54-64.

31. K. Kakiage, Y. Aoyama, T. Yano, K. Oya, J. Fujisawa and M. Hanaya, Chem. Commun., 2015, 51, 15894-15897.

32. S. Mathew, A. Yella, P. Gao, R. Humphry-Baker, B. F. E. Curchod, N. Ashari-Astani, I. Tavernelli, U. Rothlisberger, M. K. Nazeeruddin and M. Gratzel, Nat. Chem., 2014, 6, 242-247.

33. Z. Yao, M. Zhang, H. Wu, L. Yang, R. Li and P. Wang, J. Am. Chem. Soc., 2015, 137, 3799-3802.

34. K. Kakiage, Y. Aoyama, T. Yano, T. Otsuka, T. Kyomen, M. Unno and M. Hanaya, Chem. Commun., 2014, 50, 6379-6381.

35. N. Heo, Y. Jun and J. H. Park, Sci. Rep., 2013, 3, 1712. 36. H. Ellis, V. Leandri, A. Hagfeldt, G. Boschloo, J. Bergquist and D. Shevchen-

ko, J. Mass Spectrom., 2015, 50, 734-739. 37. U. Bach, D. Lupo, P. Comte, J. E. Moser, F. Weissortel, J. Salbeck, H. Spreit-

zer and M. Gratzel, Nature, 1998, 395, 583-585. 38. M. Gratzel, Mrs Bulletin, 2005, 30, 23-27. 39. A. Kojima, K. Teshima, Y. Shirai and T. Miyasaka, J. Am. Chem. Soc., 2009,

131, 6050-6051. 40. M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami and H. J. Snaith, Sci-

ence, 2012, 338, 643-647. 41. H. S. Kim, C. R. Lee, J. H. Im, K. B. Lee, T. Moehl, A. Marchioro, S. J. Moon,

R. Humphry-Baker, J. H. Yum, J. E. Moser, M. Gratzel and N. G. Park, Sci. Rep., 2012, 2, 591.

42. T. B. Song, Q. Chen, H. P. Zhou, C. Y. Jiang, H. H. Wang, Y. Yang, Y. S. Liu, J. B. You and Y. Yang, J. Mater. Chem. A, 2015, 3, 9032-9050.

43. Y. C. Hsiao, T. Wu, M. X. Li, Q. Liu, W. Qin and B. Hu, J. Mater. Chem. A, 2015, 3, 15372-15385.

44. G. Hodes, Science, 2013, 342, 317-318. 45. M. Sessolo and H. J. Bolink, Science, 2015, 350, 917-917. 46. J. Burschka, N. Pellet, S. J. Moon, R. Humphry-Baker, P. Gao, M. K. Nazee-

ruddin and M. Gratzel, Nature, 2013, 499, 316-319. 47. Y. X. Zhao and K. Zhu, J. Mater. Chem. A, 2015, 3, 9086-9091. 48. F. M. Li, C. X. Bao, H. Gao, W. D. Zhu, T. Yu, J. Yang, G. Fu, X. X. Zhou

and Z. G. Zou, Mater. Lett., 2015, 157, 38-41. 49. Q. Chen, H. P. Zhou, Z. R. Hong, S. Luo, H. S. Duan, H. H. Wang, Y. S. Liu,

G. Li and Y. Yang, J. Am. Chem. Soc., 2014, 136, 622-625. 50. W. Y. Nie, H. H. Tsai, R. Asadpour, J. C. Blancon, A. J. Neukirch, G. Gupta, J.

J. Crochet, M. Chhowalla, S. Tretiak, M. A. Alam, H. L. Wang and A. D. Mohite, Science, 2015, 347, 522-525.

51. N. J. Jeon, J. H. Noh, Y. C. Kim, W. S. Yang, S. Ryu and S. Il Seol, Nat. Ma-ter., 2014, 13, 897-903.

79

52. D. Bi, W. Tress, M. I. Dar, P. Gao, J. Luo, C. Renevier, K. Schenk, A. Abate, F. Giordano, J. Correa Baena, J. Decoppet, S. M. Zakeeruddin, M. K. Nazee-ruddin, M. Grätzel, A. Hagfeldt, Science Advance, 2016, DOI: 10.1126/sciadv.1501170

53. C. G. Wu, C. H. Chiang, Z. L. Tseng, M. K. Nazeeruddin, A. Hagfeldt and M. Gratzel, Energ. Environ. Sci., 2015, 8, 2725-2733.

54. J. P. C. Baena, L. Steier, W. Tress, M. Saliba, S. Neutzner, T. Matsui, F. Giordano, T. J. Jacobsson, A. R. S. Kandada, S. M. Zakeeruddin, A. Petrozza, A. Abate, M. K. Nazeeruddin, M. Gratzel and A. Hagfeldt, Energ. Environ. Sci., 2015, 8, 2928-2934.

55. Z. Yu and L. C. Sun, Adv. Energy Mater., 2015, 5, 1500213. 56. S. Ameen, M. A. Rub, S. A. Kosa, K. A. Alamry, M. S. Akhtar, H. S. Shin, H.

K. Seo, A. M. Asiri and M. K. Nazeeruddin, Chemsuschem, 2016, 9, 10-27. 57. G. R. Berdiyorov, F. El-Mellouhi, M. E. Madjet, F. H. Alharbi, F. M. Peeters

and S. Kais, Sol. Energ. Mat. Sol .C., 2016, 148, 2-10. 58. K. T. Butler, J. M. Frost and A. Walsh, Mater. Horiz., 2015, 2, 228-231. 59. D. D. Zhang, S. W. Eaton, Y. Yu, L. T. Dou and P. D. Yang, J. Am. Chem.

Soc., 2015, 137, 9230-9233. 60. G. Nedelcu, L. Protesescu, S. Yakunin, M. I. Bodnarchuk, M. J. Grotevent and

M. V. Kovalenko, Nano Lett., 2015, 15, 5635-5640. 61. T. Ding, R. F. Li, W. G. Kong, B. P. Zhang and H. Z. Wu, Appl. Surf. Sci.,

2015, 357, 1743-1746. 62. N. J. Jeon, J. H. Noh, W. S. Yang, Y. C. Kim, S. Ryu, J. Seo and S. I. Seok,

Nature, 2015, 517, 476-480. 63. K. Wang, Z. Q. Liang, X. Q. Wang and X. D. Cui, Adv. Electron Mater., 2015,

1, 1500089. 64. D. Wang, M. Wright, N. K. Elumalai and A. Uddin, Sol. Energ. Mat. Sol. C.,

2016, 147, 255-275. 65. Y. G. Rong, L. F. Liu, A. Y. Mei, X. Li and H. W. Han, Adv. Energy Mater.,

2015, 5, 1501066. 66. S. N. Habisreutinger, T. Leijtens, G. E. Eperon, S. D. Stranks, R. J. Nicholas

and H. J. Snaith, Nano Lett., 2014, 14, 5561-5568. 67. A. Y. Mei, X. Li, L. F. Liu, Z. L. Ku, T. F. Liu, Y. G. Rong, M. Xu, M. Hu, J.

Z. Chen, Y. Yang, M. Gratzel and H. W. Han, Science, 2014, 345, 295-298. 68. W. Chen, Y. Wu, Y. Yue, J. Liu, W. Zhang, X. Yang, H. Chen, E. Bi, I. Ash-

raful, M. Grätzel, L. Han, Science 2015, 350, 944. 69. C. T. Weisspfennig, D. J. Hollman, C. Menelaou, S. D. Stranks, H. J. Joyce, M.

B. Johnston, H. J. Snaith and L. M. Herz, Adv. Funct. Mater., 2014, 24, 668-677.

70. K. Murakoshi, R. Kogure, Y. Wada and S. Yanagida, Chem. Lett., 1997, DOI: Doi 10.1246/Cl.1997.471, 471-472.

71. J. Zhang, N. Vlachopoulos, Y. Hao, T. W. Holcombe, G. Boschloo, E. M. Johansson, M. Gratzel and A. Hagfeldt, Chemphyschem, 2016, 17, 1441-1445.

72. H. Yu, S. Zhang, H. Zhao, G. Will and P. Liu, Electrochim. Acta, 2009, 54, 1319-1324.

73. C. Jiang, M. Y. Leung, W. L. Koh and Y. Li, Thin Solid Films, 2011, 519, 7850-7854.

74. C. Jiang, W. L. Koh, M. Y. Leung, W. Hong, Y. Li and J. Zhang, J. Solid State Chem., 2013, 198, 197-202.

75. J. Li, X. Chen, W. Xu, C.-Y. Nam and Y. Shi, Thin Solid Films, 2013, 536, 275-279.

76. T. S. Su, T. Y. Hsieh, C. Y. Hong and T. C. Wei, Sci. Rep., 2015, 5, 16098.

80

77. L. Kavan, M. Zukalova, O. Vik and D. Havlicek, Chemphyschem, 2014, 15, 1056-1061.

78. P. Lellig, M. A. Niedermeier, M. Rawolle, M. Meister, F. Laquai, P. Muller-Buschbaum and J. S. Gutmann, Phys. Chem. Chem. Phys., 2012, 14, 1607-1613.

79. P. Lellig, M. Meister, J. W. Ochsmann, M. A. Niedermeier, M. Rawolle, F. Laquai, P. Muller-Buschbaum and J. S. Gutmann, Springerplus, 2015, 4, 502.

80. A. K. Chandiran, M. K. Nazeeruddin and M. Grätzel, Adv. Funct. Mater., 2014, 24, 1615-1623.

81. B. Peng, G. Jungmann, C. Jäger, D. Haarer, H.-W. Schmidt and M. Thelakkat, Coord. Chem. Rev., 2004, 248, 1479-1489.

82. W. Ke, G. Fang, J. Wan, H. Tao, Q. Liu, L. Xiong, P. Qin, J. Wang, H. Lei, G. Yang, M. Qin, X. Zhao and Y. Yan, Nat. Commun., 2015, 6, 6700.

83. L. Gouda, R. Gottesman, A. Ginsburg, D. A. Keller, E. Haltzi, J. Hu, S. Tirosh, A. Y. Anderson, A. Zaban and P. P. Boix, J. Phys. Chem. Lett., 2015, 6, 4640-4645.

84. L. Kavan, N. Tétreault, T. Moehl and M. Grätzel, J. Phys. Chem. C, 2014, 118, 16408-16418.

85. a. L. M. P. Petra J. Cameron, J. Phys. Chem. B, 2003, 107, 14394. 86. E. A. S. N. Kopidakis, N. G. Park, J. Van de Lagemaat, A. F. Frank, J. Phys.

Chem. B, 2000, 104, 3930-3936. 87. O. l. L. Dloczik, I. Lauermann, L. M. Peter, E. A. Ponomarev, G. Redmond, N.

J. Shaw, I. Uhlendorf, J. Phys. Chem. B, 1997, 101, 10281-10289. 88. L. Herz, J. Phys. Chem. C, 2010, 114, 1365. 89. J. Bouclé and J. Ackermann, Polymer International, 2012, 61, 355-373. 90. N. C. Nicolas Te´ treault, Endre Horvath, Peng Wang,Thomas Moehl, Shaik

M. Zakeeruddin, Je La´szlo Forro, remie Brillet, Rita Smajda, Arnaud Magrez, Stephane Bungener, and Michael Gratzel, ACS Nano, 2010, 4, 7644.

91. X. Sun, Q. Sun, Y. Li, L. Sui and L. Dong, Phys. Chem. Chem. Phys., 2013, 15, 18716-18720.

92. P. Chen, J. Brillet, H. Bala, P. Wang, S. M. Zakeeruddin and M. Grätzel, J. Mater. Chem., 2009, 19, 5325.

93. W. Zhang, R. Zhu, L. Ke, X. Liu, B. Liu and S. Ramakrishna, Small, 2010, 6, 2176-2182.

94. H. Melhem, P. Simon, L. Beouch, F. Goubard, M. Boucharef, C. Di Bin, Y. Leconte, B. Ratier, N. Herlin-Boime and J. Bouclé, Adv. Energy Mater., 2011, 1, 908-916.

95. G. Sadoughi, R. Mohammadpour, A. Irajizad, N. Taghavinia, S. Dadgostar, M. Samadpour and F. Tajabadi, Curr. Appl. Phys., 2013, 13, 371-376.

96. M. Pazoki, J. Oscarsson, L. Yang, B. W. Park, E. M. J. Johansson, H. Rensmo, A. Hagfeldt and G. Boschloo, RSC Adv., 2014, 4, 50295-50300.

97. S. H. Ahn, W. S. Chi, D. J. Kim, S. Y. Heo and J. H. Kim, Adv. Funct. Mater., 2013, 23, 3901-3908.

98. D. Hwang, S. M. Jo, D. Y. Kim, V. Armel, D. R. MacFarlane and S. Y. Jang, ACS Appl. Mater. Interfaces, 2011, 3, 1521-1527.

99. S. Colella, E. Orgiu, I. Bruder, A. Liscio, V. Palermo, B. Bruchmann, P. Samorì and P. Erk, Adv. Energy Mater., 2014, 4, 1301362.

100. L. Schmidt-Mende and J. L. MacManus-Driscoll, Materials Today, 2007, 10, 40-48.

101. C. Xu, J. Wu, U. V. Desai and D. Gao, Nano Lett., 2012, 12, 2420-2424. 102. L. Schlur, A. Carton, P. Lévêque, D. Guillon and G. Pourroy, J. Phys. Chem.

C, 2013, 117, 2993-3001.

81

103. M. Boucharef, C. Di Bin, M. S. Boumaza, M. Colas, H. J. Snaith, B. Ratier and J. Boucle, Nanotechnology, 2010, 21, 205203.

104. C. O. Richard L. Willis, Brian O' Regan, Thierry Lutz, Jenny Nelson, and James R. Durrant, J. Phys. Chem. B, 2002, 106, 7605.

105. K. Z. Song Rim Jang, Min Jae Ko, Kyungkon Kim, Chulhee Kim, Nam Gyu Park, and Arthur. J. Frank, ACS Nano, 2011, 10, 8267.

106. j. h. y. PETER CHEN, Fillippo De Angelis, Edoardo Mosconi, Simona Fantac-ci, Soo jin Moon, Robin Humphry Baker, Jaejung Ko, M. K. Nazeeruddin, and Michael Gratzel, Nano Lett., 2009, 9, 2487.

107. U. B. Lukas Schmidt Mende, Robin Humphry Baker, Tamotsu Horiuchi, Hide-toshi Miura, Seigo Ito, Satoshi Uchida and Michael Gratzel, Adv. Mater., 2003, 17, 813.

108. S. Tatay, S. A. Haque, B. O'Regan, J. R. Durrant, W. J. H. Verhees, J. M. Kroon, A. Vidal-Ferran, P. Gaviña and E. Palomares, J. Mater. Chem., 2007, 17, 3037-3044.

109. M. H. K. Ute B. Cappel, Neil G. Pschirer, Felix Eickemeyer, Jan Schoneboom, Peter Erk, Gerrit Boschloo, and Anders Hagfeldt, J. Phys. Chem. Lett., 2009, 113, 14595.

110. J. H. Y. Soo Jin Moon, Robin Humphry Baker, Karl Martin Karlsson, Daniel P. Hagberg, Tannia Marinado, Anders Hagfeldt, Licheng Sun, Michael Gratzel, Md. M. Nazeeruddin, J. Phys. Chem. C, 2009, 113, 16816.

111. M. Wang, S.-J. Moon, D. Zhou, F. Le Formal, N.-L. Cevey-Ha, R. Humphry-Baker, C. Grätzel, P. Wang, S. M. Zakeeruddin and M. Grätzel, Adv. Funct. Mater., 2010, 20, 1821-1826.

112. X. Jiang, K. M. Karlsson, E. Gabrielsson, E. M. J. Johansson, M. Quintana, M. Karlsson, L. Sun, G. Boschloo and A. Hagfeldt, Adv. Funct. Mater., 2011, 21, 2944-2952.

113. A. Dualeh, F. De Angelis, S. Fantacci, T. Moehl, C. Yi, F. Kessler, E. Bara-noff, M. K. Nazeeruddin and M. Grätzel, J. Phys. Chem. C, 2012, 116, 1572-1578.

114. W. H. Nguyen, C. D. Bailie, J. Burschka, T. Moehl, M. Grätzel, M. D. McGe-hee and A. Sellinger, Chem. Mater., 2013, 25, 1519-1525.

115. A. Abate, M. Planells, D. J. Hollman, S. D. Stranks, A. Petrozza, A. R. S. Kan-dada, Y. Vaynzof, S. K. Pathak, N. Robertson and H. J. Snaith, Adv. Energy Mater., 2014, 4, 1400146.

116. E. Gabrielsson, H. Ellis, S. Feldt, H. Tian, G. Boschloo, A. Hagfeldt and L. Sun, Adv. Energy Mater., 2013, 3, 1647-1656.

117. M. F. Xu, D. F. Zhou, N. Cai, J. Y. Liu, R. Z. Li and P. Wang, Energ. Environ. Sci., 2011, 4, 4735-4742.

118. D. P. Hagberg, X. Jiang, E. Gabrielsson, M. Linder, T. Marinado, T. Brinck, A. Hagfeldt and L. Sun, J. Mater. Chem., 2009, 19, 7232.

119. P. Liu, B. Xu, K. M. Karlsson, J. Zhang, N. Vlachopoulos, G. Boschloo, L. Sun and L. Kloo, J. Mater. Chem. A, 2015, 3, 4420-4427.

120. S. R. Jang, J. H. Yum, C. Klein, K. J. Kim, P. Wagner, D. Officer, M. Gratzel and M. K. Nazeeruddin, J. Phys. Chem. C, 2009, 113, 1998-2003.

121. C. Y. Chen, M. Wang, J. Y. Li, N. Pootrakulchote, L. Alibabaei, C. H. Ngoc-le, J. D. Decoppet, J. H. Tsai, C. Gratzel, C. G. Wu, S. M. Zakeeruddin and M. Gratzel, ACS Nano, 2009, 3, 3103-3109.

122. Y. X. Zhao and K. Zhu, Chem. Soc. Rev., 2016, 45, 655-689. 123. M. Z. Liu, M. B. Johnston and H. J. Snaith, Nature, 2013, 501, 395-398. 124. W. Feng, T. Umeda, A. Fujii, X. Wang and K. Yoshino, Jpn. J. Appl. Phys.,

2004, 43, 3473-3478.

82

125. L. Schmidt-Mende and M. Grätzel, Thin Solid Films, 2006, 500, 296-301. 126. H. Han, U. Bach, Y.-B. Cheng and R. A. Caruso, Appl. Phys. Lett., 2007, 90,

213510. 127. I. K. Ding, N. Tétreault, J. Brillet, B. E. Hardin, E. H. Smith, S. J. Rosenthal, F.

Sauvage, M. Grätzel and M. D. McGehee, Adv. Funct. Mater., 2009, 19, 2431-2436.

128. C. Olson, D. Veldman, K. Bakker and F. Lenzmann, International Journal of Photoenergy, 2011, 2011, 1-11.

129. J. Melas-Kyriazi, I. K. Ding, A. Marchioro, A. Punzi, B. E. Hardin, G. F. Burkhard, N. Tétreault, M. Grätzel, J.-E. Moser and M. D. McGehee, Adv. Energy Mater., 2011, 1, 407-414.

130. M. Rawolle, K. Sarkar, M. A. Niedermeier, M. Schindler, P. Lellig, J. S. Gut-mann, J. F. Moulin, M. Haese-Seiller, A. S. Wochnik, C. Scheu and P. Muller-Buschbaum, ACS Appl. Mater. Interfaces, 2013, 5, 719-729.

131. P. Docampo, A. Hey, S. Guldin, R. Gunning, U. Steiner and H. J. Snaith, Adv. Funct. Mater., 2012, 22, 5010-5019.

132. R. K. Kei Murakoshi, Yuji Wada, Shozo Yanagida, Chem. Lett., 1997, 471-473.

133. T. K. Yasuteru Saito, Yuji Wada, Shozo Yanagida, Synthetic Met., 2002, 131, 185-187.

134. Y. Saito, Electrochem. Commun., 2004, 6, 71-74. 135. Y. Saito, T. Azechi, T. Kitamura, Y. Hasegawa, Y. Wada and S. Yanagida,

Coord. Chem. Rev., 2004, 248, 1469-1478. 136. N. Fukuri, Y. Saito, W. Kubo, G. K. R. Senadeera, T. Kitamura, Y. Wada and

S. Yanagida, J. Electrochem. Soc., 2004, 151, A1745. 137. J. B. Xia, N. Masaki, M. Lira-Cantu, Y. Kim, K. J. Jiang and S. Yanagida, J.

Am. Chem. Soc., 2008, 130, 1258-1263. 138. X. Z. Liu, W. Zhang, S. Uchida, L. P. Cai, B. Liu and S. Ramakrishna, Adv.

Mater., 2010, 22, E150-E155. 139. N. Vlachopoulos, J. Zhang and A. Hagfeldt, Chimia (Aarau), 2015, 69, 41-51. 140. W. Zhang, Y. Cheng, X. Yin and B. Liu, Macromol. Chem. Phys., 2011, 212,

15-23. 141. Z. Yu and L. Sun, Adv. Energy Mater., 2015, 5, 1500213. 142. J. Burschka, A. Dualeh, F. Kessler, E. Baranoff, N. L. Cevey-Ha, C. Y. Yi, M.

K. Nazeeruddin and M. Gratzel, J. Am. Chem. Soc., 2011, 133, 18042-18045. 143. D. Y. Chen, W. H. Tseng, S. P. Liang, C. I. Wu, C. W. Hsu, Y. Chi, W. Y.

Hung and P. T. Chou, Phys. Chem. Chem. Phys., 2012, 14, 11689-11694. 144. M. Xu, Y. Rong, Z. Ku, A. Mei, X. Li and H. Han, J. Phys. Chem. C, 2013,

117, 22492-22496. 145. A. Abate, T. Leijtens, S. Pathak, J. Teuscher, R. Avolio, M. E. Errico, J. Kirk-

patrik, J. M. Ball, P. Docampo, I. McPherson and H. J. Snaith, Phys. Chem. Chem. Phys., 2013, 15, 2572-2579.

146. B. Xu, J. Huang, H. Agren, L. Kloo, A. Hagfeldt and L. Sun, Chemsuschem, 2014, 7, 3252-3256.

147. B. Xu, E. Gabrielsson, M. Safdari, M. Cheng, Y. Hua, H. Tian, J. M. Gardner, L. Kloo and L. Sun, Adv. Energy Mater., 2015, 5, 1402340.

148. J. Zhang, B. Xu, M. B. Johansson, M. Hadadian, J. P. Correa Baena, P. Liu, Y. Hua, N. Vlachopoulos, E. M. J. Johansson, G. Boschloo, L. Sun and A. Hag-feldt, Adv. Energy Mater., 2016, DOI: 10.1002/aenm.201502536.

149. K. Sveinbjornsson, K. Aitola, X. Zhang, M. Pazoki, A. Hagfeldt, G. Boschloo and E. M. Johansson, J. Phys. Chem. Lett., 2015, 6, 4259-4264.

83

150. J. Zhang, L. Yang, Y. Shen, B. Park, Y. Hao, E. M. J. Johansson, G. Boschloo, L. Kloo, E. Gabrielsson, L. Sun, A. Jarboui, C. Perruchot, M. Jouini, N. Vla-chopoulos, and A. Hagfeldt, J. Phys. Chem. C, 2014, 118, 16591-16601

151. J. Zhang, A. Jarboui, N. Vlachopoulos, M. Jouini, G. Boschloo, A. Hagfeldt, Electrochim. Acta, 2015, 179, 220-227

152. S. Caramori, S. Cazzanti, L. Marchini, R. Argazzi, C. A. Bignozzi, D. Marti-neau, P. C. Gros and M. Beley, Inorg. Chim. Acta, 2008, 361, 627-634.

153. J. Zhang, L. Häggman, M. Jouini, A. Jarboui, G. Boschloo, N. Vlachopoulos, A. Hagfeldt, Chemphyschem, 2014, 15, 1043-1047

154. J. Zhang, N. Vlachopoulos, M. Jouini, M. Johansson, X. Zhang, M. K. Nazee-ruddin, G. Boschloo, E. Johansson, A. Hagfeldt, Nano Energy, 2016, 19, 455-470

155. J. Zhang, B. Xu, M. Johansson, N. Vlachopoulos, G. Boschloo, L. Sun, E. Johansson, A. Hagfeldt, ACS Nano, 2016, 10, 6816-6825

156. J. Zhang, Y. Hua, B. Xu, L. Yang, P. Liu, M. Johansson, N. Vlachopoulos, L. Kloo, G. Boschloo, E. Johansson, L. Sun, A. Hagfeldt, Adv. Energy Mater., 2016, 1601062

157. U. B. Cappel, E. A. Gibson, A. Hagfeldt, G. Boschloo, J. Phys. Chem. C, 2009, 113, 6275-6281

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 1408

Editor: The Dean of the Faculty of Science and Technology

A doctoral dissertation from the Faculty of Science andTechnology, Uppsala University, is usually a summary of anumber of papers. A few copies of the complete dissertationare kept at major Swedish research libraries, while thesummary alone is distributed internationally throughthe series Digital Comprehensive Summaries of UppsalaDissertations from the Faculty of Science and Technology.(Prior to January, 2005, the series was published under thetitle “Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology”.)

Distribution: publications.uu.seurn:nbn:se:uu:diva-300802

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2016