Akademisk avhandling som med tillstånd av Kungl Tekniska Högskolan i Stockholm framlägges till offentlig granskning för avläggande av teknisk doktorsexamen Tisdagen den 7 Juni kl 10.00 i FB42, KTH, Roslagstullsbacken 21, AlbaNova, Stockholm. Avhandlingen försvaras på engelska. Opponent är Dr. Erik Johansson, Uppsala Universitet, Sweden.

Strategies for Performance Improvement of

Quantum Dot Sensitized Solar Cells

Jing Huang 黄 静

Doctoral Thesis

Stockholm 2016

Strategies for Performance Improvement of Quantum Dot Sensitized Solar Cells Thesis for Philosophy Doctor Degree Department of Theoretical Chemistry & Biology School of Biotechnology Royal Institute of Technology

ISBN 978-91-7595-995-5 ISSN 1654-2312 TRITA-BIO Report 2016:11 © Jing Huang, 2016 Universitetsservice US AB, Stockholm

Abstract

Quantum dot sensitized solar cells (QDSCs) constitute one of the most promising low-cost solutions  that  currently  are  explored  for  the  world’s  needs  of clean and renewable energy. Efficient, low-toxic and stable QDSCs for large-scale applications have therefore formed the subject for much of the solar cell research during recent years. This circumstance also forms the basic motivation for this thesis, where the results of my studies to improve the efficiency and stability of green QDSCs are presented and discussed.

The surface condition of quantum dots (QDs) is always crucial to the performance of QDSCs, since surface ligands can influence the loading amount of QDs on photoanodes, and that the surface defects can induce charge recombination in the solar cells. In this thesis work, a hybrid passivation approach was firstly utilized to improve the photovoltaic performance of CdSe QDs. After hybrid passivation by mercaptopropionic acid (MPA) and iodide ions, the loading efficiency of the QDs was increased with the ligands of MPA, and the surface defects on the QDs were reduced by the iodide ions, both aspects contributing to a dramatic enhancement in the efficiency of the CdSe based QDSCs. This hybrid passivation strategy was then employed for low-toxic CuInS2 QDs, and was also demonstrated as an effective way to modify the surface state of the CuInS2 QDs and improve the performance of the QDSCs based on CuInS2 QDs.

To improve the stability of the QDSCs, solid state quantum dot sensitized solar cells (ss-QDSCs) based on CuInS2 QDs were investigated. In addition to the hybrid passivation, increasing the pore size of the TiO2 active layer and changing the composition of the CuInS2 QDs were also found to be useful approaches to improve the performance of the ss-QDSCs based on CuInS2 QDs. Furthermore, for the most used hole transport material- 2,2´,7,7´-tetrakis-(N,N-di-p-methoxyphenylamine)-9,9´-spirobifluorene (Spiro-OMeTAD) - in solid state solar cells, silver bis(trifluoromethanesulfonyl)imide was shown to be an effective p-type dopant to increase its conductivity and to improve the performance of the solar cells based on it.

Keywords: quantum dot sensitized solar cells, passivation, solid state quantum dot sensitized solar cells, performance improvement, low-toxic, hole transport materials, P-type dopant

Abbreviations

AgTFSI silver bis(trifluoromethanesulfonyl)imide

APCE absorbed photon-to-current efficiency

CBD chemical bath deposition

DSCs dye-sensitized solar cells

EQE external quantum efficiency

FF fill factor

FTIR Fourier transform infrared spectroscopy

HOMO highest occupied molecular orbital

HTMs hole transport materials

IPCE incident photon-to-current efficiency

IS impedance spectroscopy

Jsc short-circuit current

J-V current-voltage

LHE light harvesting efficiency

MEG multiple exciton generation

MPA mercaptopropionic acid

OA oleic acid

OAm oleylamine

ODE 1-octadecene

PLQY photoluminescence quantum yield

PSCs perovskite solar cells

P3HT poly(3-hexylthiophene)

QDs quantum dots

QDSCs quantum dot sensitized solar cells

SILAR successive ionic layer adsorption and reaction

Spiro-OMeTAD 2,2´,7,7´-tetrakis-(N,N-di-p-methoxyphenylamine)-9,9´-

spirobifluorene

ss-QDSCs solid-state quantum dot sensitized solar cells

TBAI tetrabutylammonium iodide

TEM transmission electron microscopy

TOP tri-n-octylphosphine

UV-Vis ultraviolet-visible

Voc open-circuit voltage

XPS X-ray photoelectron spectroscopy

XRD powder X-ray diffraction

η power conversion efficiency

List of Publications

This thesis is based on the following papers, referred to in the text by their Roman numerals I-IV:

I. Improved Performance of Colloidal CdSe Quantum Dot-Sensitized Solar Cells by Hybrid Passivation Jing Huang, Bo Xu, Chunze Yuan, Hong Chen, Junliang Sun, Licheng Sun and Hans Ågren ACS Appl. Mater. Interfaces, 2014, 6, 18808-18815

II. Photovoltaic Performance Improvement Strategies for Colloidal CuInS2 QDs: Hybrid Passivation vs ZnS Shell Jing Huang, Bo Xu, Asghar Jamshidi Zavaraki, Licheng Sun and Hans Ågren Submitted for publication, 2016

III. Efficiency Enhancement of Solid-State Quantum Dot-Sensitized Solar Cells by Quality Optimizing of Colloidal CuInS2 Quantum Dots

Jing Huang, Bo Xu, Yongfei Ji, Shizhao Xiong, Aleksandar Matic, Yi Luo, Licheng Sun and Hans Ågren.

Submitted for publication, 2016

IV. AgTFSI as p-Type Dopant for Efficient and Stable Solid-State Dye-Sensitized and Perovskite Solar Cells

Bo Xu, Jing Huang, Hans Ågren, Lars Kloo, Anders Hagfeldt and Licheng Sun.

ChemSusChem., 2014, 7, 3252-3256

Papers not included in this thesis:

V. Improved Photocurrent in Quantum Dot Sensitized Solar Cells by Employing Alloy PbxCd1-xS QDs as Photosensitizers Chunze Yuan, Lin Li, Jing Huang, Zhijun Ning, Licheng Sun, Hans Ågren

Revision, Nanomaterials. 2016

VI. Novel Small Molecular Materials Based on Phenoxazine Core Unit for Efficient Bulk Heterojunction Organic Solar Cells and Perovskite Solar Cells Ming Cheng, Cheng Chen, Xichuan Yang, Jing Huang, Fuguo Zhang, Bo Xu, and Licheng Sun Chem. Mater., 2015, 27, 1808–1814

Table of Contents

Abstract Abbreviations List of publications

1. Introduction ..................................................................................... 1

1.1. Solar energy and photovoltaics ............................................................. 1 1.2. Quantum dot sensitized solar cells ........................................................ 4

1.2.1. Quantum dots ........................................................................................... 4 1.2.2. Quantum dot sensitized solar cells ............................................................ 6 1.2.3. Solid-state quantum dot sensitized solar cells ........................................... 8 1.2.4. Perovskite solar cells ................................................................................ 9

1.3. Colloidal QDs and their application in QDSCs ..................................... 11 1.3.1. Colloidal QDs ...........................................................................................11 1.3.2. Colloidal QDs in solar cells ......................................................................14

1.4. Performance improvement strategies for colloidal QDSCs .................. 15 1.4.1. Increasing loading amount of colloidal QDs .............................................15 1.4.2. Decreasing surface defects on colloidal QDs ...........................................16 1.4.3. Replacing the polysulfide electrolyte with solid state HTMs ......................17

1.5. The aim of this thesis ........................................................................... 18 2. Experimental Methods .................................................................. 19

2.1. Synthesis of colloidal QDs ................................................................... 19 2.2. Ligand exchange of the as-synthesized colloidal QDs......................... 20 2.3. Fabrication of solar cells ...................................................................... 20

2.3.1. Fabrication of liquid QDSCs .....................................................................21 2.3.2. Fabrication of ss-QDSCs .........................................................................21 2.3.3. Fabrication of perovskite solar cells .........................................................21

2.4. Characterization................................................................................... 22 2.4.1. Structural characterization .......................................................................22 2.4.2. Optical characterization ...........................................................................22 2.4.3. Photovoltaic characterization ...................................................................23

3. Passivation strategies for liquid QDSCs ....................................... 27 3.1. Introduction .......................................................................................... 27 3.2. Hybrid passivation for colloidal CdSe based QDSCs (Paper I) ............ 29 3.3. Strategies for photovoltaic performance improvement for colloidal CuInS2 QDs: hybrid passivation vs ZnS shell (Paper II) .................................. 35 3.4. Conclusions ......................................................................................... 40

4. Improvement strategies for ss-QDSCs ......................................... 41 4.1. Introduction .......................................................................................... 41

4.2. Efficiency enhancement of CuInS2 based ss-QDSCs (Paper III) ......... 42 4.3. AgTFSI as p-type dopant for perovskite solar cells (Paper IV) ............ 48 4.4. Conclusions ......................................................................................... 50

5. Concluding Remarks .................................................................... 51 6. Future Outlook .............................................................................. 52 Acknowledgements Appendices References

1

1. Introduction

The significance of energy is apparent in human life, especially in the modern time when almost every aspect in our daily life is related to energy consumption. However, with the explosion of the world’s population and the depletion of fossil fuels, energy crisis is threatening the whole world now. In addition to a possible energy crisis, the combustion of fossil fuels has induced serious environmental pollution to many countries, from the past to the present, such as the Great Smog of 1952 in London, and the extremely high level of PM 2.5 in Beijing in recent years. Therefore, it is urgent to exploit renewable and clean energy to replace the traditional fossil fuels and keep the development of our society free from this environmental threat.

1.1. Solar energy and photovoltaics

Solar energy is a promising solution to the problems of the energy crisis and environmental pollution, since solar energy is highly abundant, renewable and clean. The total radiated power of solar energy received by the Earth is around 1.7×1017 W. Due to the atmospheric absorption and scattering, the final amount of solar light that strikes the  Earth’s  surface  is  8.0×1016 W, which is still much larger  than  human’s  current  energy  consumption.1 Furthermore, solar energy is one of the most equitably distributed energy resources, and is not highly related with economic wealth. Therefore, solar energy is the ideal energy resource for mankind to use.

Photovoltaic cells (solar cells) are devices that directly convert solar energy to electricity. Since the first practical photovoltaic cell (silicon based solar cell) developed by Bell Laboratories,2 photovoltaic technology has been rapidly progressed. Until now, photovoltaic technology can be classified to two categories: wafer based solar cells and thin-film solar cells, as shown in Figure 1.1 The efficiency progress for each kind of solar cells are presented in Figure 2.

2

Figure 1. Typical structures of photovoltaic device.1

From Figure 1, we find that the thin-film technology is further divided into two groups – commercial thin-film and emerging thin-film technology. The emerging thin-film solar cells, also categorized as the third-generation photovoltaics, contain organic solar cells, dye-sensitized solar cells (DSCs), quantum dot solar cells and perovskite solar cells. These kinds of solar cells employ nanostructured materials to achieve desired optical and electronic properties, and offer the potential to accomplish devices with visible transparency, flexibility and high weight-specific power [W•g-1], together with low-cost fabrication techniques, making them promising candidates in the large-scale application of photovoltaics. At the same time, these solar cells also attract great interest from scientists and researchers, and their efficiency has been significantly improved in recent years, especially for quantum dot solar cells and perovskite solar cells (Figure 2).

3

Figure 2. Certified record efficiency for different kinds of solar cells.

(http://www.nrel.gov/ncpv/images/efficiency_chart.jpg)

4

1.2. Quantum dot sensitized solar cells

1.2.1. Quantum dots

Since the concept of nanotechnology was first introduced by Richard Feynman in   his   famous   talk   and   paper   “There’s Plenty   of   Room   at   the   Bottom”, nanotechnology and nanomaterials have attracted extensive attention, especially in material science and medical science. Nanomaterials refer to the materials with at least one dimension size below 100 nm. Because of the small size, nanomaterials may exhibit widely different chemical, physical and electronic properties from that of the bulk materials. Quantum dots (QDs) constitute one kind of most popular nanomaterials in both scientific research and industrial applications. By definition, QDs are inorganic semiconductors structured at the nanoscale, normally with the size between 1-10 nm.3 Due to the ultra-small size, QDs show distinct advantages, such as size tunable effect, large extinction coefficient, and potential of multiple exciton generation, which make QDs promising candidates for a variety of applied researches, for instance in the biomedical area4, photonics5, and for renewable energy exploration6, 7.

Size tunable effect (quantum confinement effect)

Because the size of QDs is confined in all three dimensions, the movement of charge carriers is restricted   in   “a   small   box”,   resulting   in a quantum confinement. Especially when the size of the QDs decreases to a critical size, which is twice the size of its exciton Bohr radius, the excitons (electron-hole pairs) are further squeezed. At this time, the electronic density will split into discrete energy levels from the continuous energy band of the bulk material. As a result, the QD bandgap will be increased with the increment of the quantum confinement effect, by following the effective-mass approximation8:

𝐸 , = 𝐸 , + ∆𝐸 = 𝐸 , + ħ + − . (1.1)

where Eg,QD and Eg,bulk is the bandgap energy of the QD and bulk material, me and mh is the effective mass of electron and hole in the solid, respectively, and R is the radius of the QD. Therefore, when the size of the QD gradually

5

decreases, the bandgap will be enlarged, leading to a blue-shift of the QD absorption and emission bands, as shown in Figure 3. For applications in solar cells, QDs with different energy levels and light harvesting ability could be obtained by just tuning their size, to fabricate efficient devices.

Figure 3. A scheme to show that the bandgap and optical property change of the QDs associated with their size.

Multiple exciton generation

Multiple exciton generation (MEG) is another unique property of QDs, which means that two or more electron-hole pairs can be generated from one QD upon excitation by one photon. Specifically, when a semiconductor absorbs a photon with energy larger than its bandgap Eg, a hot exciton with excess energy, ∆𝐸 = ℎ𝜈 − 𝐸 , will be produced. The excess energy can be converted into kinetic energy before cooling, and if ∆𝐸 > 𝐸 , it will generate another one or more electron-hole pairs through the process of impact ionization (Figure 4).9 It is fundamental that this impact ionization process is faster than the cooling or other relaxation processes of the hot carriers. QDs with strong quantization effects have been verified to significantly affect the relaxation dynamics of the photogenerated carriers, and are thus more likely to exhibit the MEG phenomenon.10 So far, MEG behavior has been demonstrated in colloidal PbS and PbSe QDs.11 With the behavior of MEG, the external

6

quantum efficiency and photocurrent generated from QDs could be significantly enhanced (as shown in Figure 5).12 Therefore, the maximum theoretical conversion efficiency of quantum dot based solar cells can boost to 44%.13

Figure 4. MEG process in a QD through impact ionization.10

Figure 5. An external quantum efficiency that peaked at 114% from MEG in PbSe QD-based solar cells with enhanced photocurrent, and the associated internal

quantum efficiency was 130%.12

1.2.2. Quantum dot sensitized solar cells

Benefitting from the advantages of high extinction coefficient, broad and tunable light absorption, low-cost solution processed preparation, and the potential of MEG, QDs have been more and more employed in solar cells as photosensitizers, to construct quantum dot sensitized solar cells (QDSCs).

7

Figure 6. Basic structure of QDSCs that shows the primary components and electron transfer processes in the device under illumination.

Figure 6 shows a basic configuration of QDSCs, which is mainly composed by a quantum dot-sensitized photoanode, electrolyte and a counter electrode. Upon illumination, electrons are generated from the excited QDs, and then inject to the conduction band of a wide bandgap semiconductor, followed by diffusion to the front contact. At the same time, holes are transferred to the electrolyte for reduction of the photooxidized QDs, while electrons from the counter electrode can regenerate the electrolyte.

The photoanode, which is the working electrode, is usually fabricated by deposition of QDs on a mesoporous nanocrystalline substrate, which is made by wide bandgap semiconductors (such as TiO2 and ZnO). The mesoporous nanostructured semiconductor film provides a microscopic surface area, allowing for a large loading amount of the sensitizers. The QD deposition methods mainly include successive ionic layer adsorption and reaction (SILAR), chemical bath deposition (CBD), direct adsorption and linker assisted adsorption of pre-synthesized colloidal QDs. The conventional choices of QDs for the QDSCs are CdS QDs14, 15 and CdSe QDs16, 17, and later on core/shell structured CdSe/CdS QDs18 and alloyed CdSexTe1-x QDs19 have attracted much attention in the fabrication of QDSCs. In addition, small bandgap semiconductors, such as PbS20, PbSe21 and HgS22, have also usually been employed in QDSCs. Due to the toxicity of Cd, Pb and Hg, there is a

8

tendency to replace these QDs with CuInS2 QDs23, 24, CuInSe2 QDs25, and CuInSxSe1-x QDs26,   to   fabricate  “green”  QDSCs.  So far, the record efficiency for QDSCs is 11.6%, which is based on the low-toxic Zn-Cu-In-Se QDs27.

Although the I-/I3- redox couple is used as a common electrolyte in DSCs with

excellent performance, it is not suitable for QDSCs, because the I-/I3- redox

couple can cause serious corrosion to the inorganic QDs, resulting in fast degradation of the photoanodes.28 Instead, the most used electrolyte in QDSCs is polysulfide redox couple in aqueous solution, which has been demonstrated to regenerate the oxidized QDs efficiently.29 However, because of the highly negative redox potential of the polysulfide electrolyte, the photovoltage of the QDSCs is limited. In order to overcome this limitation, numerous new types of liquid electrolytes have been tried, such as the modified polysulfide electrolyte [(CH3)4N]2S/[(CH3)4N]2Sn,30 an organic electrolyte based on McMT−/BMT redox couple,31 and the Co(o-phen)3

2+/3+ complex-based electrolyte17. Unfortunately, the overall performance of these redox couples in QDSCs still cannot be comparable with the conventional polysulfide electrolyte. Therefore, the search for suitable electrolytes for QDSCs is still a big challenge for scientists.

The task of a counter electrode is to collect electrons from the external circuit and then reduce the oxidized species in the electrolyte. Therefore, the counter electrode should be catalytically active toward the redox couple in the electrolyte. For the polysulfide electrolyte in QDSCs, nanostructured metal chalcogenides, such as CoS,32 Cu2S,33 FeS2

34 and CuInS2,35 are often employed to make the counter electrode active to catalytically reduce the polysulfide electrolyte. Among these materials, Cu2S based counter electrodes show the highest electrocatalytic activity to the reduction of Sn

2- species in the polysulfide electrolyte, and have thus become the most popular choice for fabrication of QDSCs.

1.2.3. Solid-state quantum dot sensitized solar cells

Though the performance of QDSCs have achieved great progress in recent years, the commercialization of QDSCs has remained limited mainly owing to the utilization of liquid electrolytes, which always suffer from leakage and evaporation problems in the long run. Therefore, it is necessary to replace the

9

liquid electrolyte in QDSCs with a solid state electrolyte, such as hole transport materials (HTMs) and hole conductors, to construct solid-state quantum dot sensitized solar cells (ss-QDSCs).

Figure 7. Device structure of a ss-QDSC.

Figure 7 presents a typical structure of a ss-QDSC. Except that the liquid electrolyte is replaced by the HTM layer and the corresponding back contact is changed to silver or gold, the structure of the ss-QDSCs is similar with the configuration of the liquid QDSCs. However, because of the pore filling problem of the HTM, the mesoporous TiO2 layer in a solid state solar cell has to be much thinner than that in a liquid one, which may result in a lower harvesting capability of the photoanode.

So far, several chalcogenide semiconductors, such as PbS,14, 36, 37 CdS,38, 39 CdSe17, 40 and Sb2S3

41-44, have been successfully employed in ss-QDSCs. Among these nanostructures, Sb2S3 QDs were considered as the most efficient sensitizers in ss-QDSCs that have been reported, with a PCE of 3.1% when 2,2´,7,7´-tetrakis-(N,N-di-p-methoxyphenylamine)-9,9´-spirobifluorene (Spiro-OMeTAD) was employed as HTM,42 and the PCE of 5.13% when poly(3-hexylthiophene) (P3HT) was used as HTM and co-light-absorber simultaneously.44

1.2.4. Perovskite solar cells

The perovskite solar cell (PSC) is also an emerging thin-film photovoltaic device which has attracted extensive attention in the recent five years.

10

Perovskites are a class of compounds that have the same crystal structure as CaTiO3, and the most used one in perovskite solar cells is the organometal halide perovskite CH3NH3PbX3 (X can be a halogen element Cl, Br, I, or a combination of them), the structure of which is shown in Figure 8. Similarly with QDs, perovskites are mostly employed as light harvesting materials in solar cells. Due to their panchromatic light absorption and ambipolar behavior, the performance of perovskite solar cells is quite promising.

Figure 8. Illustration of perovskite structure ABX3, where A=CH3NH3, B= Pb, and X= Cl, Br or I.45

The configuration of a typical PSC is similar with that of a ss-DSC and a ss-QDSC, in which the perovskite nanocrystals grow on a mesoporous metal oxide layer, and then a layer of HTM is applied, followed by thermal evaporation of a noble metal layer as back contact. Although perovskite was first incorporated in liquid solar cells using an iodide/triiodide-based electrolyte in 2009 with the efficiency lower than 4%,46 the breakthrough was achieved in 2012 when perovskite was employed in solid state devices with the efficiencies over 9%.45, 47 After great efforts made by scientists from all over the world in the last four years, the efficiency of PSCs has been improved to 20.8%,48 motivating lots of companies and governments to put PSCs in large-scale fabrication and application.

11

Figure 9. Schematic representation of PSC device structure (left), and a cross-sectional SEM image of a PSC device (right).45

1.3. Colloidal QDs and their application in QDSCs

1.3.1. Colloidal QDs

There mainly are two kinds of QDs that can be employed in QDSCs. One kind is the QDs prepared from in-situ methods, such as SILAR and CBD. These QDs are just composed by inorganic nanoparticles, which are formed while attaching on the TiO2 nanoparticles. The other ones are QDs made from an ex-situ approach, which are pre-synthesized before deposition on the TiO2 electrodes. These QDs are normally colloidal QDs, which consist of inorganic nanoparticles coated with a layer of organic ligand molecules, as shown in Figure 10.

Figure 10. Illustration of a colloidal QD, composed of an inorganic nanoparticle capped by a layer of organic ligand molecules.

12

Compared with the QDs made from the in-situ method, the hybrid nature of colloidal QDs offers great flexibility in engineering their chemical and physical properties, since both components of the colloidal QDs can be independently optimized to achieve desired properties. For example, the size, shape, composition and crystallinity of colloidal QDs can be extensively exploited from a controllable synthesis pathway.26 It is notable that the multiple exciton generation has been observed from nearly monodispersed colloidal QDs, attributed to the desired quality of these nanostructures.12, 49 Furthermore, the organic ligand layer opens up the potential of surface chemistry manipulation, resulting in the possibility of tailoring more properties. As the size of the colloidal QD decreases, the surface to volume ratio dramatically increases, leading to stronger interaction between the surface atoms and the ligand molecules. Consequently, the colloidal QDs can be easily dispersed in solvents, making the processing of QDs in solution possible, which is a great advantage of colloidal QDs over nanomaterials fabricated by other techniques, such as molecular beam epitaxy. These features have made colloidal QDs become promising materials for applications in biomedical imaging, light-emitting diodes (LEDs), solar cells, and so on.

Synthesis and ligand exchange for colloidal QDs

Generally, size- and shape- controlled synthesis of colloidal QDs can be achieved by the thermal decomposition method, which can be accomplished by either the hot-injection method or the one-pot method (also knowns as the heat-up method). For the hot-injection approach, the nucleation occurs upon a rapid injection of a room temperature precursor into a pre-heated surfactant, followed by reaction cooling and growth stage commencing (Figure 11a).50 For the one-pot method, all precursors and ligand are present from the beginning of the reaction, followed by steadily heating to trigger the nucleation and growth stages (Figure 11b).51

13

Figure 11. Schematic diagrams of synthetic methods for colloidal QDs: (a) hot-injection method and (b) one-pot heating-up method.6

Due to the synthesis with organic surfactants (e.g. oleic acid, oleylamine, tri-n-octylphosphine), colloidal QDs are always capped by a hydrophobic ligand layer (i.e. long alkyl chain), rendering them dispersible in organic solvents. However, some applications require the QDs to be water-dispersible, such as biomedical imaging, thus that the native hydrophobic ligand need to be exchanged by hydrophilic or charged ones. In addition, for the application in QDSCs, it has been demonstrated that colloidal QDs coated with bifunctional ligand molecule - mercaptopropionic acid (MPA), can be deposited on the TiO2 electrodes more efficiently,52, 53 which is attributed to the anchoring carboxyl group. More importantly, although the long-chain ligand can stabilize the colloidal QDs in solution, they act as an insulating medium around the QDs, which is detrimental for the application in solar cells of the colloidal QDs. Exchanging the initial ligand with small organic molecules can lead to dramatic improvements in device performance.26 Therefore, ligand exchange is a valuable tool for property modification of colloidal QDs. Figure 12 illustrates a ligand exchange reaction for colloidal QDs in solution, which can offer the

14

ligand a high degree to access the surface of the QDs, and maximize the opportunity of ligand exchange.

Figure 12. Schematic diagrams of ligand exchange reaction.

1.3.2. Colloidal QDs in solar cells

The first report of employing colloidal QDs in solar cells appeared in 1998, in which InP colloidal QDs were used as sensitizers in QDSCs.54 After that the development of QDSCs based on colloidal QDs underwent a slow progressing period, probably due to the low loading amount of the colloidal QDs on the photoelectrodes. Recently, benefitting from the linker-assisted assembly from bifunctional ligand, the surface coverage of colloidal QDs has been dramatically increased. Combining with the high quality of the colloidal QDs, the efficiency of the resulting QDSCs has exhibited great improvement. Until now, most of the highly efficient QDSCs (with efficiency > 6%) are based on colloidal QDs,18, 24, 55-57 and the record efficiency of the QDSCs is also prepared from colloidal Zn-Cu-In-Se QDs, which amount to 11.6%.27

In addition to the application in QDSCs, colloidal QDs are also used to construct depleted heterojunction colloidal quantum dot solar cells. Inspired by sensitized solar cells, the depleted heterojunction architecture employs a highly doped n-type metal oxide and a p-type colloidal QD film to form a p-n heterojunction - the typical structure of a depleted heterojunction colloidal quantum dot solar cell is shown in Figure 13.58 Benefitting from the new ligand strategies for colloidal QDs,58-60 which were developed   by   Sargent’s  

15

group, the performance of the depleted heterojunction colloidal quantum dot solar cells have been advanced rapidly, now reaching an efficiency as high as 10.8%.61

Figure 13. Structure of a depleted heterojunction colloidal quantum dot solar cell (a, Inset: photograph of a typical device), and cross-sectional SEM image of the device

(b, scale bar is 500 nm).58

1.4. Performance improvement strategies for colloidal QDSCs Although colloidal QDs have many attracting advantages, and although indeed much progress has been made for the performance improvement of the QDSCs in recent years, the best conversion efficiency of QDSCs still cannot be comparable with that of their analogue DSCs (amounting to 13%62). The major reasons for this could be the low loading amount of colloidal QDs on photoelectrodes, serious recombination caused by high surface trap density on the QDs, and low photovoltage due to the use of the polysulfide electrolyte. In order to further improve the performance of the QDSCs, in our opinion, efforts can be made along the following lines.

1.4.1. Increasing loading amount of colloidal QDs

Due to the large size of QDs (compared to the dye molecules), the deposition of the QDs on a metal oxide matrix, to form a well-covered sensitization, is difficult, and the native long alkyl ligand which capped on the colloidal QDs make the loading efficiency even lower. The poor sensitization of QDs not only limits the light harvesting capability of the photoelectrodes, but also induces charge recombination between the metal oxide and the electrolyte. Ex

16

situ linker-assisted QD deposition,53 in which the QDs undergo a ligand exchange process with small and bifunctional molecules, such as MPA, before sensitization (illustrated in Figure 14), has significantly enhanced the surface coverage of the colloidal QDs to 34%.52 However, this is still low for achieving a compact sensitization layer. Therefore, a substantial deposition of high quality colloidal QDs onto mesoporous matrices remains a challenge for research.

Figure 14. Scheme of in situ ligand exchange linker-assisted (a) and ex situ ligand exchange linker-assisted routes for the deposition of colloidal QDs on TiO2

electrodes (b).53

1.4.2. Decreasing surface defects on colloidal QDs

Figure 15. Illustration of photo-generated electrons trapped by surface states on QDs in a QDSC device.

17

Surface atoms of QDs have fewer neighbors than their interior counterparts, resulting in dangling bonds on the surface atoms. These unshared atomic orbitals create energy levels (surface states) within the bandgap of the QDs, and these states are detrimental to the quantum yield of the QDs, since they can cause non-radiative recombination of the excitons. When the QDs with extensive surface trap states are employed in QDSCs, the photo-generated electrons can be trapped by the surface states and then recombine with the holes, rather than injecting to the TiO2 nanoparticles (Figure 15). As a result, these surface defects can act as recombination centers in the devices, which can reduce charge carrier populations and lower the open-circuit voltage of the solar cells. Therefore, a reduction of the surface defects is critical to improve the performance of QDSCs.

Surface passivation is a process that can eliminate the dangling bonds and optimize the surface quality of QDs. Generally, the surface of QDs can be passivated by overcoating a shell of a wide bandgap semiconductor, to increase their photoluminescence quantum yield (PLQY).63, 64 Recently,  Sargent’s  group  have established an atomic passivation method for heterojunction quantum dot solar cells, by using halide anions to effectively passivate the surface defects on colloidal PbS QDs.58-60 This atomic ligand passivation can even remedy the trap states deep in the bandgap, improving charge carrier diffusion in the quantum dot films and enhancing the photovoltaic performance of the devices.65, 66 However, the atomic passivation strategy has not been applied to passivate the surface states of colloidal QDs for QDSCs until now. In addition, effective surface passivation methods for the colloidal QDSCs are still under exploration.

1.4.3. Replacing the polysulfide electrolyte with solid state HTMs

Although the polysulfide electrolyte has been demonstrated to efficiently regenerate the oxidized QDs in QDSCs, the open-circuit voltage of the QDSCs is still limited by its highly negative redox potential.29 Besides, due to the risk of leakage and evaporation, this aqueous electrolyte solution can induce a stability issue of the solar cells, which is detrimental for real applications. These two problems could be alleviated by employing solid state HTMs to replace the liquid polysulfide electrolyte in the solar cells. Solid state HTMs, which always are large conjugated organic molecules or conjugated polymers

18

(such as Spiro-OMeTAD and P3HT), can form a thin film on the electrodes, and without liquid component in the devices, the stability of solar cells could be significantly increased. Furthermore, the highest occupied molecular orbital (HOMO) levels of the HTMs are always more positive, making the resultant ss-QDSCs theoretically show higher open-circuit voltage. Although there have been many studies of ss-QDSCs, their performance remained unsatisfactory.

1.5. The aim of this thesis The aim of this thesis is to develop efficient, stable and green QDSCs. To realize this aim, the thesis was divided into three steps. The first step was to find an effective passivation strategy for the most investigated colloidal CdSe QDSCs to improve their photovoltaic performance, and then in the second step, this passivation strategy was applied for the green CuInS2 QDSCs. For a more comprehensive study, the most used shell coating passivation to the colloidal CuInS2 QDs was also investigated as a comparison. In the last step, the green colloidal CuInS2 QDs with surface passivation were employed in the ss-QDSCs configurations, in order to develop low-toxic and stable solar cells. For the further performance improvement of the ss-QDSCs, some other methods have been tried, such as enlarging the pore size of the TiO2 mesoporous film, optimizing the quality of the colloidal CuInS2 QDs and using doping to increase the conductivity of the HTMs.

19

2. Experimental Methods

2.1. Synthesis of colloidal QDs

In this thesis, two kinds of colloidal QDs are used - toxic CdSe QDs and low-toxic CuInS2 QDs. Both of these two kinds of colloidal QDs were synthesized by following literature methods with few modifications. The CdSe QDs were prepared via the one-pot method67 and the CuInS2 QDs were synthesized by the hot-injection approach24. All reactions were performed in an inert atmosphere by employing the standard Schlenk techniques.

Synthesis of colloidal CdSe QDs. A Se precursor solution and a Cd precursor solution were prepared before synthesizing the CdSe QDs. The Se precursor solution was obtained by dissolving Se powder in a mixture solution of tri-n-octylphosphine (TOP) and paraffin, and the Cd precursor solution was prepared by dissolving CdO in oleic acid (OA) and paraffin. Then the Se precursor solution was added into the Cd precursor solution at room temperature and vacuumed at 110 oC, followed by heated to 280 oC under Ar atmosphere. After stayed at 280 oC for 10 min, the reaction was terminated by cooling down the reaction solution, and the obtained QDs could easily disperse in an organic solvent after purification, such as in toluene and chloroform. These as-synthesized CdSe QDs exhibit a size range of 4.0 ± 0.3 nm (Figure 16) with the cubic structure (Paper I, Figure S1).

Figure 16. A TEM image of the obtained CdSe QDs.

20

Synthesis of colloidal CuInS2 QDs. CuI, In(OAc)3, oleylamine (OAm) and 1-octadecene (ODE) were firstly mixed and heated to 180 oC under Ar atmosphere. Then sulfur precursor solution, obtained by dissolving sulfur powder in OAm, was injected into the above mixed solution at 180 oC. After maintaining at this temperature for 20 min, the reaction was quenched by removing the heater. Similar to the colloidal CdSe QDs, the obtained CuInS2 QDs could easily disperse in toluene or chloroform. From the TEM image, we find that the size of the as-synthesized CuInS2 QDs is around 7 nm, with pyramidal shape.

Figure 17. A TEM image of the as-synthesized CuInS2 QDs.

2.2. Ligand exchange of the as-synthesized colloidal QDs

All the obtained colloidal QDs from synthesis were hydrophobic, and ligand exchange reactions were applied to them in order to modify their surface states. Specifically, the ligand solution, formed by dissolving a ligand reagent in methanol with the pH value of the solution adjusted to value of 12 using NaOH solution, was dropped into the QDs stock solution under strong stirring. After stirring for a certain period, the QDs transferred to the inorganic phase, and were then collected and purified.

2.3. Fabrication of solar cells

Three kinds of solar cells were studied and fabricated in this thesis: liquid QDSCs, ss-QDSCs and perovskite solar cells.

21

2.3.1. Fabrication of liquid QDSCs

Photoanodes. TiO2 mesoporous electrodes with a transparent layer and a scattering layer, which are prepared by a screen printing method with commercial paste, were employed as photoanode substrates in the QDSCs. Then the QDs were deposited onto the TiO2 electrodes by pipetting the QDs aqueous solution on TiO2 films and keeping for a certain time before washing away the excess QDs. After the deposition, the QD-sensitized electrodes were coated with two layers of ZnS by the SILAR method.

Cu2S counter electrodes. The counter electrodes were obtained by immersing polished brass foils in HCl solution, followed by sulfidation with a prepared polysulfide electrolyte solution.

Polysulfide electrolyte and assembling of solar cells. The polysulfide solution consist of 2 M Na2S, 2 M S, and 0.2 M KCl in methanol-water solution. Then a solar cell was constructed by assembling a QD-sensitized photoanode and a Cu2S counter electrode with a Surlyn film. The polysulfide electrolyte was then injected by vacuum backfilling.

2.3.2. Fabrication of ss-QDSCs

Photoanodes. TiO2 mesoporous electrodes that were used in ss-QDSCs just contained a thin active layer, without any scattering layer. Then the following preparation of the QD-sensitized photoanodes was similar with that of the liquid QDSCs.

HTM layer and Ag back contact. An HTM film was subsequently formed by applying HTM solution on the QD-sensitized photoanodes with the spin-coating approach in the open air. Finally, an Ag back contact was deposited onto the HTM film by thermal evaporation in a vacuum chamber, to complete the device construction.

2.3.3. Fabrication of perovskite solar cells

A 400 nm Al2O3 scaffold, which is prepared by the spin-coating method, was used as photoanode substrate for the PSC. Then the CH3NH3PbI3-XClX perovskite precursor solution was spin-coated on the substrate and then kept at

22

100 oC for 30 min, to allow for the growth of the perovskite QDs. After that, the HTM solution with corresponding additives, was applied on the electrodes to form an HTM film, followed by thermal evaporation of the Ag back contact in a vacuum chamber to conclude the device fabrication.

2.4. Characterization

2.4.1. Structural characterization

Transmission electron microscopy (TEM) is always used to examine the size, shape, structure and disperse state of the QDs. With energy-dispersive X-ray spectroscopy (EDS) on the TEM, the composition of the QDs could be obtained. TEM images and EDS spectra were obtained from a JEOL JEM-2100 microscope at 200kV.

Powder X-ray diffraction (XRD) is employed to determine the crystal structure of the QDs, and all XRD patterns were recorded from an X’Pert  PANalytical  PRO MRD using Cu Kα1 radiation (λ  = 1.54056 Å).

X-ray photoelectron spectroscopy (XPS) is a surface-sensitive technique that measures the elemental composition of the samples; therefore, the composition of the QDs could be determined accurately from XPS results. XPS analysis was performed with Al-Kα   radiation   (200  W,   13   kV)   as   X-ray source at a pressure of 10-9 Torr, and the resulting photoelectron spectra were collected by a PHI 5800 system from Physical Electronics.

Fourier transform infrared spectroscopy (FTIR) is used to recognize chemical species that can absorb characteristic frequencies in the infrared region, thus it could be used to detect the organic ligand on the QDs. FTIR spectra were measured by a Biorad FTS 375 spectrometer.

2.4.2. Optical characterization

Ultraviolet-Visible (UV-Vis) absorption spectroscopy was performed to investigate the light-harvesting ability of the QDs to disperse in solution, as well as the absorption of the sensitized photoelectrodes that can reflect the loading amount of the QDs on the electrodes. The UV-Vis spectra were recorded by using a Lambda 750 UV-Vis spectrophotometer.

23

Fluorescence spectroscopy is a technique that measures the emission of light from a substance upon excitation. Fluorescence spectra of the QDs solution were collected by a setup from Ocean Optics with spectrometer of USB4000-UV-VIS-ES.

2.4.3. Photovoltaic characterization

Current-voltage characterization

The current-voltage (J-V) measurement is one of the most important characterization techniques for performance evaluation of solar cells, and from which the power conversion efficiency of solar cells can be determined. The J-V measurements are conducted by turning the external load from zero (short-circuit condition) to infinite load (open-circuit condition) under AM 1.5G illumination by a sun simulator. A typical J-V curve for QDSCs is presented in Figure 18.

Figure 18. A J-V curve of QDSCs based on CdSe QDs.

Photovoltaic parameters, including open-circuit voltage (Voc), short-circuit current (Jsc), fill factor (FF) and conversion efficiency (η), can be derived from the J-V curves. As shown in Figure 18, the value of Voc and Jsc is respectively obtained when the current is zero and voltage is zero. The generated maximum power point (Pmax) of the solar cell is found when the product of the photocurrent and photovoltage reach their maximum values (Jmax and Vmax). Then the overall conversion efficiency of the solar cell is determined by the

24

ratio of the obtainable maximum power to the power of incident light (Pin = 100 mW•cm-2 for AM 1.5G), by following equation 2.1,

𝜂 = =  ∙    ∙   (2.1)

in which FF is the ratio of the maximum power to the product of Jsc and Voc, as shown in equation 2.2.

𝐹𝐹 =  ∙    ∙  

(2.2)

The fill factors represent the deviation of the actual performance of solar cells from the theoretical maximum power output, and the values of the fill factor are always below unity, due to losses in the devices that result from charge carrier recombination, internal resistances and some other sources68.

The J-V characteristics were monitored under the illumination of a 300 W xenon lamp with the light intensity of 100 mW•cm-2 at AM 1.5 G solar light condition, with a Keithley model 2400 digital source meter.

Incident photon-to-current efficiency

The measure of incident photon-to-current efficiency (IPCE), also known as external quantum efficiency (EQE), is used to reveal how efficiently the incident photons convert to the charge carriers collected at the electrodes, and can be calculated by dividing the photocurrent generated under monochromatic illumination by the incident light:

𝐼𝑃𝐶𝐸  (𝜆) =  ( )  ( )  =   ∙ ∙  ( )

∙  ∙  ( )  =   ∙    ( )

 ∙  ( ) (2.3)

where Jsc (λ) and Pin (λ) are the measured photocurrent and intensity of the incident light at the wavelength of λ.

The IPCE of the QDSCs, is generally determined by the light harvesting efficiency (LHE) of the sensitizers, charge injection efficiency (ϕinj) and the collection efficiency (ηcc) in the devices:

25

𝐼𝑃𝐶𝐸 = 𝐿𝐻𝐸 ∙ 𝜙 ∙ 𝜂 (2.4)

If the effect of the light harvesting efficiency is ruled out, the absorbed photon-to-current efficiency (APCE) of the solar cells could be obtained, by following equation 2.5 below,

𝐴𝑃𝐶𝐸  (λ) = 𝜙   ∙  𝜂  =    ( )  ( )  =    ( )

  ( ) (2.5)

APCE is always employed to reveal how efficiently the absorbed photons are converted to electrons on the electrodes, and is more useful for the comparison of the charge injection and collection from the QDs.

IPCE measurements were performed on a computer-controlled set-up, which is composed of a Xenon lamp, a monochromator and a potentiosat.

Impedance spectroscopy

Impedance spectroscopy (IS) is a useful tool to analyze electron transfer processes in DSCs and QDSCs, from which the resistance of different charge transfer processes can be obtained and the chemical capacitance at the interfaces could be derived. Figure 19a shows the transmission line model for a DSC, which also can be applied to model the physical processes occurring in QDSCs.69

Figure 19. Transmission line model for DSCs and QDSCs (a) and Nyquist plot of CdSe based QDSCs under dark condition at -0.53V forward bias (b).

26

In Figure 19a, Rs is the series resistance for the transport resistance of the conducting glass and the connection setup. Rtr is the electron transport resistance through the mesoporous TiO2 layer, and Rrec is the recombination resistance at the TiO2/electrolyte interface, while Cμ represents the chemical capacitance that indicates the change of electron density as a function of the Fermi level. Zdiff is the diffusion impedance of the redox species in the electrolyte solution. RCE is the charge transfer resistance at the interface between the electrolyte and counter electrode, and CCE represents the corresponding interfacial capacitance. Accordingly, the first semicircle in the Nyquist plot of a QDSC (Figure 19b), which is in the high frequency region, arises from the charge transfer resistance for the reduction of the electrolyte at the counter electrode. The second semicircle, which is the most investigated one in this thesis, results from the charge recombination resistance at the photoanode/electrolyte interface. A bigger semicircle radius means less charge recombination in the device.

IS measurements were conducted with an impedance module from Autolab PGstat12 potentiosat.

27

3. Passivation strategies for liquid QDSCs

(Paper I and II)

3.1. Introduction

In QDSCs, the major two factors that can limit the performance of QDSCs are the low loading amount of colloidal QDs on the photoanodes and the recombination loss at surface traps of the QDs. Both of these two factors are related to the ligand which is capping on the QDs.

Generally, the pre-synthesized QDs are always capped by long hydrocarbon chains, which can stabilize the QDs from aggregation; however, these long alkyl chains will seriously hinder the deposition of QDs on the hydrophilic photoanode substrates. Mercaptopropionic acid (MPA), a small bifunctional organic molecule, is reported to be a promising alternative to the original long alkyl ligand, which can significantly enhance the loading amount of the QDs.53 Therefore, ligand exchange with MPA could be an effective way to increase the surface coverage of QDs on the photoanodes. However, ligand exchange only with MPA cannot contribute to suppress the surface defects on QDs, which is induced by imperfect surface atomic termination. In recent years, an atomic ligand passivation strategy has been established by Sargent’s group through using halide anions to reduce the surface defects of colloidal PbS QDs.58, 60 This method can also be applied in Cd-based QDs, with improved performance in QDSCs.15

Figure 20. Scheme of hybrid passivation to QDs.

28

Therefore, in this chapter, I will employ a hybrid passivation strategy to toxic CdSe QDs and low-toxic CuInS2 QDs, through ligand exchange by MPA and iodide ions at the same time (tetrabutylammonium iodide (TBAI) will be chosen as the source of iodide ions, shown in Figure 20), to improve the photovoltaic performance of these two kinds of QDs.

29

3.2. Hybrid passivation for colloidal CdSe based QDSCs (Paper I)

In the first paper, I employed hybrid passivation to CdSe QDs by MPA and iodide ions through a solution ligand exchange reaction to replace the original ligand OA. Ligand exchange only with MPA or iodide ions was conducted as comparison, and the CdSe QDs with original ligand OA were employed as reference.

Figure 21. EDS spectra (a) and FTIR spectra (b) of CdSe QDs with different passivation.

As shown in Figure 21a, I peaks and S peak can be detected from EDS spectra when the CdSe QDs was ligand exchanged only with iodide ions and MPA, respectively. While in hybrid passivated QDs, both I peaks and S peak can be observed, indicating that iodidie ions and MPA could be introduced to the surface of CdSe QDs by the ligand exchange reaction at the same time. FTIR spectra (figure 21b) reveal that the oleate ligand (C-H vibration between 2800 cm-1 and 3000 cm-1) on the QDs was significantly reduced by hybrid passivation and MPA passivation, demonstrating that the original ligand OA was replaced after both kinds of passivation, and that it can be removed more completely by the hybrid passivation. Consequently, we can conclude that CdSe QDs could be successfully hybrid passivated by MPA and iodide ions through the solution ligand exchange reaction.

30

Figure 22. UV-Vis absorption spectra of CdSe QDs-sensitized TiO2 electrodes based on different surface passivation of QDs. Insert: photographs of (a) OA-

CdSe/TiO2 film, (b) Iodide passivated-CdSe/TiO2 film, (c) MPA passivated-CdSe/TiO2 film, (d) Hybrid passivated-CdSe/TiO2 film.

To evaluate the effect of hybrid passivation to the loading efficiency of the CdSe QDs on TiO2 films, the absorbance of the sensitized electrodes was measured and is shown in Figure 22. In comparison with the original ligand OA, MPA passivation increases the absorbance of the CdSe sensitized TiO2 films, and hybrid passivation can further improve the loading efficiency of the CdSe QDs, which is also suggested by a more reddish color of the TiO2 film when sensitized with the hybrid passivated QDs. Due to the bifounctionality of MPA molecule, the QDs became water-soluble and were capable to bind to the TiO2 electrodes more efficiently through the carboxyl group of MPA, both of which facilitate the QDs with passivation that contains MPA component to achieve higher loading efficiency on the TiO2 electrodes. The more efficient deposition of the QDs with hybrid passivation probably results from the more complete removal of the original OA ligand.

The synthesized CdSe QDs with different passivation were employed in liquid QDSCs, to examine the influence of hybrid passivation on the photovoltaic performance of the CdSe QDs. Results from J-V characteristics are shown in Figure 23 and Table 1. It can be found that hybrid passivation can significantly improve the performance of the QDSCs, with higher Jsc (10.87 mA/cm2), FF value (0.585) and desirable Voc (0.52 V), and with an overall efficiency that amounts to 3.31%, which is 41% higher than that with the original ligand OA. On the other hand, Jsc of the cells from only MPA passivation just show slight

31

increment, while only iodide atomic passivation provide lower Jsc than that from OA. This variation in Jsc is consistent with the IPCE values of the devices, which are shown in Figure 24a. However, due to the different loading amounts of QDs with different passivation, APCE (shown in Figure 24b) should preferably be used to evaluate the charge injection and collection in the devices. After ruling out the effect of different absorbance of the electrodes, the APCE value of the solar cells based on hybrid passivated CdSe QDs remains the highest. Considering the fact that the surface states can significantly influence the electron injection and recombination process,70, 71 the higher APCE value of the devices with hybrid passivation may be attributed to a reduced amount of surface traps after that the CdSe QDs were hybrid passivated. The APCE value of the solar cells based on MPA capped CdSe QDs is similar to that of the original OA capped CdSe QDs, indicating that ligand exchange with MPA poses no obvious influence on the electron transfer efficiency. Using only atomic passivation provides the lowest APCE value among the tested solar cells. The reason might be the partial aggregation of the iodide ions passivated CdSe QDs (paper I, Figure S3), which could hinder the charge injection from the iodide passivated QDs to the TiO2 nanoparticles.

Figure 23. J-V curves of QDSCs based on CdSe QDs treated with different surface passivation.

32

Table 1. Photovoltaic parameters of QDSCs based on CdSe QDs treated with different surface passivation

Passivant Voc (V) Jsc (mA/cm2) FF η  (%)

OA 0.500 9.42 0.496 2.34

I 0.505 8.82 0.535 2.38

MPA 0.495 9.60 0.557 2.65

Hybrid 0.520 10.87 0.585 3.31

Figure 24. IPCE (a) and APCE (b) curves of QDSCs based on CdSe QDs treated

with different surface passivation.

In addition to the improved Jsc, gradually enhanced fill factor values can be observed from Table 1, which range from 0.496 for OA capped CdSe-QDSCs to 0.535 for the cells with iodide passivation, and then goes up to 0.557 and 0.585 for devices with MPA and hybrid passivation, respectively. This enhancement of FF value contributes to a gradually increased conversion efficiency of the devices. It is commonly believed that the recombination pathway can greatly influence the fill factor value of the QDSCs. With reduced trap density of the QDs and decreased recombination centers, the diffusion length of electrons will be improved and thus a notably higher fill factor value could be obtained.66 Following this claim, IS measurements under dark condition were performed to analyze the recombination resistance of the devices, and the results are shown in Figure 25. Consistent with the lowest FF

33

value, the recombination resistance in the cells based on OA capped CdSe QDs is the smallest, and then becomes larger when the QDs were modified with iodide ions, indicating that iodide ions can reduce the surface traps on the QDs and suppress the recombination in the devices to some extent. The recombination resistance further increases after the CdSe QDs were ligand exchanged by MPA, and an even higher FF value is obtained. The small size and high wettability nature of MPA, together with the hole extraction ability of the thiol group in MPA72, 73, both promote the hole extraction from the photoexcited QDs and reduce the possibility of charge recombination in the devices. When hybrid passivation is performed for CdSe QDs, MPA and iodide ions get bound to the surface of the QDs at the same time, not only facilitating the hole transfer, but also modifying the surface states and forming a more compact ligand layer.58 As a result, the solar cells based on hybrid passivated CdSe QDs exhibit the largest recombination resistance and the highest FF value.

Figure 25. Nyquist plots from IS characterization of QDSCs based on CdSe QDs treated with different surface passivation under dark condition at -0.53V forward

bias.

Though hybrid passivation can significantly improve the performance of the CdSe based QDSCs, the presence of too much TBAI during ligand exchange will decrease the efficiency of the solar cells, as shown in Figure 26 (Paper I, Figure 7 and Table 2), since it would interfere with the binding of MPA to the QDs, which is unfavorable for the adsorption of the QDs onto the TiO2 electrodes (Paper I, Figure 8). Therefore, the choice of the amount of TBAI is crucial for the optimal effect of the hybrid passivation.

34

Figure 26. J-V curves of QDSCs based on hybrid passivation with different amount of TBAI.

In conclusion, hybrid passivation with MPA and iodide ions is a useful way to improve the photovoltaic performance of QDSCs. MPA can increase the loading efficiency of colloidal CdSe QDs and promote the hole extraction from the QDs. Simultaneously, the iodide ions can eliminate the surface traps on the CdSe QDs and thus suppress the charge recombination in the devices. As a result, a significant enhancement of efficiency of the colloidal CdSe-QDSCs is achieved after the hybrid passivation treatment.

35

3.3. Strategies for photovoltaic performance improvement for colloidal CuInS2 QDs: hybrid passivation vs ZnS shell (Paper II)

In this section, hybrid passivation was applied to the low-toxic CuInS2 QDs to improve their photovoltaic performance. Traditional passivation with ZnS shell coating was also carried out for the CuInS2 QDs, and comparisons between these two kinds of passivation have been made. Zn-10 and Zn-20 in the following part refer to CuInS2 QDs with different ZnS shell thickness, when the shell coating time was 10 min and 20 min, respectively.

Figure 27. Fluorescence spectra of CuInS2 QDs solution and CuInS2 QDs with ZnS shell coating solution.

Fluorescence spectra was used to evaluate the passivation with ZnS shell coating to the pre-synthesized CuInS2 QDs. Results (Figure 27) show that the PL intensity of CuInS2 QDs was significantly improved by overcoating with a ZnS layer, and prolongation of the shell coating time could further increase the PL intensity. According to literature, surface states that cause non-radiative decay of the excited CuInS2 QDs could be minimized by shell coating, resulting in longer PL lifetime and higher PL intensity.23, 63, 64 Therefore, we suppose that ZnS shell coating can reduce the surface trap density on the pre-synthesized CuInS2 QDs, and that thicker ZnS shell could eliminate the surface defects more effectively. Because of the hole extracting effect of the thiol group in MPA, almost no fluorescence could be detected from the hybrid passivated CuInS2 QDs.

36

Figure 28. UV-Vis absorption spectra of sensitized TiO2 electrodes based on CuInS2 QDs with different passivation. The corresponding photographs are inserted.

The influence of different passivation strategies on the loading efficiency of the CuInS2 QDs was subsequently compared. Except for hybrid passivated QDs, all other kinds of CuInS2 QDs were ligand exchanged with MPA before depositing on the TiO2 electrodes to initially improve the loading amount of the CuInS2 QDs. As shown in Figure 28, the loading amount of the CuInS2 QDs with hybrid passivation is slightly higher than the untreated CuInS2 QDs, which is similar to the results for CdSe QDs, and can be attributed to the more complete removal of the original capping ligand on the QDs. In contrast, ZnS shell coating lowers the loading efficiency of the CuInS2 QDs, possibly due to that the size of the CuInS2 QDs were increased by the shell coating (Paper II, Figure 4, 8 nm and 8.5 nm for Zn-10 and Zn-20, respectively, while the size of the untreated CuInS2 QDs is around 7nm), which is unfavorable for the deposition of the QDs on the mesoporous TiO2 substrate. Therefore, hybrid passivation is more helpful for the deposition of CuInS2 QDs on TiO2 electrodes.

37

Figure 29. Nyquist plots from IS measurement of QDSCs based on CuInS2 QDs with different passivation under dark conditions at -0.55 V.

Figure 29 shows IS characterization results for the QDSCs based on CuInS2 QDs with different passivation. The recombination resistance of the solar cells based on the untreated CuInS2 QDs is the smallest. Treatment with hybrid passivation significantly increases the recombination resistance of the devices, indicating that the iodide ions in hybrid passivation can remedy the surface traps on the CuInS2 QDs and suppress the recombination occurring at the surface of the QDs. The recombination in the devices is further suppressed by ZnS shell coating, in which Zn-20 performs even better than Zn-10, demonstrating that the surface defects on the CuInS2 QDs could be eliminated by ZnS shell coating to a larger extent. Consequently, for the surface passivation of CuInS2 QDs, ZnS shell coating is a more effective approach than hybrid passivation.

Figure 30. J-V curves of QDSCs based on CuInS2 QDs with different passivation.

38

J-V characterization was finally carried out to evaluate the photovoltaic performance of the different kinds of CuInS2 QDs when employed in QDSCs. The J-V curves are shown in Figure 30 and the corresponding parameters from the characterization are summarized in Figure 31. From the overall conversion efficiency of the solar cells, it is clearly found that hybrid passivation can significantly enhance the efficiency of the solar cells, which amounts to 4.7%, while the solar cells based on ZnS shell coated CuInS2 QDs show inferior performance, compared with the devices with original CuInS2 QDs. Through comparison between the photovoltaic parameters of the solar cells, we find that the efficiency enhancement of the solar cells based on hybrid passivated CuInS2 QDs mainly results from the increased Jsc and Voc. The enhancement of Jsc can be attributed to the improved loading efficiency and promoted charge transfer efficiency (indicated by APCE values of the solar cells, which is shown in Paper II, Figure 5b) of the hybrid passivated QDs, and the increment of Voc could benefit from improved Jsc and suppressed recombination in the devices. For the passivation with ZnS shell coating, the photocurrent and photovoltage of the cells is dramatically decreased, which is probably caused by the inefficient loading of the ZnS shell coated CuInS2 QDs, and by the electron injection barrier between the CuInS2 QDs and TiO2 nanoparticles created by the ZnS shell.24 Though higher FF values are obtained by the solar cells with ZnS shell coating, which could be related to a more suppressed recombination at the surface of the QDs, the overall conversion efficiency of the solar cells is still lower than that based on the original CuInS2 QDs.

39

Figure 31. Photovoltaic parameters (PCE, Jsc, Voc and FF) from J-V

characterization of QDSCs based on CuInS2 QDs with different passivation.

Therefore, though ZnS shell coating can eliminate the surface defects of CuInS2 QDs more effectively, it will decrease the loading efficiency of the QDs and hinder the electron injection in the devices, which is unfavorable for the photovoltaic performance improvement of the CuInS2 QDs. In contrast, hybrid passivation can enhance the loading efficiency of the CuInS2 QDs, promote the electron transfer and suppress the recombination in the device, contributing to the performance enhancement of the QDSCs based on CuInS2 QDs.

40

3.4. Conclusions

Hybrid passivation with mercaptopropionic acid (MPA) and iodide ions is an effective strategy to improve the photovoltaic performance of colloidal QDs, both for toxic CdSe QDs and low-toxic CuInS2 QDs. The surface coverage of the colloidal QDs can be increased by MPA, which also promotes the hole extraction from the photo-excited QDs. At the same time, iodide ions can reduce the surface defects on the QDs and thus suppress the charge recombination in the devices. Although ZnS shell coating also is a good way to passivate the surface defects on CuInS2 QDs, it cannot effectively enhance the photovoltaic performance of CuInS2 QDs in QDSCs due to the less efficient loading of the shell coated QDs. Therefore, hybrid passivation is still a better way for improving the performance improvement of CuInS2 based QDSCs.

41

4. Improvement strategies for ss-QDSCs

(Paper III and IV)

4.1. Introduction

Although QDSCs have great potential for large-scale applications of photovoltaics, the risk of leakage and evaporation of the liquid electrolyte always limits the long-term stability of the solar cells. Therefore, solid state HTMs present a viable alternative to the liquid electrolyte in QDSCs, and to construct solid state QDSCs (ss-QDSCs). However, the most commonly used HTMs, such as Spiro-OMeTAD and P3HT, usually are organic molecules with big size, and will have problems of inferior pore filling74, 75, low conductivity and charge mobility76 in ss-QDSCs. Therefore, new strategies need to be developed to improve the performance of ss-QDSCs.

Comparing with liquid QDSCs, the active layer of ss-QDSCs has to be much thinner to alleviate the pore filling problem of HTMs in the metal oxide matrix. However, a thin active layer will dramatically reduce the loading amount of the colloidal QDs, which will limit the light absorption of the photoanode and result in low photocurrent and efficiency of the ss-QDSCs. In this chapter, the results of efforts to enhance the photocurrent by increasing the loading amount and optimizing the quality of the CuInS2 QDs will be shown.

In addition, compared to other HTMs, though Spiro-OMeTAD exhibit the best performance in ss-DSCs and PSCs, it has been demonstrated that the performance of ss-DSCs and PSCs is still limited by the low charge mobility of Spiro-OMeTAD. In this chapter, a new p-type dopant, silver bis(trifluoromethanesulfonyl)imide (AgTFSI), which was employed to enhance the conductivity of Spiro-OMeTAD and improve the conversion efficiency of the PSCs, will also be presented.

42

4.2. Efficiency enhancement of CuInS2 based ss-QDSCs (Paper III)

In this work, colloidal CuInS2 QDs were employed to construct low-toxic ss-QDSCs, with Spiro-OMeTAD as HTMs. Increasing the pore size of the TiO2 active layer was tried to promote the loading amount of the QDs. The quality of the QDs was optimized to increase the quantum yield of the QDs and suppress the recombination in the devices.

Figure 32. Energy level diagram of TiO2, as-synthesized CuInS2 QDs and Spiro-OMeTAD in the ss-QDSCs

Due to the I-III-VI nature, CuInS2 QDs always exhibit off-stoichiometry, and that the composition of CuInS2 will alter its band gap. From XPS results (Paper III, Figure 2), we find that the stoichiometric ratio of Cu : In : S in our synthesized CuInS2 QDs is 1 : 2.53 : 4.14, which implies the existence of Cu vacancies in the QDs. Combined with DFT calculations and the effective-mass approximation, the calculated valence band of the as-synthesized QDs locates at -5.63 eV, and the resultant conduction band resides at -3.95 eV. As shown in Figure 32, the difference between the conduction bands of CuInS2 QDs and TiO2 is 0.25 eV, facilitating the photo-generated electrons from the QDs injecting into the TiO2 nanoparticles. Meanwhile, the energy gap between the valance band of CuInS2 QDs and HOMO of Spiro-OMeTAD, which amounts to 0.51 eV, also favors the hole transfer in the device. Therefore, the band positions of our CuInS2 QDs is energetically favorable for the charge transfer in the solar cells, making these CuInS2 QDs theoretically suitable as sensitizers in ss-QDSCs.

43

Figure 33. TEM images of CuInS2 QDs sensitized 20 nm TiO2 nanoparticles (a) and 30 nm TiO2 nanoparticles (b), and corresponding absorption spectra of the CuInS2

QDs sensitized TiO2 photoanodes (c).

Due to the limitation of thin active layer in ss-QDSCs, combining with the low loading nature of colloidal QDs, the deposition amount of CuInS2 QDs on TiO2 photoelectrodes is really unsatisfactory. In order to increase the loading amount of the QDs, paste with larger TiO2 nanoparticles (30 nm) was used to enlarge the pore size of the TiO2 active layer during the fabrication of the photoanodes. Compared to the most used paste with 20 nm TiO2 nanoparticles, the surface coverage of CuInS2 QDs on the photoanodes with 30 nm TiO2 nanoparticles was significantly improved, which can be observed by the TEM image in Figure 33b, and the absorbance of the sensitized electrodes was dramatically enhanced (Figure 33c). The improved loading amount of the QDs leads to an increased IPCE value and photocurrent of the resultant solar cells, contributing to an enhanced efficiency of the devices (Paper III, Figure 5).

Figure 34. Fluorescence spectra of different CuInS2 QDs - QD0, QD1, QD2.

44

Though the loading amount of CuInS2 QDs was enhanced for the photoanodes with larger pore size, the photocurrent and efficiency of the solar cells was still low. Optimizing the quality of the QDs to improve their quantum yield and suppress the recombination in the devices could be another approach to enhance the performance of the solar cells. Therefore, some synthesis parameters was changed, trying to get optimized CuInS2 QDs. For convenience, the original synthesized CuInS2 QDs will be referred as QD0. We shortened the reaction time to get smaller QDs (QD1), which may further promote the loading amount of the QDs and increase the driving force of electron injection from the QDs to TiO2. However, though QD1 indeed exhibit smaller size (5 nm, shown in Paper III, Figure S4), the absorbance of the resultant sensitized photoanodes was not improved (Paper III, Figure S5). QD2 are CuInS2 QDs synthesized with a smaller Cu/In ratio, which can create larger Cu deficiency in the QDs and then could influence the photoelectric property of the CuInS2 nanostructures.63, 77 From Figure 3, we can find that compared to QD0, the emission peak of QD1 shows slight blueshift, which is consistent with the smaller size of QD1; however, the PL intensity of QD1 is lower, implying that QD1 may not be ripened well, and that there might be numerous surface defects on the QDs. In contrast, QD2 presents a dramatically increased PL intensity with blueshifted emission. This phenomenon could be ascribed to the larger Cu deficiency that is reported to promote the internal-related emission process and widen the band gap of the CuInS2 QDs.63, 64, 77, 78

Figure 35. IPCE spectra (a) and J-V curves (b) of ss-QDSCs based on different kinds of CuInS2 QDs - QD0, QD1, QD2, and hybrid passivated QD2.

45

Figure 35 and Table 2 present the photovoltaic performance of the ss-QDSCs based on QD0, QD1, QD2 and QD2 with hybrid passivation (this will be discussed later). Owing to the extensive surface defects on QD1, which can hinder the charge injection from the QDs and cause serious recombination at the interfaces, both of the IPCE value and Jsc of the cells based on QD1 are lower than that based on QD0, together with the inferior Voc and FF value, leading to much lower conversion efficiency. In contrast, the performance of the solar cells based on QD2 is significantly improved, resulting from enhanced Jsc, higher Voc and increased FF value. According to literature, the larger Cu deficiency in QD2 could promote the internal-related emission and suppress the surface-related recombination, thus a prolonged emission lifetime and a higher emission yield could be obtained.63, 64 Furthermore, due to the domination of internal and surface defects of the CuInS2 QDs, the trend of charge separation yield in these QDs is parallel to their emission yield trend, thus enhanced photocurrent and reduced recombination in the devices could be achieved in QD2.23 Therefore, promoted internal defects and reduced surface defects could be a good way to optimize the quality of CuInS2 QDs, and be beneficial to the performance of CuInS2 based ss-QDSCs.

Table 2. Photovoltaic parameters of ss-QDSCs based on different kinds of CuInS2 QDs - QD0, QD1, QD2, and hybrid passivated QD2.

sample Voc (V) Jsc (mA/cm2) FF η  (%)

QD0 0.530 3.39 0.395 0.71

QD1 0.485 1.99 0.373 0.36

QD2 0.565 4.09 0.443 1.02

QD2-passivation 0.560 4.44 0.498 1.24

Though surface defects on CuInS2 QDs can be eliminated to some extent by the larger Cu deficiency in QD2, they are still inevitable owing to the imperfect surface atomic termination. In order to further reduce the surface traps, hybrid passivation was applied to QD2, which offers further increase of the Jsc and FF value, contributing to an improved PCE of 1.24%. The surface defect elimination by hybrid passivation can also be observed from the IS characterization, as shown in Figure 36. The recombination resistance of the

46

solar cells is firstly enlarged by the larger Cu deficiency in QD2, and then the recombination in the devices is further suppressed by hybrid passivation of QD2, demonstrating that the surface defects of the QDs is suppressed step by step. Similar with previous results, the gradually increased FF value of the solar cells based on different CuInS2 QDs is consistent with the increasing tendency of recombination resistance in the devices, benefitting from the suppressed surface defects.

Figure 36. Nyquist plots from IS measurements of the ss-QDSCs based on different kinds of CuInS2 QDs - QD0, QD1, QD2, and hybrid passivated QD2.

Figure 37. Durability of ss-QDSCs based on CuInS2 QDs.

47

Finally, the stability of the CuInS2 based ss-QDSCs was evaluated, and the results are shown in Figure 37. After aging in the open air for 380 hours, the Voc and FF almost keep the same level (decrease is less than 10%), and the Jsc just show slight decrease, resulting in the degradation of only 27% for the efficiency of the solar cells. Comparing with the liquid QDSCs, the ss-QDSCs presented significantly improved stability.

Therefore, increasing the pore size of the TiO2 active layer, which can enhance the loading amount of the QDs, and optimizing the quality of the CuInS2 QDs, that can suppress the surface defects and promote the quantum yield of the QDs, could be two effective approaches to improve the photocurrent and conversion efficiency of ss-QDSCs based on CuInS2 QDs.

48

4.3. AgTFSI as p-type dopant for perovskite solar cells (Paper IV)

In this part, AgTFSI (shown in Figure 38a) was used as a p-type dopant for Spiro-OMeTAD in PSCs, to improve the conductivity of Spiro-OMeTAD and the photovoltaic performance of the PSCs.

Figure 38. (a) Molecular structure of AgTFSI, and photographs of Spiro-MeOTAD films with different (molar) doping ratios of AgTFSI. (b) J–V curves of Spiro-

MeOTAD films with different doping ratios of AgTFSI. (c) Conductivity of Spiro- MeOTAD films with different doping concentration of AgTFSI.

From figure 38a, we can find that the color of Spiro-OMeTAD films become gradually darker when increased concentration of AgTFSI was applied, something that can be ascribed to that the Spiro-OMeTAD was gradually oxidized in the film. As determined in the literature, the oxidation potential of Spiro-OMeTAD is 0.63V vs. NHE,76 and that of Ag+ is 0.8 V vs. NHE. The potential gap may provide a driving force for a one-electron oxidation reaction between Spiro-OMeTAD and Ag+, as

49

As a result, an organic salt of Spiro-OMeTAD+TFSI- is formed, which leads to increased charge carrier density and reduced series resistances in the electrodes, resulting in improved conductivity of the electrodes. As shown in figure 38b and 38c, the conductivity of Spiro-OMeTAD films are significantly increased upon addition of AgTFSI, and further increased with the increasing doping ratio, indicating that AgTFSI could be a p-type dopant for the conductivity improvement of Spiro-OMeTAD.

Figure 39. J–V characteristic curves (a) and IPCE spectra (b) of the perovskite-based devices obtained using different doping substances.

Then this kind of p-type dopant was applied in the most popular PSCs, which is based on the inorganic nanocrystals - perovskites. The commonly oxygen doped devices were employed as a reference. The J-V curves and corresponding IPCE spectra are presented in Figure 39, with the photovoltaic parameters are summarized in Table 3. Compared to the oxygen doping, the devices with AgTFSI doping shows an enhanced conversion efficiency of 12.0%, and that this enhancement mainly originates from the increment of FF value. Due to the improved conductivity of Spiro-OMeTAD in the devices by AgTFSI doping, the series resistance of the solar cells is significantly reduced, thus a higher FF value can be obtained.

50

Table 3. Photovoltaic parameters for PSCs based on different doping methods.

Doping Method Voc (V) Jsc (mA/cm2) FF η  (%)

AgTFSI 0.92 20.95 0.62 12.0

oxygen 0.92 19.95 0.55 10.1

Therefore, this commercial silver based organic salt, AgTFSI can be an effective p-type dopant for the most used HTM - Spiro-OMeTAD, which can increase the conductivity of Spiro-OMeTAD film, and then improve the performance of the PSCs based on this kind of HTM.

4.4. Conclusions

For ss-QDSCs, increasing the pore size of the TiO2 active layer can enhance the loading amount of the QDs, and optimization of the quality of CuInS2 QDs through controlling Cu deficiency and hybrid passivation can effectively improve the quantum yield of the QDs. Both ways contribute to a significant enhancement of the photocurrent of the ss-QDSCs. In addition, AgTFSI can be used as an effective p-type dopant for improving the conductivity of Spiro-OMeTAD, which is also a useful way to improve the performance of solid state solar cells.

51

5. Concluding Remarks

The aim of this thesis was to develop efficient, stable and green quantum dot sensitized solar cells (QDSCs). The development of strategies for improving the performance of the QD solar cells, the employment of high-quality and low-toxic QDs in QDSCs and the evolution of efficient solid state QDSCs (ss-QDSCs) were the main tasks of this thesis.

From Chapter 3 and 4 (papers I, II and III), I found that hybrid passivation was an effective strategy to improve the photovoltaic performance of QDs, both for toxic CdSe QDs and low-toxic CuInS2 QDs, and both for liquid and solid state QDSCs. By increasing the loading amount of the QDs on the photoelectrodes and reducing the surface defects of the QDs, hybrid passivation can effectively enhance the photocurrent and fill factor of the solar cells, leading to a significant improvement of the device performance.

The photovoltaic performance of CuInS2 based solar cells can be improved by properly increasing the Cu deficiency in CuInS2 QDs (paper III). Larger Cu deficiency can promote internal defects and suppress surface defects in the QDs, resulting in a higher quantum yield of the CuInS2 QDs.

For ss-QDSCs, increasing the pore size of the TiO2 active layer by larger TiO2 nanoparticles can efficiently enhance the loading amount of the QDs and improve the pore filling of the HTMs (paper III). AgTFSI was found to be an effective p-type dopant for Spiro-OMeTAD in solid state perovskite solar cells (paper IV).

52

6. Future Outlook

Although the performance of QDSCs has been significantly improved in the late five years, the efficiency of QDSCs can still not meet the requirements for large-scale applications, especially so for the stable ss-QDSCs. Future efforts should still focus on the development of ss-QDSCs based on low-toxic QDs, and can be made by considering the following aspects:

Substrates for photoanodes: from previous results, we found that the photocurrent of ss-QDSCs was much lower than that of liquid QDSCs. This can be mainly ascribed to the low loading amount of QDs on the photoanodes, resulting from the thinner TiO2 active layer in the ss-QDSCs. Optimizing the TiO2 paste to get suitable pore size of the photoanodes and increase the loading ability of the substrates could be a good way. In addition, ZnO nanorods can also be tried as substrates for ss-QDSCs, which can facilitate the penetration both for HTMs and QDs.

Quality of QDs: the quality of QDs can dramatically influence the performance of solar cells, through affecting Jsc, Voc and FF at the same time. Firstly, in addition to the CuInS2 QDs that are investigated in the thesis, low-toxic QDs with smaller bandgap could be chosen as photosensitizers in ss-QDSCs, such as CuInSe QDs and ZnCuInSe QDs, to increase the light harvesting of the photoanodes. The size of the QDs could also be optimized for the ss-QDSCs, to get high loading efficiency without significantly decreasing the light harvesting capability. Last but not the least, more efficient passivation methods are still needed to be explored, to further improve the quality of QDs and suppress the charge recombination in the solar cells.

Choice of HTMs: though Spiro-OMeTAD can work with QDs in ss-QDSCs, their comparatively big size restrict the thickness of the photoanode. In addition, the manufacturing cost of Spiro-OMeTAD also limits its large-scale application. Therefore, small molecular HTMs should also be tried in ss-QDSCs.

53

With the improvement of the each component in ss-QDSCs, I believe that the efficiency of ss-QDSCs can become comparable with liquid QDSCs and DSCs, and that they finally can contribute to the commercialization of the third generation solar cells.

54

Acknowledgements

I would like to express my sincere gratitude to the people who greatly contributed to my PhD study in the last four years:

Professor Hans Ågren, my main supervisor, for providing me the opportunity to pursue my PhD study in our wonderful department. I really appreciate your professional guidance on my thesis work, endless patience and constant encouragement on my experimental work and manuscripts writing. I also would like to thank for your support on allowing me to attend international conferences.

Professor Licheng Sun, my co-supervisor, for your kind support and collaboration on the characterization work in my study.

My special thanks also go to Dr. Chunze Yuan, for guiding me to make the first piece of solar cell, and offering me lots of advice and discussion on my experiments.

Many thanks to Dr. Bo Xu, Dr. Hong Cheng, Dr. Yonfei Ji, Dr. Shizhao Xiong (Chalmers), Dr. Asghar Jamshidi Zavaraki, for your all nice collaboration on my PhD studies.

I also want to express my appreciation to the great help and discussion on my study from Dr. Xin Li, Dr. Na Kong, Dr. Cheng Chen, Dr. Lei Wang, Dr. Miao Yang, Jiajia Gao, Jinbao Zhang (UU).

My gratitude also goes to the members in the Department of Theoretical Chemistry and Biology, who give me advice on everything and create a pleasant working environment: Professor Hans Ågren, Professor Boris Minaev, Professor Faris Gel'mukhanov, Professor Yi Luo, Professor Olav Vahtras, Professor Lars Thylen, Dr. Mårten Ahlquist, Dr. Yaoquan Tu, Dr. Zilvinas Rinkevicius, Dr. Arul Murugan, Dr. Stefan Knippenberg, Dr. Viktor Kimberg, Dr. Xin Li, Dr. Jaime Axel Rosal Sandberg, Dr. Chunze Yuan, Dr. Weijie Hua, Dr. Yongfei Ji, Dr. Lijun Liang, Dr. Rocio Marcos, Dr. Sai Duan, Dr. Guanjun Tian, Dr. Ting Fan, Dr. Yan Wang, Dr. Xianqiang Sun, Dr. Xinrui Cao, Dr.

55

Liqin Xue, Dr. Ying Wang, Dr. Quan Miao, Ignat Harczuk, Guanglin Kuang, Xu Wang, Wei Hu, Zhengzhong Kang, Vinicius Vaz de Cruz, Rafael Carvalho Couto, Junfeng Li, Rongfeng Zou, Shaoqi Zhan, Zhen Xie, Matti Vapa, Tuomas Loytynoja, Nina Bauer.

Big thanks to Nina Bauer for the administrative help during all the four years.

In addition, I would like to give my thanks to Yan Wang and all my friends who live in Kungshamra, for always giving me joyful time and making me enjoy the life in Sweden.

My thanks also go to China Scholarship Council (CSC), for the financial support in the last four years.

Finally, I want to express my deep gratitude to my husband Bo Xu and my parents, Weixing Huang and Dongying Zhou, for your endless love and for always supporting my study and life.

56

Appendix A

The following is a description of my contribution to Publications I to IV, as requested by KTH. Paper I. I undertook all the quantum dots synthesis, device fabrication, major characterization and analysis work, and wrote the manuscript. Paper II. I undertook all the quantum dots synthesis, device fabrication, major characterization and analysis work, and wrote the manuscript. Paper III. I undertook all the quantum dots synthesis, major device fabrication, major characterization and analysis work, and wrote the manuscript. Paper IV. I took part of the work on device fabrication, characterization and discussion.

57

Appendix B

1. J. Jean, P. R. Brown, R. L. Jaffe, T. Buonassisi and V. Bulovic, Energy Environ. Sci., 2015, 8, 1200-1219. 2. D. M. Chapin, C. S. Fuller and G. L. Pearson, J. Appl. Phys., 1954, 25, 676-677. 3. H. S. Mansur, Wiley Interdisciplinary Reviews: Nanomedicine and Nanobiotechnology, 2010, 2, 113-129. 4. S. E. Lohse and C. J. Murphy, J. Am. Chem. Soc., 2012, 134, 15607-15620. 5. J. Q. Grim, L. Manna and I. Moreels, Chem. Soc. Rev., 2015, 44, 5897-5914. 6. G. H. Carey, A. L. Abdelhady, Z. Ning, S. M. Thon, O. M. Bakr and E. H. Sargent, Chem. Rev., 2015, 115, 12732-12763. 7. G. Chen, J. Seo, C. Yang and P. N. Prasad, Chem. Soc. Rev., 2013, 42, 8304-8338. 8. Y. Wang and N. Herron, The Journal of Physical Chemistry, 1991, 95, 525-532. 9. M. C. Beard, J. M. Luther, O. E. Semonin and A. J. Nozik, Acc. Chem. Res., 2013, 46, 1252-1260. 10. A. J. Nozik, Physica E: Low-dimensional Systems and Nanostructures, 2002, 14, 115-120. 11. R. J. Ellingson, M. C. Beard, J. C. Johnson, P. Yu, O. I. Micic, A. J. Nozik, A. Shabaev and A. L. Efros, Nano Lett., 2005, 5, 865-871. 12. O. E. Semonin, J. M. Luther, S. Choi, H.-Y. Chen, J. Gao, A. J. Nozik and M. C. Beard, Science, 2011, 334, 1530-1533. 13. M. C. Hanna and A. J. Nozik, J. Appl. Phys., 2006, 100, 074510. 14. H. Lee, H. C. Leventis, S.-J. Moon, P. Chen, S. Ito, S. A. Haque, T. Torres, F. Nüesch, T. Geiger, S. M. Zakeeruddin, M. Grätzel and M. K. Nazeeruddin, Adv. Funct. Mater., 2009, 19, 2735-2742. 15. M. S. de la Fuente, R. S. Sánchez, V. González-Pedro, P. P. Boix, S. G. Mhaisalkar, M. E. Rincón, J. Bisquert and I. Mora-Seró, J. Phys. Chem. Lett., 2013, 4, 1519-1525. 16. J. Huang, B. Xu, C. Yuan, H. Chen, J. Sun, L. Sun and H. Ågren, ACS Appl. Mater. Interfaces, 2014, 6, 18808-18815. 17. H. Lee, M. Wang, P. Chen, D. R. Gamelin, S. M. Zakeeruddin, M. Grätzel and M. K. Nazeeruddin, Nano Lett., 2009, 9, 4221-4227.

58

18. Z. Pan, H. Zhang, K. Cheng, Y. Hou, J. Hua and X. Zhong, ACS Nano, 2012, 6, 3982-3991. 19. Z. Pan, K. Zhao, J. Wang, H. Zhang, Y. Feng and X. Zhong, ACS Nano, 2013, 7, 5215-5222. 20. G. Niu, L. Wang, R. Gao, B. Ma, H. Dong and Y. Qiu, J. Mater. Chem., 2012, 22, 16914-16919. 21. N. Guijarro, T. Lana-Villarreal, T. Lutz, S. A. Haque and R. Gómez, J. Phys. Chem. Lett., 2012, 3, 3367-3372. 22. J. W. Lee, D. Y. Son, T. K. Ahn, H. W. Shin, I. Y. Kim, S. J. Hwang, M. J. Ko, S. Sul, H. Han and N.-G. Park, Sci. Rep., 2013, 3. 23. D. H. Jara, S. J. Yoon, K. G. Stamplecoskie and P. V. Kamat, Chem. Mater., 2014, 26, 7221-7228. 24. Z. Pan, I. Mora-Seró, Q. Shen, H. Zhang, Y. Li, K. Zhao, J. Wang, X. Zhong and J. Bisquert, J. Am. Chem. Soc., 2014, 136, 9203-9210. 25. W. Li, Z. Pan and X. Zhong, J. Mater. Chem. A, 2015, 3, 1649-1655. 26. H. McDaniel, N. Fuke, N. S. Makarov, J. M. Pietryga and V. I. Klimov, Nat Commun, 2013, 4, 2887. 27. 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, 138, 4201-4209. 28. Y. Tachibana, H. Y. Akiyama, Y. Ohtsuka, T. Torimoto and S. Kuwabata, Chemistry Letters, 2007, 36, 88-89. 29. V. Chakrapani, D. Baker and P. V. Kamat, J. Am. Chem. Soc., 2011, 133, 9607-9615. 30. L. Li, X. Yang, J. Gao, H. Tian, J. Zhao, A. Hagfeldt and L. Sun, J. Am. Chem. Soc., 2011, 133, 8458-8460. 31. Z. Ning, H. Tian, C. Yuan, Y. Fu, L. Sun and H. Ågren, Chemistry – A European Journal, 2011, 17, 6330-6333. 32. Z. Yang, C. Chen, C. Liu, C. L. Li and H. T. Chang, Adv. Energy Mater., 2011, 1, 259-264. 33. M. Deng, S. Huang, Q. Zhang, D. Li, Y. Luo, Q. Shen, T. Toyoda and Q. Meng, Chem. Lett., 2010, 39, 1168-1170. 34. J. Xu, H. Xue, X. Yang, H. Wei, W. Li, Z. Li, W. Zhang and C. S. Lee, Small, 2014, 10, 4754-4759. 35. J. Y. Chang, S. C. Chang, S. H. Tzing and C. H. Li, ACS Appl. Mater. Interfaces, 2014, 6, 22272-22281.

59

36. S. H. Im, H.-J. Kim, S. W. Kim, S.-W. Kim and S. I. Seok, Energy Environ. Sci., 2011, 4, 4181-4186. 37. S. Kim, S. H. Im, M. Kang, J. H. Heo, S. I. Seok, S.-W. Kim, I. Mora-Sero and J. Bisquert, Phys. Chem. Chem. Phys., 2012, 14, 14999-15002. 38. T. P. Brennan, P. Ardalan, H.-B.-R. Lee, J. R. Bakke, I. K. Ding, M. D. McGehee and S. F. Bent, Adv. Energy Mater., 2011, 1, 1169-1175. 39. C.-F. Chi, P. Chen, Y.-L. Lee, I. P. Liu, S.-C. Chou, X.-L. Zhang and U. Bach, J. Mater. Chem., 2011, 21, 17534-17540. 40. I. Barcelo, J. M. Campina, T. Lana-Villarreal and R. Gomez, Phys. Chem. Chem. Phys., 2012, 14, 5801-5807. 41. J. A. Christians and P. V. Kamat, ACS Nano, 2013, 7, 7967-7974. 42. S.-J. Moon, Y. Itzhaik, J.-H. Yum, S. M. Zakeeruddin, G. Hodes and M. Grätzel, J. Phys. Chem. Lett., 2010, 1, 1524-1527. 43. J. A. Chang, S. H. Im, Y. H. Lee, H.-j. Kim, C.-S. Lim, J. H. Heo and S. I. Seok, Nano Lett., 2012, 12, 1863-1867. 44. J. A. Chang, J. H. Rhee, S. H. Im, Y. H. Lee, H.-j. Kim, S. I. Seok, M. K. Nazeeruddin and M. Gratzel, Nano Lett., 2010, 10, 2609-2612. 45. M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami and H. J. Snaith, Science, 2012, 338, 643-647. 46. A. Kojima, K. Teshima, Y. Shirai and T. Miyasaka, J. Am. Chem. Soc., 2009, 131, 6050-6051. 47. 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. Grätzel and N.-G. Park, Sci. Rep., 2012, 2, 591. 48. D. Bi, W. Tress, M. I. Dar, P. Gao, J. Luo, C. Renevier, K. Schenk, A. Abate, F. Giordano, J.-P. Correa Baena, J.-D. Decoppet, S. M. Zakeeruddin, M. K. Nazeeruddin, M. Grätzel and A. Hagfeldt, Science Advances, 2016, 2, e1501170. 49. G. S. He, L.-S. Tan, Q. Zheng and P. N. Prasad, Chem. Rev., 2008, 108, 1245-1330. 50. J. Park, J. Joo, S. G. Kwon, Y. Jang and T. Hyeon, Angew. Chem. Int. Ed., 2007, 46, 4630-4660. 51. S. G. Kwon, Y. Piao, J. Park, S. Angappane, Y. Jo, N.-M. Hwang, J.-G. Park and T. Hyeon, J. Am. Chem. Soc., 2007, 129, 12571-12584. 52. H. Zhang, K. Cheng, Y. M. Hou, Z. Fang, Z. X. Pan, W. J. Wu, J. L. Hua and X. H. Zhong, Chem. Commun., 2012, 48, 11235-11237.

60

53. W. Li and X. Zhong, J. Phys. Chem. Lett., 2015, 6, 796-806. 54.  A.  Zaban,  O.  I.  Mićić,  B.  A.  Gregg  and  A.  J.  Nozik,  Langmuir, 1998, 14, 3153-3156. 55. S. Jiao, Q. Shen, I. Mora-Seró, J. Wang, Z. Pan, K. Zhao, Y. Kuga, X. Zhong and J. Bisquert, ACS Nano, 2015, 9, 908-915. 56. J. Wang, I. Mora-Seró, Z. Pan, K. Zhao, H. Zhang, Y. Feng, G. Yang, X. Zhong and J. Bisquert, J. Am. Chem. Soc., 2013, 135, 15913-15922. 57. K. Zhao, Z. Pan, I. Mora-Seró, E. Cánovas, H. Wang, Y. Song, X. Gong, J. Wang, M. Bonn, J. Bisquert and X. Zhong, J. Am. Chem. Soc., 2015, 137, 5602-5609. 58. A. H. Ip, S. M. Thon, S. Hoogland, O. Voznyy, D. Zhitomirsky, R. Debnath, L. Levina, L. R. Rollny, G. H. Carey, A. Fischer, K. W. Kemp, I. J. Kramer, Z. Ning, A. J. Labelle, K. W. Chou, A. Amassian and E. H. Sargent, Nat. Nanotech., 2012, 7, 577-582. 59. Z. Ning, Y. Ren, S. Hoogland, O. Voznyy, L. Levina, P. Stadler, X. Lan, D. Zhitomirsky and E. H. Sargent, Adv. Mater., 2012, 24, 6295-6299. 60. J. Tang, K. W. Kemp, S. Hoogland, K. S. Jeong, H. Liu, L. Levina, M. Furukawa, X. Wang, R. Debnath, D. Cha, K. W. Chou, A. Fischer, A. Amassian, J. B. Asbury and E. H. Sargent, Nat Mater, 2011, 10, 765-771. 61. M. Liu, F. P. G. de Arquer, Y. Li, X. Lan, G.-H. Kim, O. Voznyy, L. K. Jagadamma, A. S. Abbas, S. Hoogland, Z. Lu, J. Y. Kim, A. Amassian and E. H. Sargent, Adv. Mater., 2016, DOI: 10.1002/adma.201506213. 62. S. Mathew, A. Yella, P. Gao, R. Humphry-Baker, F. E. CurchodBasile, N. Ashari-Astani, I. Tavernelli, U. Rothlisberger, K. NazeeruddinMd and M. Grätzel, Nat. Chem., 2014, 6, 242-247. 63. Y.-K. Kim, S.-H. Ahn, K. Chung, Y.-S. Cho and C.-J. Choi, J. Mater. Chem., 2012, 22, 1516-1520. 64. L. Li, A. Pandey, D. J. Werder, B. P. Khanal, J. M. Pietryga and V. I. Klimov, J. Am. Chem. Soc., 2011, 133, 1176-1179. 65. S. M. Thon, A. H. Ip, O. Voznyy, L. Levina, K. W. Kemp, G. H. Carey, S. Masala and E. H. Sargent, ACS Nano, 2013, 7, 7680-7688. 66. D. Zhitomirsky, O. Voznyy, L. Levina, S. Hoogland, K. W. Kemp, A. H. Ip, S. M. Thon and E. H. Sargent, Nat Commun, 2014, 5, 3803. 67. L. Liao, H. Zhang and X. Zhong, J. Lumin., 2011, 131, 322-327. 68. P. V. Kamat, K. Tvrdy, D. R. Baker and J. G. Radich, Chem. Rev., 2010, 110, 6664-6688.

61

69. V. González-Pedro, X. Xu, I. Mora-Seró and J. Bisquert, ACS Nano, 2010, 4, 5783-5790. 70. J. Z. Zhang, Acc. Chem. Res., 1997, 30, 423-429. 71. F. Hetsch, X. Xu, H. Wang, S. V. Kershaw and A. L. Rogach, J. Phys. Chem. Lett., 2011, 2, 1879-1887. 72. Y. Tan, S. Jin and R. J. Hamers, J. Phys. Chem. C, 2012, 117, 313-320. 73. C. Bullen and P. Mulvaney, Langmuir, 2006, 22, 3007-3013. 74. B. Li, L. Wang, B. Kang, P. Wang and Y. Qiu, Sol. Energy Mater. Sol. Cells, 2006, 90, 549-573. 75. L. Schmidt-Mende and M. Grätzel, Thin Solid Films, 2006, 500, 296-301. 76. B. Xu, H. Tian, D. Bi, E. Gabrielsson, E. M. J. Johansson, G. Boschloo, A. Hagfeldt and L. Sun, J. Mater. Chem. A, 2013, 1, 14467-14470. 77. M. Uehara, K. Watanabe, Y. Tajiri, H. Nakamura and H. Maeda, J. Chem. Phys., 2008, 129, 134709. 78. Y. Hamanaka, T. Kuzuya, T. Sofue, T. Kino, K. Ito and K. Sumiyama, Chem. Phys. Lett., 2008, 466, 176-180.