Metal-Oxide Based Interlayers for Organic and Perovskite Photovoltaics

Mehrad Ahmadpour

University of Southern Denmark

Faculty of Engineering

Mads Clausen Institute

A thesis submitted for the degree of

Doctor of Philosophy

September 2017

Abstract

Thin-film solar cells including both organic (OSC) and perovskites (PSC) devices have emerged as an excellent alternative to the traditional silicon wafer technology in the field of photovoltaic technologies. The basic materials used in these devices provides the solar cells with unique properties such as low weight, semi-transparency, mechanical flexibility and potentially low cost, which in turn opens up for completely new application areas. OSC has recently achieved Power Conversion Efficiencies (PCE) of more than 13%, and PSC has shown an outstanding PCE of more than 22%, whereas in both cases, their stability still lacks significant improvements.

This work is dedicated to research on improving the performance, including device stability, of organic and perovskite solar cells, using novel metal-oxide based interlayers. The thesis will initial focus on the fabrication and optimization of DBP/C70 organic cells, where new developments towards integration of OSC for usage in low-power consuming electronics will be shown. Specifically it is shown how efficient OSC reaching very high output voltage above 6V can be developed by multi-stacked devices, applying efficient interfaces between each sub-cell, and by optimized the thickness in the individual sub-cells. Integration of novel reactive sputtered Molybdenum oxide (MoOx) layers will be demonstrated as a new method to improve the stability of these organic solar cells, while maintaining a high device performance, compared to cells made thermal evaporation of MoOx hole transport layer. This is obtained by development of crystalline films with high work functions, which in the end lead to the improved device stability.

Based on the analysis on OSC devices, new metal-oxide based interlayers for PSC devices are also investigated. Especially adaptation of NiO:Cu hole-transport layers are investigated alongside solvent engineering approaches for developing high performing PSC devices. At the end, it is show that copper in combination with specific electron transport layers leads to PSC with high performance and improved device stability.

Resume

Tyndfilms solceller inklusiv bade organiske (OSC) og perovskite (PSC) devices er blevet til lovende alternativer til den traditionelle silicium wafer teknologi inden fotovoltaiske enheder. De basale materialer der anvendes i disse devices giver solcellerne unikke egenskaber såsom lav vægt, gennemsigtighed, mekanisk fleksibilitet og potential billig pris, hvilket giver anledning til helt nye applikationsområder. OSC har for nyligt opnået effektiviteter på mere end 13% of PSC har vist fremragende effektiviteter på mere end 22%, I begge teknologier halter stabiliteten dog stadig efter.

Dette arbejde er rettet mod forskning omkring forbedring af ydeevne, inklusiv stabilitet, af organiske og perovskite solceller ved brug af nye metal-oxid baserede interfacelag. Denne afhandling vil I begyndelsen have fokus på fabrikation og optimering af DBP/C70 organiske solceller, hvor ny udvikling indenfor integration af organiske solceller til forsyning lav effekt elektronik vil blive vist. Specifikt vises det hvordan effektive OSC der opnår meget høje output spændinger omkring 6V kan blive udviklet vha. tandem devices, ved anvendelse af effektive interfacelag, og optimering af tykkelsen af de enkelte sub-celler. Integration af nye reaktiv sputteret Molybdæn oxid (MoOx) lag demonstreres som en ny metode til at forbedre stabiliteten af disse organiske solceller, imens der opnås høje effektiviteter sammenlignet med celler der er fremstillet ved termisk pådampning af MoOx hul transport lag. Dette opnås ved at udvikle krystallinske film med meget højt løsrivelsesarbejde, hvilket i sidste ende resulterer I den forbedrede device stabilitet.

Baseret på OSC device analysen undersøges nye metal-oxid baserede interfacelag til PSC devices ligeledes. Specielt adaptionen af NiO:Cu hul transport lag undersøges sammen med optimering af opløsningsmidler til udvikling af højeffektive PSC devices. Til slut vises det hvordan kobber sammen med specifikke elektron transport lag fører til PSC med høj effektivitet og forbedret stabilitet.

Acknowledgements

Firstly, I would like to thank Associate Professor Morten Madsen for giving me an opportunity as a PhD student to be part of his research group. Whenever I was facing a predicament, he was always there helping me to find the most easy and straight forward ways towards solving the challenges. His experience and moral boosting comments have always encouraged me to deliver my best in any kind of situation. I have always felt emboldened to have him as my PhD supervisor - who was more like a friend.

Furthermore, I would like to thank all my colleagues at nanoSYD center, MCI; especially, the PhD and postdoc fellows in the OPV group (Andre, Bhushan, Mina, Golnaz, Vida and Per) for their constant assistance, advices and always supporting attitude. We learnt from each other, solved problems together and overcame difficulties where each of the nanometer in this path would not have been possible without each and everyone’s motivation and dedication.

I would also like to thank Dr. Shahzada Ahmad and Ms. Laura Calio from THINFACE network, for providing me the opportunity to work at Abengoa, Spain during my secondment stay. Moreover, I would like to convey a very special thanks to Dr. Afshin Hadipour (Senior scientist at IMEC, Belgium) and Prof. Tom Aernouts (Group leader, Thin-film photovoltaics group, IMEC, Belgium) for giving me the chance to carry out my research work on Perovskites solar cells at IMEC under their supervision.

Last but not the least, most of the credit goes to my life-partner, Sharin (Soudabeh) for her never ending support and for all the sacrifices she had to make. Without her I would definitely not be where I am today…

Contents Introduction ....................................................................................................................................................... 4

1-1 Organic Solar Cells ............................................................................................................................. 5

1-1-1 Solution processing.................................................................................................................... 5

1-1-2 Thermal evaporation ................................................................................................................. 6

1-2 Material Selection.............................................................................................................................. 7

1-3 Perovskite solar cells ......................................................................................................................... 8

1-4 Transient Metal-oxide interlayers ..................................................................................................... 9

1-5 Outlines............................................................................................................................................ 10

1-6 Reference ......................................................................................................................................... 11

Experimental.................................................................................................................................................... 14

2-1 Reactive sputtering of metal-oxides ................................................................................................ 15

2-2 Cluster Deposition System ............................................................................................................... 16

2-3 Thermal Evaporation for organic solar cell ..................................................................................... 19

2-4 Perovskite Solar Cells ....................................................................................................................... 20

2-5 Characterization of Solar Cells ......................................................................................................... 21

2-5-1 Spectral Response of Solar Cell ............................................................................................... 22

2-5-2 Morphology ............................................................................................................................. 23

2-5-3 Optical characterization .......................................................................................................... 25

2-5-4 Electrical characterization ....................................................................................................... 25

2-6 Reference ......................................................................................................................................... 26

DBP-C70 Solar Cell............................................................................................................................................. 28

3-1 OSC principles and material systems ............................................................................................... 29

3-2 Device Architecture ......................................................................................................................... 31

3-3 DBP:C70 Organic Solar Cells .............................................................................................................. 34

3-3-1 Materials .................................................................................................................................. 34

3-3-2 Development of DBP:C70 solar cells ......................................................................................... 36

3-3-3 Fabrication Steps ..................................................................................................................... 37

3-3-4 DBP/C70: Configuration ............................................................................................................ 38

3-3-5 DBP/C70:: Fabrication ............................................................................................................... 40

3-4 Multi-junction DBP/C70 solar cells ................................................................................................... 41

3-5 Reference ......................................................................................................................................... 46

Molybdenum-Oxide Interlayers for Small Molecule Organic Solar Cells ........................................................ 50

4-1 Sputtering Molybdenum Oxide film formation ............................................................................... 51

4-2 Molybdenum Oxide film formation and properties ........................................................................ 51

4-3 Molybdenum oxide as interlayer for organic solar cells ................................................................. 55

4-3-1 Thermal evaporation of MoOx layers ..................................................................................... 55

4-3-2 Sputtered deposited MoOx layers ............................................................................................ 57

4-3-3 Post-Annealing of the Molybdenum Oxide Films .................................................................... 60

4-4 Stability of OSC devices employing sputtered MoOx layers ............................................................ 67

4-5 Reference ......................................................................................................................................... 70

Transport Layers for Perovskite Solar Cell ....................................................................................................... 74

5-1 Perovskite solar cells ....................................................................................................................... 75

5-2 Hole Transport Layers in Perovskite Solar Cells .............................................................................. 77

5-2-1 DBP .......................................................................................................................................... 77

5-2-2 MoOx ........................................................................................................................................ 78

5-3 Development of efficient PIN perovskite solar cells........................................................................ 81

5-4 HTL and ETL optimization in PIN Perovskite solar cells ................................................................... 83

5-5 Metal electrode optimization in PIN Perovskite solar cells ............................................................. 85

5-6 Reference ......................................................................................................................................... 88

Conclusion and Outlook .................................................................................................................................. 92

6-1 Conclusions ...................................................................................................................................... 93

6-2 Outlook ............................................................................................................................................ 94

Chapter 1 Introduction

Introduction

4

Solar power has been attracting a lot of attention during the past few decades mainly due to the fact that it generates electricity without producing side-products such as Greenhouse gases or radioactive wastes like other energy generation sources. The market share of solar cells has been occupied with mono-crystalline Silicon solar cells, however, the trend is changing now towards thin-film photovoltaics. Although Thin-film photovoltaic technologies have been introduced recently compared to Silicon PV, it has achieved a lot of progress in short time compared to other technologies achieving more than 13% in Power Conversion Efficiency[1].

Organic electronics is part of material science that deals with organic semiconductors, polymers or monomers (small molecules), used as an active medium in devices. The term organic refers to certain group of molecules that contains Carbon atoms, usually combined with other elements such as Hydrogen and/or Nitrogen. Organic Semiconductors are very attractive materials for electronics fabrication, because of their favorable properties. Organic semiconducting devices have attracted a lot of attention during the past decade because of their advantages over inorganic ones such as their ease of manufacturing, their abundance and since they are cheaper and easier to purify and manufacture.

There are numerous applications using organic semiconductors, amongst them Organic Light Emitting Diodes (OLEDs), which have already made it to the market where there are several applications using OLEDs such as displays, lighting application, etc. Other devices types are Organic Solar Cells (OSCs), where the organic semiconductors are used to convert light to electricity. OSCs, as in the case of OLEDs, have been studied a lot in the recent decades, and made a lot of progress since the first cell reaching 1% power conversion efficiency have been introduced in 1987[2].

Recently, a new class of solar cell called Perovskite Solar Cells has been introduced. They adopt a Perovskite crystal structures that contains a hybrid organic-inorganic lead compound as their absorber layers. Similar to organic semiconductors they show tunable band-gaps[3,4], but reaching superior transport characteristics for electron and holes[5]. The efficiency of these devices has been increased since they were first introduced in 2009 with 3.8% [6] to more than 22% in the year 2017[7], which has been certified by National Renewable Energy Laboratory (NREL) for small scale cells.

1-1 Organic Solar Cells

Organic Solar Cells (OSCs) is a thin-film solar cell technology that uses organic semiconductors as their active materials to generate electricity. The organic semiconductors are divided into the two main categories of polymers and small molecules. The manufacturing processes for OSC are much easier compared to inorganic materials, and they consume much less material due to their high absorption coefficients. Depending on the organic semiconductor and their properties, either wet processing or thermal evaporation can be adopted for device fabrication.

1-1-1 Solution processing

if molecular weight of a compound is high or if the material decomposes during vacuum heating, usually the best method to produce a film is by solution processing. Organic polymers can usually be categorized in this field. In this case the organic materials are dissolved in one or few solvents and a film can be produced

Introduction

5

by spin coating, ink jet Printing, spraying coating, etc. to form the thin-film. Depending on the fabrication method, material and solvent, the quality, thickness and morphology of the final film can vary.

1-1-2 Thermal evaporation

Small Molecules usually have low sublimation temperature and they are suitable for vacuum evaporation. Evaporating of an organic molecule requires a vacuum chamber, where molecules can be evaporated either by Physical Vapor Deposition (PVD) or Chemical Vapor Deposition (CVD). In this thesis, PVD, which include both evaporation and sputtering to form thin-film layer, are used. These evaporation technics require a very low pressure in order to obtain a defect free and smooth final film, since the long mean free path for deposition of the molecules via a molecule gas is required.

Figure 1.1 shows some example of commonly used Small Molecule Organic semiconductors for OSC fabrication. The materials are divided into 3 categories: electron donors, electron acceptor and exciton blocking layers materials.

Figure 1.1 From the top: ZnPc, Pentacen are Donor materials, C60 and C70 are acceptor organics and BCP and Bphen are Exciton Blocking layer.

Pentacene

Zinc Phthalocyanine

Introduction

6

Both electron donor and acceptor materials are active materials that form photo-generated excitons (electron-hole pairs) upon light absorption. In the cells, the photo-generated exciton are separated at the interface between the two different materials, leading to free carriers that are transported towards their respective electrode, guided by the build-in field in the cells. In terms of device fabrication, small molecules hold some advantages over polymers:

• Simple synthesis: the synthesis steps to produce monomers are usually less than 5 steps whereas for polymers it is usually more.

• Higher Purity: Monomers have well-defined molecular weight that make it feasible for several purification technics such as re-crystalization or sublimation to be utilized to purify them.

• Better control over structure and morphology by controling the deposition properties such as substrate temperature, magnetic field, carrier gas, etc.

• No need for toxic solvents • Less complexity when making multilayer structure.

1-2 Material Selection

• The most important parameter for choosing a donor/acceptor pair to fabricate OSC is their ionic potentials (HOMOaccpetor>HOMODonor) and electron affinity (LUMOAcceptor>LUMODonor); these conditions confirms that right potential will be present inside a junction and that the current will flow in the right way. Also the binding energy of an Exciton should have enough energy to make the charge seperation a energetically favorable (EEX>HOMODonor-LUMOAcceptor). This differenece determines the Voc of the junction.

• Absorption spectrum should be maximize over the solar spectrum. This way most of the spectrum will be absorbed and possbily converts to electricity. Figure 1.2 shows absorption spectrum of DBP, C70 and an OSC consisting both where it clearly shows that their absorption covers almost the whole visible spectrum starting from 350nm to 700nm.

• Right electrodes can play an important role in the perfromance of final device but it is not only their conductivity which is important but their work-function is basically defines the built-in potential of the junction. Generally, the anaode work-function should match Donor’s HOMO level and vise versa.

• Last, the materials need to have right sublimation temperature (in case of small-molecules) or right soluability (in case of polymers) plus they should be thermally stable.

In this work, solar cells based on the small molecule DBP as electron donor and C70 as electron acceptor are developed, having a particular focus on the introduction on new metal-oxide based transport layers for obtaining high efficiency and especially high stability.

Introduction

7

350 400 450 500 550 600 650 7000.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Abs

orpt

ion

(A.U

.)

Wavelength (nm)

DBP C70

DBP+C70

Figure 1.2 Absorption spectrum of DBP, C70 and a junction consists of both

1-3 Perovskite solar cells

A new concept of thin-film photovoltaics rises from the idea behind Dye-Sensitized Solar Cells (DSSCs), which has been introduced by the end of 20th century [13]. DSSC implements an organic semiconductor, a sensitizing metal oxide anode, which could be regenerated with a liquid iodide electrolyte or a solid hole extraction layer that is typically an organic molecule. If the dye in DSSCs gets replaced with a perovskite absorber, the final product is a Perovskite Solar Cell (PSC), which was introduced successfully for the first time in 2009, leading to 3.8% efficient cells [14]. The Perovskite itself is a crystal, which can be seen in Figure 1.3, and perovskite solar cells use the crystals to form the active light absorbing layer from these crystals.

Figure 1.3 Structure of an ABX3 Perovskite crystal. In MAPI perovskite, MA is A, Pb is X and I would be B.

The PSC is typically made from a hybrid organo-lead halide, although there have been reports on lead-free PSC, but their efficiency is only around 5% PCE[8]). The organic part of the structure is an amine derivative compound with a halide like CH3NH3I or in short MAI. The inorganic part is a lead salt most commonly used

Introduction

8

salt is PbI2. As for OSCs, evaporation[9–11] and solution processing[12–14] is normally adopted to fabricate PSC cells, which needs to be followed by an annealing step to finalize the formation of the perovskite crystal. The base solvent for PSC is normally a polar solvent called Dimethylformamide DMF, which can effectively dissolve the lead salt. In addition to DMF other solvents can be used in order to further assist in dissolving the cations, or help modifying the boiling point of the final solution which will be further explained in chapter 5.

The record efficiency today for perovskite solar cells is around 22% [7], which however requires the use of costly and unstable Spiro-OMeTAD hole transport layers in combination with gold electrodes, which is not viable for any commercial route. In this work, alternative metal-oxide based hole transport layers are investigated in combination with new electron transport layers.

1-4 Transient Metal-oxide interlayers

Transition metal oxides (TMOs) such as Molybdenum oxide presents interesting properties due to their high workfunctions and tunable optoelectronic properties through a variaty of compositions. The composition can be controlled by the temperature (during or after the deposition), by the growth method and by controlling the background enrivonment during growth. The crystaline structure can also be changed by changing either the aformentioned parameters, which in turn can affect the energy levels positions of the oxide, i.e. the fermi energy level and the conduction and valence band levels. These properties make them ideal candidate for many applications including gas sensing, light-emitting devices, lithium batteries and last but not least their use in organic semiconductor devices either as charge injection layer or as charge transport layer[15–17].

The energy alignment of transition metal oxides is related to their electro-chemical potential equilibration, and is of utmost importance when they are used for charge injection/collection to/from organic molecules to electrodes[18]. In organic devices, the main property that makes transition metal oxides so interesting is their ability to exchange and transport charge through close energy level alignment to electron donor and acceptor materials[19]. When metal-oxides face organic molecules, they have the ability to decrease the hole injection barrier at the interface which result in a decrease in the contact resistance between organic molecule and electrodes. Metal oxides can both affect the HOMO/LUMO of the organic molecules and from that result in a better connection between the organic layer and the respective electrode by either decreasing HOMO or increasing the LUMO[20]. Due to their tenability in their physical, chemical and electronic properties they have been widely used in those applications where electron injection/extraction properties required[21].

For hole transport metal-oxides, which is investigated in this work, there is a distinct change in the energy level alignment at the interface when the substrate work function becomes equal to the ionization energy of the organic molecule. If the work function of the substrate surpass the ionization of the organic molecule, a minimum and constant HOMO offset energy for hole transport is formed. This HOMO offset does not depend on the electronic structure of the substrate[22]. For this reason, the high work functions of many different transition metal-oxides make them very interesting candidates as hole contacting materials for many different organic materials. This is further motivated by the fact that one can manipulate their optoelectronic properties via defect engineering, which is based on changing the

Introduction

9

occupancy of the d band within the forbidden gap of the TMO[23], leading to completely modified electrical and optical properties.

In this work, reactive sputtering of Molybdenum oxide (MoOx) thin films is conducted to develop hole transport layers that can be implemented in both organic and hybrid solar cells. The work is focused on tuning the properties of the MoOx films in a manner that allow for efficient and stable hole transport in these devices.

1-5 Outlines

This work is gathered into 5 main chapters. First, in chapter 2, the different experimental techniques that have been employed throughout this work is explained. This includes both fabrication techniques for the organic solar cells developed at SDU NanoSYD (the perovskite solar cells are in this work developed at the collaboration partners in IMEC and at Abengoa Research), as well as characterization techniques employed for both films and devices. In chapter 3, the main focus is placed on development of small molecule solar cells from DBP and C70, having a special focus on development of multi-junction solar cells with high open-circuit voltages. Chapter 4 will be focused towards Molybdenum oxide hole transport layers, which was studied as hole transport layer in the developed small molecule organic solar cells. Chapter 5 will be dedicated to results related to Perovskite solar cells investigating again different hole-transport materials, based on the experience from chapter 4, and also optimizing electron transport and active layers, results which points towards directions for obtaining more stable Perovskite solar cells. At the end, conclusions and some suggestions for future works are provided.

Introduction

10

1-6 Reference

[1] http://www.heliatek.com/en/press/press-releases/details/heliatek-sets-new-organic-photovoltaic-world-record-efficiency-of-13-2, (n.d.).

[2] C.W. Tang, Two-layer organic photovoltaic cell, Appl. Phys. Lett. 48 (1986) 183–185. doi:10.1063/1.96937.

[3] G.E. Eperon, S.D. Stranks, C. Menelaou, M.B. Johnston, L.M. Herz, H.J. Snaith, Formamidinium lead trihalide: a broadly tunable perovskite for efficient planar heterojunction solar cells, Energy Environ. Sci. 7 (2014) 982. doi:10.1039/c3ee43822h.

[4] J.H. Noh, S.H. Im, J.H. Heo, T.N. Mandal, S. Il Seok, Chemical Management for Colorful, Efficient, and Stable Inorganic–Organic Hybrid Nanostructured Solar Cells, Nano Lett. 13 (2013) 1764–1769. doi:10.1021/nl400349b.

[5] S.D. Stranks, G.E. Eperon, G. Grancini, C. Menelaou, M.J.P. Alcocer, T. Leijtens, L.M. Herz, A. Petrozza, H.J. Snaith, Electron-Hole Diffusion Lengths Exceeding 1 Micrometer in an Organometal Trihalide Perovskite Absorber, Science (80-. ). 342 (2013) 341–344. doi:10.1126/science.1243982.

[6] A. Kojima, K. Teshima, Y. Shirai, T. Miyasaka, Organometal Halide Perovskites as Visible-Light Sensitizers for Photovoltaic Cells, J. Am. Chem. Soc. 131 (2009) 6050–6051. doi:10.1021/ja809598r.

[7] W.S. Yang, B.-W. Park, E.H. Jung, N.J. Jeon, Y.C. Kim, D.U. Lee, S.S. Shin, J. Seo, E.K. Kim, J.H. Noh, S. Il Seok, Iodide management in formamidinium-lead-halide–based perovskite layers for efficient solar cells, Science (80-. ). 356 (2017) 1376–1379. doi:10.1126/science.aan2301.

[8] P. V. Kamat, J. Bisquert, J. Buriak, Lead-Free Perovskite Solar Cells, ACS Energy Lett. 2 (2017) 904–905. doi:10.1021/acsenergylett.7b00246.

[9] C. Gao, J. Liu, C. Liao, Q. Ye, Y. Zhang, X. He, X. Guo, J. Mei, W. Lau, Formation of organic–inorganic mixed halide perovskite films by thermal evaporation of PbCl 2 and CH 3 NH 3 I compounds, RSC Adv. 5 (2015) 26175–26180. doi:10.1039/C4RA17316C.

[10] L.K. Ono, M.R. Leyden, S. Wang, Y. Qi, Organometal halide perovskite thin films and solar cells by vapor deposition, J. Mater. Chem. A. 4 (2016) 6693–6713. doi:10.1039/C5TA08963H.

[11] P. Fan, D. Gu, G.-X. Liang, J.-T. Luo, J.-L. Chen, Z.-H. Zheng, D.-P. Zhang, High-performance perovskite CH3NH3PbI3 thin films for solar cells prepared by single-source physical vapour deposition, Sci. Rep. 6 (2016) 29910. doi:10.1038/srep29910.

[12] W. Nie, H. Tsai, R. Asadpour, J.-C. Blancon, A.J. Neukirch, G. Gupta, J.J. Crochet, M. Chhowalla, S. Tretiak, M.A. Alam, H.-L. Wang, A.D. Mohite, High-efficiency solution-processed perovskite solar cells with millimeter-scale grains, Science (80-. ). 347 (2015) 522–525. doi:10.1126/science.aaa0472.

[13] G. Adam, M. Kaltenbrunner, E.D. Głowacki, D.H. Apaydin, M.S. White, H. Heilbrunner, S. Tombe, P. Stadler, B. Ernecker, C.W. Klampfl, N.S. Sariciftci, M.C. Scharber, Solution processed perovskite solar cells using highly conductive PEDOT:PSS interfacial layer, Sol. Energy Mater. Sol. Cells. 157 (2016) 318–325. doi:10.1016/j.solmat.2016.05.011.

[14] N. Rolston, B.L. Watson, C.D. Bailie, M.D. McGehee, J.P. Bastos, R. Gehlhaar, J.-E. Kim, D. Vak, A.T. Mallajosyula, G. Gupta, A.D. Mohite, R.H. Dauskardt, Mechanical integrity of solution-processed

Introduction

11

perovskite solar cells, Extrem. Mech. Lett. 9 (2016) 353–358. doi:10.1016/j.eml.2016.06.006.

[15] G. Wang, T. Jiu, P. Li, J. Li, C. Sun, F. Lu, J. Fang, In situ growth of columnar MoO3 buffer layer for organic photovoltaic applications, Org. Electron. 15 (2014) 29–34. doi:10.1016/j.orgel.2013.10.015.

[16] A. Arfaoui, B. Ouni, S. Touihri, A. Mhamdi, A. Labidi, T. Manoubi, Effect of annealing in a various oxygen atmosphere on structural, optical, electrical and gas sensing properties of MoxOy thin films, Opt. Mater. (Amst). 45 (2015) 109–120. doi:10.1016/j.optmat.2015.03.017.

[17] H. Simchi, B.E. McCandless, T. Meng, W.N. Shafarman, Structural, optical, and surface properties of WO3 thin films for solar cells, J. Alloys Compd. 617 (2014) 609–615. doi:10.1016/j.jallcom.2014.08.047.

[18] T. Matsushima, Y. Kinoshita, H. Murata, Formation of Ohmic hole injection by inserting an ultrathin layer of molybdenum trioxide between indium tin oxide and organic hole-transporting layers, Appl. Phys. Lett. 91 (2007) 253504. doi:10.1063/1.2825275.

[19] M.T. Greiner, M.G. Helander, W.-M. Tang, Z.-B. Wang, J. Qiu, Z.-H. Lu, Universal energy-level alignment of molecules on metal oxides, Nat. Mater. 11 (2011) 76–81. doi:10.1038/nmat3159.

[20] S. Tokito, K. Noda, Y. Taga, Metal oxides as a hole-injecting layer for an organic electroluminescent device, J. Phys. D. Appl. Phys. 29 (1996) 2750–2753. doi:10.1088/0022-3727/29/11/004.

[21] H. Tang, F. Li, J. Shinar, Bright high efficiency blue organic light-emitting diodes with Al2O3/Al cathodes, Appl. Phys. Lett. 71 (1997) 2560–2562. doi:10.1063/1.119325.

[22] C. Tengstedt, W. Osikowicz, W.R. Salaneck, I.D. Parker, C.-H. Hsu, M. Fahlman, Fermi-level pinning at conjugated polymer interfaces, Appl. Phys. Lett. 88 (2006) 53502. doi:10.1063/1.2168515.

[23] M.T. Greiner, Z.-H. Lu, Thin-film metal oxides in organic semiconductor devices: their electronic structures, work functions and interfaces, NPG Asia Mater. 5 (2013) e55. doi:10.1038/am.2013.29.

Chapter 2 Experimental

Experimental

14

In this chapter, the techniques used to fabricate and characterize both individual films and devices throughout this thesis are described. Only the basics regarding the methods used are explained, while at the respective chapter, the details of the detailed conditions, parameters and processes are explained. The focus of this chapter will be on reactive DC-sputter deposition of molybdenum oxide, thermal evaporation of organic and inorganic materials, basic characterization of organic solar cells, morphological characterization from Scanning Electron Microscopy (SEM) and Atomic Force Microscopy (AFM), as well as External Quantum Efficiency (EQE), Photoluminescence (PL) and absorption measurements of the respective films and devices. All these technics and methods are used throughout this work, in order to better understand the specific microscopic and macroscopic properties of the materials, films and devices.

2-1 Reactive sputtering of metal-oxides

Physical Vapor Deposition (PVD) is a common thin-film production technique, which has many variations. Sputter deposition offers strong control over several parameters of the final film. During a sputtering process, the energy of the formed ions will be transferred to the surface atoms of a target, which provides the target atoms with enough energy to be knocked-off from the target, and travel through the vacuum system towards the samples. In DC sputtering, the target is kept at a negative bias, while the substrate is positively charge, generating an electric field inside the low pressure chamber. The carrier gas, argon, will be ionized due to high electric, radio frequency field inside the chamber. The positively charged atoms in plasma will hit the negatively charged target, transferring the momentum to the target atoms, releasing them to be deposited on the sample. The pressure inside the deposition chamber should be low enough to guarantee a relatively free path for the sputtered atoms yet it should be high enough to maintain the plasma.

The main advantage of the DC sputtering technique is the high thickness control, along with good coverage and uniformity across the sample surfaces. In addition, the possibility of introducing other gases along the deposition process, such as Nitrogen and Oxygen, allows for the formation of nitrides and oxides of the target. It is also possible to control the temperature of the substrate during the deposition, which affects the properties of the final film. Tuning the temperature could for example lead to differences in the a films microstructure, e.g. amorphous, semi-crystalline or fully crystalline films[1].

Figure 2.1. DC sputtering system placed in the clean-room at MCI with two sputtering head inside the main-chamber.

Experimental

15

Figure 2.1 shows the DC sputtering deposition system Cryofox Explorer 600 from Polyteknik1 placed inside the NanoSYD cleanroom at SDU. The system has 1 DC and 1 RF sputter source, and in this work the metal-oxide films are formed from a Molybdenum target placed in the DC-sputtering part. In this system, it is possible to control the sputter power, the carrier gas flow rate as well as the flow rate of oxygen to form oxide films of different compositions.

2-2 Cluster Deposition System

At SDU NanoSYD, Mads Clausen Institute, we have been using an Ultra High Vacuum (UHV) Cluster Deposition system (Figure 2.2) manufacture by Polytechnic Company, for the development of organic solar cells. The cluster system consists of 3 deposition chambers, a transfer chamber and a load-lock. At MCI, the load-lock is directly connected to a Nitrogen Glovebox in order to minimize exposure of samples to contaminations, oxygen and humidity. Starting from left-hand side of the load-lock, there is an thermal evaporation deposition chamber using boats, which can be used for metal and metal-oxide deposition, the sub-sequent thermal evaporation deposition chamber uses crucibles for organic material deposition. The last chamber is a RF-sputtering chamber that can be employed for metal-oxide sputtering. Note that in this project, the sputtered metal-oxide films are conducted in the already optimized DC sputter system at the SDU Cleanroom. All these deposition systems are connected to a central transfer chamber, where a robotic arm is transferring samples between the different deposition chambers and load lock. Everything is controlled through a control a control unit, from where sample recipes are loaded. Crystal Quartz microbalances (QCM), using the piezoelectric sensitivity of the quartz crystal by monitoring its oscillation frequency in respond to the amount of mass deposit on it, are used to measure the film thicknesses with high precision in orders of a tenth of a angstrom.

Figure 2.2 Cluster system connected to Nitrogen glovebox with its controller unite.

1 http://www.polyteknik.com/index.php

Experimental

16

Figure 2.2 shows the complete cluster deposition system used in this project, which contains 3 different deposition chambers, plus a transfer chamber and a load-lock. All of the chambers are held under ultra-low vacuum, in the order of 10-8mbar. A robotic arm in the transfer chamber transfers the samples from the load-lock, which is connected to a Nitrogen glovebox, to the respective deposition chambers. The cluster system has 3 slots for loading samples, where on each set of sample there is a possibility to deposition an infinite number of layers in any combination of chambers. The vacuum is only broken between depositions when there is a need to change the shadow mask between any consecutive layers. During that time, the samples are kept in Nitrogen, otherwise they are under constant vacuum, which is a great advantage due to the sensitivity of the organic materials to Oxygen and humidity.

Load-Lock (LL)

The load lock of the cluster system is the chamber used to load and unload samples to the system. LL is mainly pumped with a rough pump to allow for transfer of the samples in to the high vacuum Transfer chamber, providing a vacuum in this chamber in the order of 10-2mbar. The LL has 3 loading levels, which makes it possible to run 3 different recipes for 10cm or 4inch sample sizes each time.

Transfer Chamber (TC)

The largest chamber in the system used for distributing samples amongst the differ deposition chambers is the TC. Since TC is a cross connection among all chambers, to minimize cross contamination in the system, it maintains the highest vacuum level; the normal operating pressure in the TC is around 10-9mbar using a turbo pump in combination with a rough pump. The robotic arm in the middle of TC has the role of placing samples in the respective deposition chamber according to the loaded recipe. With the help of optical windows on top of TC it is possible to follow the movement of wafers during their transport. Calibration of the robotic arm has been performed frequently during this work in order to ensure precise delivery of samples to the respective chambers.

Thermal Chamber (Thermal)

The Thermal chamber is mostly used to evaporate metals using deposition boats, or metal oxide powders, using crucibles and crucible holders. This chamber is also pumped down using a turbo pump but the vacuum level is a bit lower than TC as mentioned already, and the normal operating vacuum level in our thermal chamber is 10-8mbar.

The Thermal Chamber uses a transformer to provide a constant output voltage of 1V and high current to evaporate metals. Current will flow through a boat or a crucible holder to generate heat causing the materials to evaporate. The sample holder on top of the chamber keeps the samples facing the deposition sources while rotating it to generate films with improved uniformity. The thermal chamber contains two deposition sources that shares one transformer. It has an optical window to monitor samples and sources during the whole process (Figure 2.3).

Experimental

17

Figure 2.3 Thermal chamber of the cluster system. The first chamber after LL which is mostly used to deposit transporting metal oxides or metallic electrodes.

Organic Chamber (Organic)

The organic chamber is employed for deposition of various different organic small molecules. As organic semiconductors have lower deposition temperature compared to metals, in this chamber the crucibles are heated using a coil heater around each crucible. This way, the heat distribution is also more uniform, which is essential for controlled vacuum evaporation of small organic molecules, and it helps the material to sublimate easily without decomposing material in the crucible. The operating pressure is similarly 10-8mbar. The organic chamber has two power supplies, which control 6 material sources in total, both power supplies can be controlled simultaneously and with the help of 3 crystal quarts (1 for each 2 source) co-evaporation could be done (Figure 2.4).

Ceramic crucibles are used to evaporate materials that are placed inside the ovens. To refill the materials, we take the clean crucibles under a film-hood where each material is added to added from its sealed pack to the respective crucible. To void opening the chamber often we usually add a generous amount of material inside, this way, we make sure that we will not over-expose materials to the ambient. To remove the possible contamination from the freshly refilled materials, we usually run few test deposition and check the thickness and optical absorption.

Experimental

18

Figure 2.4 Organic chamber for depositing of our Organic Semiconductors. The materials are placed in ceramic crucibles and annealed using a metallic coil around in the oven.

The sample holder on top of the chamber is connected to a stepper motor which controls the rotation, in order to obtain uniform final films. The thickness uniformity in this chamber is 10% across the 10cm sample holder. There is a heater close to the samples with a certain distance from the samples, which can heat the samples up to 250oC. The cluster system controller thus allows annealing the samples during evaporation of a material.

2-3 Thermal Evaporation for organic solar cell

Another PVD evaporation method is thermal evaporation. Organic films are typically formed from this method, where a thermal sublimation process takes place transferring the organic molecules from a solid state directly to a gas phase in vacuum. The ceramic crucible filled with a certain material is placed inside an oven, which is heated up from the coil around the crucible providing a very good control over the temperature. The transferred heat from the oven to the crucible causes the organic molecule to sublimate and deposited on a substrate, which usually is placed on top of the source with a certain distance. The process is similar to processes normally used in organic molecular beam epitaxy, where the materials typically are evaporated from Knudsen type cells. Here, to ensure the large-area with uniform coverage needed for devices, large opening of the crucibles are used instead.

Experimental

19

Figure 2.5 Cluster system’s Organic Chamber’s sample holder, sample heater, shutter are also depicted in the lower picture in schematic form.

Figure 2.5 shows the complete organic evaporation sources used in this project, which contains 5 sources for different materials. The samples are placed in a 100mm stainless steel shadow mask, which sits on the sample holder facing the sources Figure 2.2. Rotation of the sample holder results in a uniform film on top of the samples, where the distance from the source reduces the thickness variation on the samples horizontally.

The Organic Solar Cells (OSCs) usually contains 5 layers, including 2 interfacial transport layers, 2 active electron donor/acceptor layers and a metallic contact, which can be deposited in 2 runs, due to the fact that we need to change the mask between the last transport layer and the metallic contact.

Sputter Chamber (Sputtering)

The last deposition chamber is the sputter chamber. The chamber consists of two RF-sputtering heads, which shares one power supply. The pressure inside the sputter chamber is also 10-8mbar during normal condition, and it also allows sample rotation and sample heating. There is crystal quartz to monitor the deposition rate. The RF-sputtering is conducted with an Argon carrier gas, that gives the possibility of depositing many different materials in this chamber, from isolating to metallic, and furthermore allows for reactive sputtering by introducing different gases such as Nitrogen and Oxygen during the sputtering process.

2-4 Perovskite Solar Cells

NiO:Cu we used as hole-transport layer for PIC structure MAPI. Ni(NO3)2 doped with Copper has been used to spin coat on the substrate inside a Nitrogen glovebox to avoid contamination on the layer; after spin-coating the samples are annealed inside the glovebox to remove the excessive solvent to be ready to transfer. They transferred to ambient (with presence of oxygen) where they annealed at almost 300oC for more than a minute to form NiO:Cu with a well-placed energy levels.

Experimental

20

To fabricate MAPI perovskite, Methylamine Iodine (MAI) powder from Sigma Aldrich has been measured according to the recipe and dried mixed with PbI2 salt from Sigmal Aldrich. This mixed then dissolves in a mixture of solvent gamma-Butyrolactone (GBL), Dimethylformamide (DMF) and Dimethyl sulfoxide (DMSO) with a mixture ratio of 6:2:2. Solution has been freshly prepared for each batch a day before and would keep on stirring overnight at 60oC. All perovskite has been spin coating using 2 step recipes which will be further explained in chapter 5 inside a Nitrogen glovebox. After spin-coating the Perovskites, the samples need to anneal at 100oC degree for around 10min prior to Electron-Transport layer deposition.

After cooling down the samples coated with Perovskite, samples has been cool down to room temperature inside a Nitrogen glovebox where multi-layer Electron transport layer consisting Zinc Oxide (ZnO), Bathocuproine (BCP) and Phenyl-C-butyric acid methyl ester (PCBM) has been used to form a proper buffer layer prior.

2-5 Characterization of Solar Cells

Any solar cell measured in dark conditions will show a diode characteristic in their current density vs. voltage (J-V) output. Under illumination, the diode characteristics of the cell shifts towards higher currents at zero voltage, due to generation of carriers inside the cell, namely generation of a photo-current. The power generated by the cell can be obtained by calculating the area in the fourth quadrant of the J-V curve under illumination, considering the maximum power point. To calculate the output power conversion efficiency of a cell, the Short Circuit Current (Jsc), Open Circuit Voltage (Voc) and Fill Factor (FF) must be properly characterized. The product of these three values divided by the input power of the cell gives the output power conversion efficiency (PCE) of the cell.

Figure 2.6 Typical JV curve which shows the definition of the Jsc, Voc, Vm and Jm values

The output voltage of a cell under illumination when no current is generated is the Voc. The Voc is the maximum voltage that a solar cell can provide to an external circuit, and assuming ohmic contacts at the electrodes, it depends on the difference between the HOMO level of the electron donor material and the LUMO level of the electron acceptor material, or more precisely the charge transfer state formed by the two materials[2]. For a cell under illumination, the maximum measureable current is the Jsc, which is

Experimental

21

obtained at zero voltage. The Jsc increase linearly with the incoming light intensity, as the number of photo-generated carriers dictates it.

The Fill Factor is a key parameter to evaluate how well a cell is performing. The FF is defined as the ration between the maximum deliverable power to an output circuit and the product of Jsc and Voc. In general increased series resistance and decreased shunt resistances can lead to reduced Fill Factors. The poor mobilities of organic semiconductors and relatively large carrier losses typically leads to lower Fill Factors in these devices compared to for example those of many inorganic solar cells.

𝐹𝐹𝐹𝐹(%) =𝑉𝑉𝑚𝑚𝐽𝐽𝑚𝑚𝑉𝑉𝑜𝑜𝑜𝑜𝐽𝐽𝑠𝑠𝑜𝑜

× 100

The ration between the maximum power generated by the cell and the optical input power (light) defines the power conversion efficiency of the cell. The maximum generated power from a cell is the product of Vm and Jm, and therefore:

𝑃𝑃𝑃𝑃𝑃𝑃(%) = 𝑉𝑉𝑚𝑚𝐽𝐽𝑚𝑚𝑃𝑃𝑖𝑖𝑖𝑖

× 100 or 𝑃𝑃𝑃𝑃𝑃𝑃(%) = 𝐹𝐹𝐹𝐹 × 𝑉𝑉𝑜𝑜𝑜𝑜𝐽𝐽𝑠𝑠𝑜𝑜𝑃𝑃𝑖𝑖𝑖𝑖

× 100

National American Society for Testing Materials (ASTM) defined Incident solar power density standards such as the E948 and International Electrotechnical Commission (IEC) 6090-1 clarifies test conditions to characterize electrical performance of a solar cell. Standard Testing Conditions (STC) are: 1. Set temperature for the device under measurement should be 25±1 2. Spectral distribution of light must be Air Mass (AM)1.5±25% 3. The irradiance on top of a solar cell is to be 1 sun±2%

Air Mass (AM) coefficient is the direct optical path through atmosphere of the Earth. In other words, AM is the ration between the length that light travels through the atmosphere and the length it should travel if the sun was directly overhead. The most used spectrum is AM 1.5 which most commercial solar simulators provide. AM 1.5 corresponds to atmospheric thickness to a solar zenith angle of 48.19o, which is the average latitude of the world.

2-5-1 Spectral Response of Solar Cell

One of the important characterization techniques for a solar cell is the spectral response of the solar cell, which is a very powerful tool to optimize the performance of a cell. The spectral response is the current per irradiated optical power in A/W, at a certain wavelength. Therefore the SR(λ) could be defined as JPH per incident power:

SR(λ)= JPH(λ)/P(λ) Also, the External Quantum Efficiency (EQE) for a device determines the conversion efficiency of an absorbed photon to a free charge in a range of wavelengths: EQE=Number of electrons in the external circuit/Number of incident photons

Experimental

22

The EQE could be calculated from Spectral Response by considering the energy of a photon in each wavelength using Planks equation (EP=hC/ λ) where h is Plank’s constant and C is the speed of light: EQE(λ)= SP(λ)(hC/q λ) The short circuit current could be taken directly from the EQE by simply integrating the product of EQE and the photon flux density:

Jsc=∫ 𝑞𝑞𝑃𝑃𝑞𝑞𝑃𝑃(λ) λℎ𝐶𝐶𝑃𝑃λ𝐴𝐴𝐴𝐴1.5𝑑𝑑λ∞

0

where 𝑃𝑃λ

𝐴𝐴𝐴𝐴1.5 is the spectral irradiance of AM1.5 spectrum. In this work, the External quantum efficiency (EQE) measurements were performed by irradiating the samples with a 150W Xe lamp through a Monochromator (VIS-NIR Newport Cornerstone 1/4m) using a fiber coupling into a Mitoyo FS-70 microscope. The external quantum efficiency measurements were carried out in air at 300K using the same conditions aforementioned. A calibrated Si photodiode (Hamamatsu S2386-44 K) was used to measure the incident power.

2-5-2 Morphology

2-5-2-1 Atomic Force Microscope (AFM) Atomic Force Microscope (AFM) has been used to map the topography of the involved organic films and stacks in this work. In general, the AFM can operate in both contact mode and tapping mode. The contact mode is a basic mode where the tip is constant contact with the scanned surface, which is not well suited for many applications due to adhesion and friction between the tip and the surface, therefore in several applications the tapping mode is been used to avoid the drawbacks of contact mode[3]. In tapping mode, the tip of the cantilever will oscillate at a frequency near the natural frequency of the tip, while it allows the tip to touch the surface of a sample to perform a scan. Doing so, the dragging force will also reduce dramatically providing a better control over the movement of the cantilever[4] along with potential less film damage, which for example is seen in organic thin-films when used in combination with contact mode scanning.

Experimental

23

Figure 2.7 Atomic Force Microscope at MCI clean-room. AFM uses to capture the morphology of the deposited films

The Dimension 3100 Nanoman AFM from Veeco (Figure 2.7) has been used to capture the topography of the films. The standard parameters that has been extracted from the films has been RMS roughness and thickness of the films, in addition, formation of crystalline features, especially for the metal-oxide films have been investigating from the captured AFM.

2-5-2-2 Scanning Electron Microscope (SEM) Scanning Electron Microscopy (SEM) provides information about the surface morphology of a sample using an Electron beam. The interaction of Electron beam with the surface of the sample generates 3 main signals including Secondary Electrons (SE), Back-Scattered Electrons (BS) and X-ray which are used to map an object in terms of morphology and composition.

Figure 2.8 Scanning Electron Microscope at the MCI clean-room used in this project for example to investigate the

cross-section of organic solar cells.

At MCI we use Scanning Electron Microscope (Hitachi S-4800 SEM) (Figure 2.8) and for imaging we combined the SE signal with BS signal to generate a high resolution image.

Experimental

24

2-5-3 Optical characterization

A spectrophotometer allows complete characterization of optical properties of a solid film or a liquid in terms of absorption, reflection and transmission. The UV/VIS/NIR spectrophotometer, Shimadzu 2700 used in this work is built from an ISR-2600 integrated sphere, to measure transmittance, reflectivity and absorption of a target material with resolution of 0.1nm from 185nm to 900nm. With the ability to make accurate transmittance measurements to 0.000001 %( 1 part in 100 million), the UV-2700 can be applied to evaluating the transmission characteristics of polarization films.

2-5-4 Electrical characterization

Current-Voltage characteristics are measured under dark conduction and under illumination by 1 sun, AM 1.5G spectrum, while the voltage is swept and current is measured from a Keithley 2400 SourceMeter. . The light source used is a Sun 3000 class AAA solar simulator from Abet Technologies. A LabVIEW program is used to control the source, shutter and to extract the dataset from the organic solar cells.

The measurements are conducted using a home-made sample holder (Figure 2.9). The sample holder keeps the sample in Nitrogen, when loaded inside the glovebox, while the samples are being measured, and it is sealed with a 4mm thick fused silica glass that has more than 90% transparency for wavelengths above 270nm. The sample holder allows measuring 4 samples (maximum 16 individual devices) per run, each sample would be connected to 12 spring-loaded gold pins for the contact, 2 per device and 4 for common electrodes (2 on each side).

Figure 2.9 The home-made sample holder that we use to characterize our 15mmx15mm samples with top-illumination.

Experimental

25

2-6 Reference

[1] S. Uthanna, V. Nirupama, J.F. Pierson, Substrate temperature influenced structural, electrical and optical properties of dc magnetron sputtered MoO3 films, Appl. Surf. Sci. 256 (2010) 3133–3137. doi:10.1016/j.apsusc.2009.11.086.

[2] K. Vandewal, Interfacial Charge Transfer States in Condensed Phase Systems, (n.d.). doi:10.1146/annurev-physchem-040215-112144.

[3] G. Binnig, C.F. Quate, C. Gerber, Atomic Force Microscope, Phys. Rev. Lett. 56 (1986) 930–933. doi:10.1103/PhysRevLett.56.930.

[4] J. Tamayo, R. García, Deformation, Contact Time, and Phase Contrast in Tapping Mode Scanning Force Microscopy, Langmuir. 12 (1996) 4430–4435. doi:10.1021/la960189l.

Chapter 3 DBP-C70 Solar Cell

DBP-C70 Solar Cell

29

3-1 OSC principles and material systems

In contrast to the inorganic semiconductors, the terms electron donor and electron acceptors are used in organic electronics to represent p-type and n-type semiconductors. The donor points to a molecule that generates excitons, and transfers (donates) electrons to the acceptor molecules, which in turn accept the electrons, due to its high electron affinity, thus leading to a separation of excitons at the donor and acceptor interface.

The main difference between organic and inorganic semiconductors in solar cells is the presence of strongly bound (Frenkel type) excitons, and the carrier mobilities of the materials. Frenkel excitons are electron-hole pairs bound with each other by a Coulomb force, before they get separated at the interface and start contributing to the photo-generated current as free charges.

Equation 1 shows the connection between the refractive index of medium to the possibility of free charge formation inside a medium. Because of the lower relative permittivity of organic materials (~3), strongly bound photo-generated excitons are generated in these materials.[1] The separation of excitons will take place at the interface between the donor/acceptor interfaces, due to the energy level difference at the interface. The rather short diffusion length of the generated Frenkel excitons means that the interface area inside organic solar cells therefore needs to be large, in order to avoid losses. After separation of excitons, the hole will be transported inside the donor material (which typically has higher hole-mobility), and electron will be transported by the acceptor material (with higher electron mobility), guided by the build-in field, towards the respective electrodes.

𝐹𝐹𝑐𝑐 = 14𝜋𝜋𝜋𝜋𝜋𝜋0

𝑞𝑞2

𝑟𝑟2 (eq. 1)

where r is the distance between the charges q, 𝜀𝜀0 is the vacuum permittivity and ε is the permittivity of the organic semiconductor.

The ionization potential (IP) is the energy required to remove an electron from the molecule in its neutral state. The donor material, which will be giving away electrons, needs to have a lower IP than the acceptor, so that electrons can be promptly extracted (IPD < IPA). The IP is equivalent to the HOMO level.

The ionization energy is the required energy for removing an electron from an atom in its steady state. A donor material which is supposed to give away electron should have ionization energy lower than the one in acceptor so that electron transfer can occur with minimum loss. On the other hand, the electron affinity is equivalent to the release energy from an atom in neutral state when it accepts an electron. For an acceptor, this energy should be lower than donor to guarantee prompt electron transfer[2,3]. If the energy difference is not sufficient for separating the electron-hole pair, then the exciton will recombine without contributing to the photo-current, and lead to loss processes. Because of the short exciton diffusion lengths, bilayer organic solar cells are usually made very thin, using suitable materials with proper energy levels, and bulk heterojunction cells are optimized towards large interface areas that lead to efficient exciton splitting, and at the same time transport pathways for hole and electrons towards the respective electrodes. The photophysical steps involved in the operation of organic solar cells, from incident photon absorption to collection of free charges at the electrodes are the following: (Figure 2.11)

DBP-C70 Solar Cell

30

1. Light absorption when a photon is absorbed by either the donor or acceptor material 2. Exciton Diffusion to the donor/acceptor interface 3. Charge transfer the separation of electron/hole pairs 4. Charge transport of free carriers to the respective electrodes 5. Charge collection at the electrodes

Figure 2.11. Conversion of light to electricity inside a bilayer OSC.

Light Absorption:

When light penetrate the transparent electrode and enters the active medium, it gets absorbed by either the donor or acceptor material. In this process, the thickness of the active layers plays an important role, due to the combined low exciton diffusion lengths and low carrier mobilities of organic semiconductors, which means that the active layers must be very thin, usually in the order of tens of nanometers.[4] However, due to high absorption coefficients of most organic materials, such thin films are also sufficient for absorbing reasonable amounts of light[5]. In addition, the bandgap of the active layers of course have to be carefully chosen. In most cases, the organic materials have high bandgaps, typically around or even exceeding 2 eV,[6] however, in recent years a lot of research results on high performance organic solar cells from low bandgap polymers have pushed this value down in energy.

Exciton Diffusion:

After a photon is converted into an exciton by light absorption, the exciton will start to diffuse inside the organic semiconductor. If it in that process ends up at the interface between the donor and acceptor materials, the energy difference at the interface will assist the pair to be separated, via a charge transfer state formed between the two different semiconductors. Because of the small exciton diffusion length, some of the photo-excited excitons will recombine before reaching at the interface. Exciton diffusion lengths are typically in the orders between 5nm to 50nm, depending on the materials.[7,8] Because of charge neutrality of the photo-generated excitons, they would not be affected by the electric field, which results in an isotropic diffusion of them in an organic semiconductor. Due to the presence of traps and defects in a material lattice, or simply because of non-crystallinity of organic semiconductors, excitons can easily recombine, even before reaching the average diffusion length.

DBP-C70 Solar Cell

31

The exciton diffusion length should be considered very carefully because only those excitons that reach the donor/acceptor interface contribute to any photo-current. This typically leads to a trade-off between light absorption and recombination mechanisms in organic solar cells. In too thin cells, light absorption is not ideal, but good transport properties are obtained, a vice versa for thick cells.

Charge transfer state:

At the interface of donor/acceptor materials, intermolecular Charge Transfer (CT) states cause the dissociation of the carriers. The charge transfer process is the result of the energy exchange for an exciton to form free carriers at the interface. The presence of CT states is due to the different energy level between the Lowest Unoccupied Molecular Orbital (LUMO) and Highest Occupied Molecular Orbital (HOMO) of the donor with LUMO and HOMO level of acceptor. In CT states, the electron-hole pair are located on two neighboring molecules of two materials,[9] but not yet turned into free charge carriers that can contribute to the photocurrent. The CT states are in case of ohmic contacts defining the open-circuit voltage in organic solar cells.

Charge transport:

Once the charges are separated at the interface, the electron will be present in the acceptor material, whereas the hole will remain in the donor. Since the electron and holes are charged particles, they are being affected by the internal electric field in the solar cell, leading to transport towards the respective electrodes. In bulk heterojunction cells, however, there is no direct path inside the organic molecular system for charge transport, which can lead to additional loss processes and reduced Fill factors. The charge transport inside an organic layer is described by a hopping mechanism instead of band transport[10]. A low internal field inside organic solar cells, the driving force for transporting charges is the gradient of the concentration of electrons and holes in each material, which causes the carriers to diffuse towards respective electrodes. A larger internal field, the process is dominated by carrier drift, and the product of mobility and lifetime (µτ) is a good measure of the efficiency of this process, which shows how far a free charge can travel in a fixed electric field before it recombines.

Charge collection

At each electrode, the charges must be collected efficiently to avoid recombination at the interface. which means that charges needs to overcome any potential energy barrier at the organic/metal interface.[11] In inorganic solar cells, charge collection is facilitated by using a highly doped semiconductor in contact with the metal, in order to provide minimum energy barriers at the electrodes. In organic electronics, electron transport and hole transport layers have been used to act as a doped layer. These layers do not assist in collecting the charges at the respective electrode, but they in some cases also avoid collecting the wrong charge at the wrong electrode, which is why they sometimes are referred to as blocking layers.

3-2 Device Architecture

During the past decades, there are mainly four used architectures in organic solar cell technology, which in this part will be explained briefly. During the introduction of all architectures, the advantages and drawbacks of each will be highlighted.

DBP-C70 Solar Cell

32

Single layer architecture: the most basic configuration is the “single layer” architecture, where there is one type of semiconductor sandwiched between two electrodes (Figure 2.11). These devices are usually labeled Schottky diodes, as the charge separation occurs at the interface between the organic semiconductor and one of the electrodes. The other contact is considered to be ohmic contact, depending of the type of semiconductor, and respecting energy level alignment. Because there is only one active material, the absorption is comparatively lower, and moreover, with the presence of both carriers in the same material, the recombination losses should be higher. The lower efficiency of these devices is also simply hampered by the fact that the field generated at the Schottky contact typically is not strong enough to efficiently split Frenkel type excitons.

Figure 2.11. Schottky structure. A sandwich of active organic semiconductor between two electrodes.

Bilayer Configuration: In bilayer cells, two separate organic semiconductor layers are sandwiched between two electrodes (Figure 2.12). In this configuration the charge transport properties are considered very well, as each layer can be optimized for optimum transport and absorption at the same time, reducing any recombination losses. For highly absorbing organic semiconductors, this can sometimes be of advantage. The main disadvantage of this architecture is the small interface between the active materials, which reduce the exciton separation possibility. For such configurations, exciton diffusion lengths must be carefully considered to achieve the most optimized cell, which I will further elaborate on throughout this chapter. This configuration is also the main configuration used in this work with DBP and C70 molecules.

Figure 2.12. Bilayer structure. Bilayer consists of an electron donor in contact with an electron acceptor with a single interface.

Bulk Heterojunction (BHJ): Instead of having two separate donor and acceptor layers, as in bilayer solar cells, bulk heterojunction solar cells consist of a single layer containing both donor and acceptor material

DBP-C70 Solar Cell

33

mixed together. The ratio between each material could vary, but the concept remains the same. By mixing the two materials, the donor acceptor interface will increase, assisting excitons to transform into free charges very fast. In this method, there is a possibility of forming an island of donor (acceptor) material inside an acceptor (donor) layer, where the charges would not have a transport path to their respective electrodes, leading to losses. More free charges are generated in this architecture, which thus provides higher current density compared to the other previously mentioned architectures. This however requires delicate optimization of the nanoscale domains that form the heterojunction inside the active layer.

PIN/NIP: the most efficient configuration for organic solar cells is in many cases the combination of the two aforementioned architectures. In such architecture, more free carriers can be generated, providing still low resistive direct paths towards the electrodes. For that, the BHJ layer needs to be sandwiched between two permeable layers for efficient charge collection at the contacts (Figure 2.13).

Figure 2.13. PIN (NIP) configuration is a BHJ junction with few addition transport layers sandwiched between two electrodes. PIN (NIP) configuration usually combines the best properties of Bilayer and BHJ and it give best

performance of two semiconductors.

Multilayer Structure: As the width of the absorption spectrum of an organic semiconductor is relatively low, it is possible to combine multiple materials to use the most of the spectrum. One of the methods to do so is to fabricate a tandem, triple or multilayer configuration.[12] Tandem configuration is mainly interesting for small molecule organic semiconductors, since the method of fabrication is more suitable for such structure. A single PIN/NIP device can provide a good performance in case of IQE (current density) and Fill-Factor, but a single device will not take well use of the solar spectrum, therefore limiting the available energy output. To overcome such limitation and to maximize both IQE and output power, two different PIN/NIP stacks of the same type but with different materials (different absorption spectrum) can be used.

A tandem cell consists of at least two PIN/NIP cells on top of each other with a recombination center in between them. As the cells are mostly connected in series, ideally the current densities of both cells are balanced in order to maximize the output power from the cell and reduce losses of free carriers. The output voltage is going to be the sum of the output voltage of the individual cells. Today, this configuration using

DBP-C70 Solar Cell

34

different types of small molecule semiconductors achieved the world record efficiency of organic solar cells with more than 13.0%.[13]

3-3 DBP:C70 Organic Solar Cells

The results of fabrication and characterization of organic solar cell based on Tetraphenyldibenzoperiflanthen (DBP) and Fullerene (C70) will be presented in this section. During the first part, I will focus on development of organic solar cells (OSC) using the cluster deposition system at SDU NanoSYD and the preparation steps for fabricating the organic solar cells. Subsequently, the results for a bilayer configuration OSC based on DBP/ C70 will be presented and at the end, I will show the capability of stacking several thin OSC on top of each other in order to achieve very high output voltages.

3-3-1 Materials

In this subsection, the employed materials used to develop the organic solar cells in this work will be shortly introduced.

Tetraphenyldibenzoperiflanthen (DBP)

DBP is a P-type (Donor) semiconductor with a symmetric structure demonstrated in Figure 3.1 (left), which only consists of Carbon and Hydrogen. DBP molecule is a planer non-polar molecule with a desired orientation which could be tuned by changing either the substrate temperature[14] or by controlling the gas-flow during the Physical vapor deposition of DBP[15].Since its patent in 2009, it has been used by many groups all over the world[16–20].

Figure 3.1. (left)Molecular structure of Tetraphenyldibenzoperiflanthen (DBP), which only consists of Carbon and Hydrogen atoms. (righ) Molecular structure of Fullerene (C70) which only consists of Carbon atoms.

The main advantage of DBP over other donor materials is its high optical absorption, which allows for decreasing the thickness of active layer, resulting in a reduction in recombination losses. Moreover, its HOMO level has been measured to be around 5.5 e.V [20], which is deep enough to reach respectable open-circuit voltages in conjunction with fullerene acceptors, the most used organic acceptor semiconductors to date. The Voc is related to the difference between the HOMO level of the donor and the LUMO level of the acceptor material, or more precisely the CT state energy for electron transfer between the donor and acceptor molecules [21].

Fullerene (C70)

DBP-C70 Solar Cell

35

Fullerene defines as any molecule containing only Carbon atoms, and which has a hollow shaped sphere, ellipsoid, tube or alike. C70 is a fullerene that contains 70 Carbon atoms shown Figure 3.1 (right). A rugby ball shaped molecule consists of 25 hexagons and 12 pentagons with a Carbon atom at the vertices of each Polygon and a bond along each polygon edge.

The most commonly known fullerenes are C60 and C70, however, there are various fullerenes with different composition Cn, where n>20. It was firstly discovered and formed using a laser evaporation technique on Graphite by Robert Curl, Harold Kroto and Richard Smalley[22], which brought them a Nobel prize in Chemistry in 1996 for their achievement. Brownish C70 powder has an energy gap of ~1.7 eV, and it has a LUMO level at 4.0 e.V. which combined with its high electron mobility and broad absorption [23,24], makes it an attractive acceptor organic semiconductor for use in organic solar cells.

Molybdenum Oxide (MoO3)

Transition metals such as Vanadium, Tungsten and Molybdenum in composition with oxygen show semiconducting behavior that depends on their precise composition, and there has been numerous studies regarding their properties. Vanadium (V2O5), Tungsten (WO3) and Molybdenum (MoO3) oxides are n-type semiconductors with deep lying conduction bands and high work functions of above 6 eV. The work function can be dramatically changed due to exposure to the environment, method of fabrication and thermal treatment during or after film formation, such as it has been demonstrated for Molybdenum oxide[25,26] which will be further detailed in chapter 4. As it has been shown by Mark T. Greiner et. al. the energy offset in hole transport between such metal-oxide and any organic semiconductors depends on the difference between the metal-oxide work function, and the ionization energy for the organic molecules[26], some of these are shown in (Figure 3.3). The dash-line inside the energy gap shows the position of the fermi level in each composition, defining thus the work function of the material.

Figure 3.3. Energy bandgap, position of HOMO and LUMO level and Fermi energy level location inside the bandgap for several metal oxides.[27]

In comparison to other transition metal oxides, MoO3 has been used the most in many electronic applications, including both organic and inorganic solar cells, due to both its relatively lower sublimation temperature, which makes it feasible to be used in vacuum deposition systems in contrast to V2O5 and

DBP-C70 Solar Cell

36

WO3, which have much higher evaporation temperatures. Molybdenum oxide is a wide bandgap n-type semiconductor with a very deep lying HOMO level of around ~9 e.V. the high work function has made it a perfect candidate to be used as hole-injection and extraction layer in organic electronics such as OLEDs and OSCs. Power conversion efficiency and stability improvements in many organic devices due to the use of Molybdenum oxide compositions have been shown in past, this will be elaborated further on in chapter 4 and 5.

Bathocuproine (BCP)

Bathocuprine (Figure 3.4) is a well-known electron transport and exciton blocking layer in many organic applications such as OLEDs[28] and OSCs[29,30]. BCP is a white yellowish powder, which is solvable in many organic solvents, but not in water.

Figure 3.4. Molecular structure of Bathocuproine (BCP).

As it is the case for many other transport and blocking layers, BCP has a wider band-gap compared to the employed active materials, which makes it almost transparent within the visible spectrum. The role of BCP between the electron acceptor and metallic electrode is to reduce exciton recombination loss and provide transport of electrons to the metallic electrode. The deep lying HOMO prevents at the same time transport of holes, and thus it works both as electron transport and hole blocking layer. Note that in pristine form, BCP is not a proper selection for electron transport in many devices, due to the very high lying LUMO level. The electron transport properties have been demonstrated to be due to BCP:metal complexes formed during electrode evaporation on top of BCP, arising from metal atom penetrating the BCP film, lowering the LUMO level to match e.g. the LUMO level of many fullerene derivatives[31].

3-3-2 Development of DBP:C70 solar cells

In this section, the fabrication process of the OSC developed in this work will be explained in details. The initial steps for sample preparation will briefly be stated before the main fabrication processes, including organic and electrode layer deposition will be detailed. The system used for characterization of the OSC will shortly be described at the end.

The fabrication and characterization of the OSC has been conducted at OPV group at SDU NanoSYD, the Mads Clausen Institute (MCI) at University of Southern Denmark. As mentioned earlier, there are two main methods to fabricate OSC depending on the molecular weight of a material. Small molecule organic semiconductors have comparatively lower sublimation temperatures, and in this work, the focus is mostly on this category. To develop the OSC used in this work, a dedicated cluster deposition system was employed, where thermal deposit of the active layers, transport layers and metal contacts was conducted.

DBP-C70 Solar Cell

37

As the cluster hatch opens to a Nitrogen glovebox, samples are loaded and un-loaded in Nitrogen environment, and this not exposed to air before transport for JV characteristics.

In thermal evaporation of organic molecules, the material is heated up from a boat or crucible inside the high vacuum system (around 10-8 mbar), in order sublimate the material at elevated temperatures to form a molecular beam of the material going from source to the samples. The molecular beam depends on the material, shape of the boat or crucible and deposition temperature, furthermore it will change depending on the vacuum level inside the vacuum process chamber.

3-3-3 Fabrication Steps

The first step in fabrication of OSC is to prepare the substrate. As the Glass samples coated with Transparent Conductive Oxide (TCO) arrives from the provider in vacuum packs, optical inspection of any defects in the TCO is first conducted. The most common TCO in OSC industry is Indium Tin Oxide (ITO) which has low sheet resistance (around 10-20 Ω/□) and high transparency in visible regime (more than 90%) compared to other commercially available TCOs. The used pre-patterned ITO coated glass samples have been ordered from Kimberly-Clark Technology (Kimtech). 700µm thick glass has been diced to 15mmx15mm samples, which have a 10mm wide stripe of ITO in the middle. The ITO layer has thickness of 100nm with average roughness of 2-3nm with sheet resistance of 15 Ω/□ (Figure 3.5).

Figure 3.5. Glass samples coated with 700 µm ITO strip in the middle. Final OSC device containing 4 individual cells. The active layers are completely separated with the help of cross limiting any cross talk between samples.

After inspection under optical microscope, sample are cleaned with the help of an ultra-sonic bath and by using Acetone (15 min) and Isopropanol (IPV) (15 min) and afterwards they are dried using a nitrogen blower. Following this, the substrates are placed in an air plasma chamber for about 20min to remove the remaining carbon residues and to increase the work function of ITO[32].

The samples then transferred to the glovebox before transfer to deposition chambers. Inside the glovebox, the samples will be placed inside 10cm stainless steel shadow mask and later the mask will be placed inside Loadlock of the cluster deposition system, to be transferred to the deposition chambers. Stainless steel shadow masks that is used for depositing organic active and transport layers (Figure 3.6 left) contains a cross in the middle of each substrate, in order to avoid cross-talk between each individual solar cell. The metal masks (Figure 3.6 right) are used for the deposition of metal electrodes, and it defines the active area of each cell. It has a wider cross in the middle to avoid shorts. The silver deposited in the areas close to the

DBP-C70 Solar Cell

38

edge of the sample is used for probing the samples using probing pins, without scratching material off or penetrating the organic layers (resulting in shorting-circuiting a device).

Figure 3.6. (left )Active material/transport layer stainless steel Deposition Mask. (right) Stainless steel electrode deposition mask. The lower patterned plate is 0,3mm thick to minimize the shadowing effects.

The control unit of the cluster system controls the robotic arm for loading and unloading the samples in any deposition chamber. Recipes have been optimized for each material, which thus can be loaded to the program for OSC fabrication, where especially the thickness, but also deposition rate of any layer is set. For most of the recipes, the organic deposition rate is kept around 0.1-0.2 Å/s, and for silver contacts rates of around 0.2-0.5 Å/s is employed. During evaporation of any material, the mask is kept rotating at 5rpm, in order to form a uniform smooth final film. (Figure 3.5)

3-3-4 DBP/C70: Configuration

Bilayer configuration containing a donor and an acceptor layer has employed to develop reference solar cells with relatively high efficiency, which arises from the rather high absorption coefficient of DBP, allowing for development of thin layer devices, and thus relatively low exciton losses[33,34] . From this simple base structure, it is possible to develop multi-junction solar cells consisting of stacks of many individual subcells, in order to increase the output voltage by connecting each consecutive stack in series. By optimizing the layer between each stack, good ohmic contacts can be achieved, limiting the losses in the stack. Simulation can also be adopted to calculate the optimized thicknesses for each single layer of the active materials, in order to increase the output power while avoiding any drop in output voltage. In this section, we will present on development of reference solar cells, multijunction solar cells, and at the end also PIN structure solar cells, all based on DBP/C70 active layers, employed molybdenum oxide both as hole transport layer ,and as part of the interlayer between multi-junction cells.

Contrary to solution processed BHJ solar cells, achieving best performing BHJ with evaporation needs a lot of optimization where slightest change in temperature of each material can dramatically alter the performance of the final device. Co-evaporation of two organic material at different temperature, where each material has to be controlled individually at their right sublimation temperature, besides dependency

DBP-C70 Solar Cell

39

of the deposition rate on factors including amount of material inside the boat, deposition temperature, sample temperature, etc., makes evaporation a difficult method in terms of reproducible results. The used cluster deposition system provides a superior control over the thickness over the 10cm shadow masks (samples) with less than 5% variation in the thickness over the whole area. Usually to have a complete understanding of the process, multiple samples are employed in a single run, depending on the sample size and design of the shadow mask. In the work presented here, 18 samples can be fitted into the cluster per run, where each have a maximum number of 4 cells giving in total a maximum of 72 cells developed in one run.

The basic bilayer configuration OSC consist of the active layer sandwiched between the molybdenum oxide hole transport layer (HTL) and BCP Electron transport layer (ETL) using ITO and silver as electrodes. A standard configuration device where the HTL is deposited on top of the ITO (Anode) and ETL is deposited on top of the active layer is used. Figure 3.8 shows the energy levels of our electrodes, transport layers and active materials in a bilayer (PIN) configuration. Note that molybdenum oxide here function as a pure hole transport layer that do not block excitons. Note also that intrinsic BCP does not conduct electrons well in such an energy level alignment scheme, however, the strong electron transport properties arise from the complex made with silver atoms penetrating the layer leading to a lower lying LUMO level fo the BCP:Ag complex[35].

Figure 3.8. Energy level alignment and positioning of our material in an OSC stack; the dimensions are not scaled to their respective thicknesses.

Pre-patterned ITO glass substrates with ITO thickness of 100-120nm has been order from the Kintech company. The samples have 2-3 nm average root-mean-square roughness and they have around 80% transmittance at 550nm. DBP, the employed electron donor material is purchased from Luminescence technology (LumTech. Taiwan), Fullerene C70 purchased from Solenne, Netherlands and MoO3 powder and BCP are bought from Sigma Aldrich.

DBP-C70 Solar Cell

40

Fullerene is the most common and used electron acceptor among many research groups in both small molecule shape as Cx or in polymeric shape as PCxbM (where x>20) derivatives. Fullerenes mostly have very high electron mobility, long exciton diffusion length and the absorption spectrum broadens with the size of the molecule. With its LUMO level (~4.1 e.V.), also leads to rather high voltages in combination with DBP. There are numerous donor materials used to fabricate small molecule OSCs, where each has their advantages and drawbacks. DBP has a strong optical absorption and deep lying HOMO level (~5.5 eV), which results in high Voc as mentioned earlier. The high absorption coefficient means that the thickness of the DBP layer can be decreased to reduce the recombination loss, while still maintain strong light absorption. The maximum reported Voc using DBP as donor has been reported by Forrest et. Al. to be 0.91 V[16].

3-3-5 DBP/C70: Fabrication

A typical characteristic of a Bilayer configuration, which in most cases provides higher FF and Voc but lower current than mixed heterojunction cells, is shown in Figure 3.9. A very high Voc of around 900 mV with FF values of around 65% is achieved in these cells, which is comparatively close to the maximum reachable values for this material combination. Later, will results from better thickness optimized bilayer cells, developed with assistance from optical simulation. The thickness optimization helps on improving the current density without influencing the voltage and FF.

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9-7

-6

-5

-4

-3

-2

-1

0

1

2

Cur

rent

Den

sity

(mA

/cm

2 )

Voltage (v)

Figure 3.9. Typical JV curve of a Bilayer DBP/C70 solar cell. The black curve is a dark curve, and the red one is measured under 1 Sun illumination.

The optical absorption for 50nm of DBP, C70, BCP and MoO3 is shown in Figure 3.10, which has been deposited on plane, polished glass samples. From the absorption spectrum it is clear that the transporting materials are almost transparent in the visible spectrum where the active layers are absorbing the most. DBP absorption spectrum is complementary to C70’s absorption spectrum, in another words, DBP has a strong absorption where C70 has a weak absorption, and vice versa.

DBP-C70 Solar Cell

41

350 400 450 500 550 600 650 7000.0

0.2

0.4

0.6

0.8

1.0

1.2

Ab

sorp

tion (

A.U.

)

Wavelength (nm)

BCP MoO3

DBP C70

350 400 450 500 550 600 650 7000.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

1.1

% Absorption Spectrum(Glass/ITO/MoO3/DBP/C70/BCP)

EQE of an OSC

Wavelength (nm)

Abs

orpt

ion

(A.U

.)

0

2

4

6

8

10

12

14

16

18

20

EQ

E (%

)

Figure 3.10. (left)Absorption of each of our materials on plan glass substrate (glass absorption has been deducted).(right) Absorption Spectrum and EQE of our fabricated OSC which shows the peak matching.

Figure 3.10 shows the absorption spectrum of a single junction OSC, along with the EQE of the cell, which demonstrates peaks matching to the absorption spectrum. The measured short-circuit current from the EQE spectrum (integration of the spectrum) is around 6.1 mA/cm2, which is relatively close to the averaged output Jsc of the same from JV curve of Figure 3.9.

3-4 Multi-junction DBP/C70 solar cells

In terms of multi-junction solar cells developed from stacks of individual DBP/C70 based subcells, solar cells non-optimized for current matching were initially investigated. Figure 3.13 shows the J-V curves of DBP/C70 multifold solar cells characterized under AM 1.5G illumination and Table 3.2 shows the details on the average extracted parameters. The active layer thicknesses of the single layer cells were adopted from the previous section, and used as a base for choosing the thicknesses of the active layers in the multifold cells (at this point non-optimized for current matching).

Figure 3.13. JV characteristics of the DBP/C70 multifold and single solar cells (best performing devices shown) investigated in the devices non-optimized for current matching. The JV curves show the increase in Voc and decrease in

Jsc as the number of cells in the multi-fold stack is increased.

0 1 2 3 4 5 6 7-6

-5

-4

-3

-2

-1

0

1

Cur

rent

Den

sity

(mA

/cm

2 )

Voltage (V)

Single 2-Fold 3-Fold 5-Fold 10-Fold

DBP-C70 Solar Cell

42

As light passes through the layers, the light intensity decreases due to photon absorption by the active materials in each sub-cell, leading to photo-generated excitons in each individual sub-cell. In addition, the exciton generation rate is affected by optical interference effects in these thin-film devices. Since all the cells in the multi-junction device are connected in series, the overall current is limited by the least current generating sub-cell, which again is the one showing the weakest light absorption. Therefore, the active layer thicknesses for the multifold cells were chosen to be thinner than in the single cells (Table 3.2).

Donor/Acceptor Voc (V) Jsc (mA/cm2) FF (%) PCE (%) Single (10/30) 0.85 6.04 64.10 3.30 2-Fold (5/20-10/30) 1.47 3.47 53.32 2.71 3-Fold (10/10) 2.28 2.25 53.87 2.75 5-Fold (10/10) 3.69 1.39 51.86 2.66 10-Fold (7/10) 6.31 0.68 47.33 2.01

Table 3.2. Average output parameters of multifold and single cells non-optimized for current matching, which shows the drop in current density and increase in open-circuit voltage as the numbers of sub-cells are increased.

To support the choice of active layer thicknesses in these non-optimized multifold cells, we investigated device parameters from different active layer thicknesses in each cell type

1st cell 2nd cell 3rd cell 4th cell 5th cell DBP C70 DBP C70 DBP C70 DBP C70 DBP C70

Single 20 30 2-fold 10 25 10 19 3-fold 10 20 5 10 10 11 5-fold 5 7 10 18 10 20 5 10 10 12

Table 3.3. Optimized active layer thicknesses in nanometers of the muti-stack devices. The accuracy of the thicknesses is 1nm.

However, at this stage, the optimization steps are only conducted for few different thicknesses for each cell configuration, as a complete optimization would require the modeling conducted later in this work. Even though the active layer thicknesses of the cells in this work, as shown in Table 3.3, are smaller for cells containing multiple sub-cells, the overall Jsc is substantially decreased for these multifold non-optimized solar cells, which can arise from the lack of current matching amongst each of the sub-cells. The effect is emphasized in figure 3.14 (left), which shows the large decrease in Jsc as a function of number of sub-cells in the non-optimized devices. On the other hand, Voc is increasing as expected and raise all the way up to 6.44V for the 10-fold cells with only minimal losses. The losses could for example arise from morphologically modified interfaces in the active layers, which at the end could result in a reduced Voc, leading to the small decrease in the device efficiency as the number of sub-cells increases in the devices (Fig. 3.14 right). Another contribution to the small efficiency loss is the Fill Factor, which show a small drop with an increase in the number of sub-cells, which could arise due to increased series resistance in the devices. In order to maximize Jsc in the devices, and thus also the device efficiency, the thickness of each sub-cell should be adjusted to obtain similar light absorption in each sub-cell, which again leads to similar exciton generation rates and thus current matching. To obtain this, we conduct in the second part of this work optical and electrical device modeling to optimize for current matching between each sub-cell.

DBP-C70 Solar Cell

43

Single

2-Fold

3-Fold

5-Fold

10-F

old0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

PCE

(%)

0

10

20

30

40

50

60

70

80

90

100

Mea

n FF

(%)

Figure 3.14. Device parameters of non-optimized DBP/C70 multifold and single solar cells. (left) The plot shows that the

Voc is increasing along with the number cells in the multi-fold device, while at the same time Jsc is decreasing. (right)The PCE and FF for the different cells show a small decrease as the device stack is increased.

To investigate the optical response within each sub-cell, and to obtain current matching in the multi-junction cells, optical modeling based on the Transfer Matrix Method[36] was implemented. We note that in this work, the main focus is placed on optimizing the single and 5-fold solar cells, although the method would be equally applicable and valuable for other multifold device stacks as well. In order to determine the photo-currents caused by the generated excitons, a drift-diffusion simulation based on an electrical model for OPV reported in Ref. [37] was employed. The Jsc of each sub-cell were obtained and chosen as the criteria for generating current matching in the sub-cells. Figure 3.15(left) demonstrates the calculated non-optimized current density of each sub-cell, which shows a substantial mismatch between the photo-generated current from each sub-cell due to absorption changes, which is also caused by the strong optical interference within these thin layers. From the performed device simulations, the limiting cell in terms of current mismatching was found. The thicknesses of each sub-cell were then tuned accordingly with the help of our modeling and implemented theoretically by the device modeling, in order to compensate for the current mismatch. In Figure 3.15(right), the modeled current density of each sub-cell for a thickness optimized 5-fold cell are shown.

Figure 3.15. (left) Modeled current density of each sub-cell for the 5-Fold cell before optimization. (Right) Modeled

current density of the same stack after optimization. Modeling conducting by Yiming Liu.

Sing

le

2-Fo

ld

3-Fo

ld

5-Fo

ld

10-F

old

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5M

ean

Voc

(V)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

Mea

n Js

c (m

A/c

m2 )

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8-3

-2

-1

0

1

2

3

Cur

rent

den

sity

(mA

/cm

2 )

Voltage (V)

subcell #1 subcell #2 subcell #3 subcell #4 subcell #5

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8-3

-2

-1

0

1

2

3

Cur

rent

den

sity

(mA

/cm

2 )

Voltage (V)

subcell #1 subcell #2 subcell #3 subcell #4 subcell #5

DBP-C70 Solar Cell

44

After the optimization, the Jsc of each sub-cell is almost identical, and the current matching among the sub-cells improved the Jsc of the total device from 1.2 mA/cm2 (calculated Jsc for non-optimized 5-fold cell) to around 2.0 mA/cm2 (calculated Jsc for optimized 5-fold cell). Similar to the 5-fold cell, optical modeling was performed for a single junction cell to compare against an optimized reference containing a single cell. According to our modeling, the optimized thickness for a single junction cell was calculated to be DBP (20nm)/C70 (30nm).

Devices containing the optimized thicknesses were fabricated based on the modeling outputs, and the results in terms of averaged output parameters are shown in Table 3.4, and the device parameters of the best performing optimized and non-optimized single and 5-fold cells are shown in Table 3.5.

Sub-cell # 1 2 3 4 5 Sub-cell Jsc before optimization (mA/cm2) 2.48 1.21 1.61 2.69 2.03 Sub-cell Jsc after optimization (mA/cm2) 2.07 2.12 2.04 2.03 2.08

Table 3.4. Average output parameters obtained from the simulation for the non-optimized 5-fold vs. optimized 5-fold cell showing the current matching among the sub-cells.

Table 3.5. Comparison between the average output parameters of the optimized and non-optimized cells.

The comparison between the J-V curves for the optimized and non-optimized 5-fold cells is shown in Figure 3.16. Clearly the optimized cells lead to substantially higher Jsc as expected from the modeling results, and the Jsc also matches the expected value obtained from modeling (around 2.1mA/cm2). The output voltages are as expected not significantly modified for the optimized devices, however, small modifications is observed which again could relate to different morphologies at the interface due to the modified thicknesses.

Figure 3.16. comparison between the JV characteristics of the non-optimized and optimized 5-fold multi-stack devices.

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0-2.5

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

Cur

rent

Den

sity

(mA/

cm2 )

Voltage (V)

Non-Optimized Optimized

Structure Voc (V) Jsc (mA/cm2) FF (%) PCE (%) Single (Non-optimized) 0.85 6.04 64.10 3.30 Single (Optimized) 0.87 6.77 61.80 3.65 5-Fold (Non-optimized) 3.69 1.39 51.86 2.66 5-Fold (Optimized) 3.57 2.05 53.45 3.87

DBP-C70 Solar Cell

45

Since these tandem cells employ the same material combination throughout the stack, the device efficiency does not increase for the multifold devices shown here. There are minor losses in Voc and FF as previously mentioned, which in this work is compensated by the optimized currents to yield similar device efficiencies for the single and 5-fold optimized cells. However, the overall large increase in voltage in these devices, up to a high as 6.44 V for the 10-fold stack, makes it an interesting technology for use in combination with low power consuming electronics.

The SEM image showing the 10-fold cell (Figure 3.17) was taken by a Scanning Electron Microscope (Hitachi S-4800 SEM). The cross section image has been done on a 10-Fold substrate that has been used to obtain J-V characteristics prior to SEM imaging and it has been cut using a diamond tip.

Figure 3.17. S EM image of the cross section of the 10-fold cell showing electrodes, active layers and the recombination layers.

DBP-C70 Solar Cell

46

3-5 Reference

[1] P.G. Nicholson, F.A. Castro, Organic photovoltaics: principles and techniques for nanometre scale characterization, Nanotechnology. 21 (2010) 492001. doi:10.1088/0957-4484/21/49/492001.

[2] N.S. Sariciftci, L. Smilowitz, A.J. Heeger, F. Wudl, Semiconducting polymers (as donors) and buckminsterfullerene (as acceptor): photoinduced electron transfer and heterojunction devices, Synth. Met. 59 (1993) 333–352. doi:10.1016/0379-6779(93)91166-Y.

[3] J.J.M. Halls, J. Cornil, D.A. dos Santos, R. Silbey, D.-H. Hwang, A.B. Holmes, J.L. Brédas, R.H. Friend, Charge- and energy-transfer processes at polymer/polymer interfaces: A joint experimental and theoretical study, Phys. Rev. B. 60 (1999) 5721–5727. doi:10.1103/PhysRevB.60.5721.

[4] Y. Tamai, H. Ohkita, H. Benten, S. Ito, Exciton Diffusion in Conjugated Polymers: From Fundamental Understanding to Improvement in Photovoltaic Conversion Efficiency, J. Phys. Chem. Lett. 6 (2015) 3417–3428. doi:10.1021/acs.jpclett.5b01147.

[5] A.W. Hains, Z. Liang, M.A. Woodhouse, B.A. Gregg, Molecular Semiconductors in Organic Photovoltaic Cells, Chem. Rev. 110 (2010) 6689–6735. doi:10.1021/cr9002984.

[6] Y. Lin, Y. Li, X. Zhan, Small molecule semiconductors for high-efficiency organic photovoltaics, Chem. Soc. Rev. 41 (2012) 4245. doi:10.1039/c2cs15313k.

[7] D.E. Markov, E. Amsterdam, P.W.M. Blom, A.B. Sieval, J.C. Hummelen, Accurate Measurement of the Exciton Diffusion Length in a Conjugated Polymer Using a Heterostructure with a Side-Chain Cross-Linked Fullerene Layer, J. Phys. Chem. A. 109 (2005) 5266–5274. doi:10.1021/jp0509663.

[8] Y. Terao, H. Sasabe, C. Adachi, Correlation of hole mobility, exciton diffusion length, and solar cell characteristics in phthalocyanine/fullerene organic solar cells, Appl. Phys. Lett. 90 (2007) 103515. doi:10.1063/1.2711525.

[9] K. Vandewal, Interfacial Charge Transfer States in Condensed Phase Systems, (n.d.). doi:10.1146/annurev-physchem-040215-112144.

[10] H. Bässler, Charge Transport in Disordered Organic Photoconductors a Monte Carlo Simulation Study, Phys. Status Solidi. 175 (1993) 15–56. doi:10.1002/pssb.2221750102.

[11] L.-M. Chen, Z. Xu, Z. Hong, Y. Yang, Interface investigation and engineering – achieving high performance polymer photovoltaic devices, J. Mater. Chem. 20 (2010) 2575. doi:10.1039/b925382c.

[12] M. Riede, C. Uhrich, J. Widmer, R. Timmreck, D. Wynands, G. Schwartz, W.-M. Gnehr, D. Hildebrandt, A. Weiss, J. Hwang, S. Sundarraj, P. Erk, M. Pfeiffer, K. Leo, Efficient Organic Tandem Solar Cells based on Small Molecules, Adv. Funct. Mater. 21 (2011) 3019–3028. doi:10.1002/adfm.201002760.

[13] http://www.heliatek.com/en/press/press-releases/details/heliatek-sets-new-organic-photovoltaic-world-record-efficiency-of-13-2, (n.d.).

[14] S. Galindo, M. Ahmadpour, L.G. Gerling, A. Marsal, C. Voz, R. Alcubilla, J. Puigdollers, Influence of the density of states on the open-circuit voltage in small-molecule solar cells, Org. Electron. Physics, Mater. Appl. 15 (2014) 2553–2560. doi:10.1016/j.orgel.2014.07.011.

[15] B. Song, S.R. Forrest, Nanoscale Control of Morphology in Fullerene-Based Electron-Conducting

DBP-C70 Solar Cell

47

Buffers via Organic Vapor Phase Deposition, Nano Lett. (2016) acs.nanolett.6b01416. doi:10.1021/acs.nanolett.6b01416.

[16] X. Xiao, J.D. Zimmerman, B.E. Lassiter, K.J. Bergemann, S.R. Forrest, A hybrid planar-mixed tetraphenyldibenzoperiflanthene/C70 photovoltaic cell, Appl. Phys. Lett. 102 (2013). doi:10.1063/1.4793195.

[17] Z. Wang, D. Yokoyama, X.-F. Wang, Z. Hong, Y. Yang, J. Kido, Highly efficient organic p-i-n photovoltaic cells based on tetraphenyldibenzoperiflanthene and fullerene C70, Energy Environ. Sci. (2012) 249–255. doi:10.1039/c2ee22952h.

[18] T. Zhuang, T. Sano, J. Kido, Efficient small molecule-based bulk heterojunction photovoltaic cells with reduced exciton quenching in fullerene, Org. Electron. 26 (2015) 415–419. doi:10.1016/j.orgel.2015.08.015.

[19] C.-W. Chen, Z.-Y. Huang, Y.-M. Lin, W.-C. Huang, Y.-H. Chen, J. Strzalka, A.Y. Chang, R.D. Schaller, C.-K. Lee, C.-W. Pao, H.-W. Lin, Morphology, molecular stacking, dynamics and device performance correlations of vacuum-deposited small-molecule organic solar cells., Phys. Chem. Chem. Phys. 16 (2014) 8852–64. doi:10.1039/c3cp55385j.

[20] D. Fujishima, H. Kanno, T. Kinoshita, E. Maruyama, M. Tanaka, M. Shirakawa, K. Shibata, Organic thin-film solar cell employing a novel electron-donor material, Sol. Energy Mater. Sol. Cells. 93 (2009) 1029–1032. doi:10.1016/j.solmat.2008.11.034.

[21] B.P. Rand, D.P. Burk, S.R. Forrest, Offset energies at organic semiconductor heterojunctions and their influence on the open-circuit voltage of thin-film solar cells, Phys. Rev. B. 75 (2007) 115327. doi:10.1103/PhysRevB.75.115327.

[22] H.W. Kroto, J.R. Heath, S.C. O’Brien, R.F. Curl, R.E. Smalley, C60: Buckminsterfullerene, Nature. 318 (1985) 162–163. doi:10.1038/318162a0.

[23] J.M. Frost, M.A. Faist, J. Nelson, Energetic Disorder in Higher Fullerene Adducts: A Quantum Chemical and Voltammetric Study, Adv. Mater. 22 (2010) 4881–4884. doi:10.1002/adma.201002189.

[24] Y. He, Y. Li, Fullerene derivative acceptors for high performance polymer solar cells, Phys. Chem. Chem. Phys. 13 (2011) 1970–1983. doi:10.1039/C0CP01178A.

[25] M.T. Greiner, M.G. Helander, W.-M. Tang, Z.-B. Wang, J. Qiu, Z.-H. Lu, Universal energy-level alignment of molecules on metal oxides, Nat. Mater. 11 (2011) 76–81. doi:10.1038/nmat3159.

[26] A.L. F. Cauduro, R. dos Reis, G. Chen, A.K. Schmid, C. Méthivier, H.-G. Rubahn, L. Bossard-Giannesini, H. Cruguel, N. Witkowski, M. Madsen, Crystalline Molybdenum Oxide Thin-Films for Application as Interfacial Layers in Optoelectronic Devices, ACS Appl. Mater. Interfaces. 9 (2017) 7717–7724. doi:10.1021/acsami.6b14228.

[27] M.T. Greiner, M.G. Helander, W.-M. Tang, Z.-B. Wang, J. Qiu, Z.-H. Lu, Universal energy-level alignment of molecules on metal oxides, Nat. Mater. 11 (2012) 76–81. doi:10.1038/nmat3159.

[28] Y. Masumoto, T. Mori, Application of organic bathocuproine-based alloy film to organic light-emitting diodes, Thin Solid Films. 516 (2008) 3350–3356. doi:10.1016/j.tsf.2007.11.082.

[29] P. Peumans, S. Uchida, S.R. Forrest, Efficient bulk heterojunction photovoltaic cells using small-

DBP-C70 Solar Cell

48

molecular-weight organic thin films, Nature. 425 (2003) 158–162. doi:10.1038/nature01949.

[30] M. Vogel, S. Doka, C. Breyer, M.C. Lux-Steiner, K. Fostiropoulos, On the function of a bathocuproine buffer layer in organic photovoltaic cells, Appl. Phys. Lett. 89 (2006) 163501. doi:10.1063/1.2362624.

[31] T. Sakurai, S. Toyoshima, H. Kitazume, S. Masuda, H. Kato, K. Akimoto, Influence of gap states on electrical properties at interface between bathocuproine and various types of metals, J. Appl. Phys. 107 (2010) 43707. doi:10.1063/1.3309278.

[32] J. Xue, S.R. Forrest, Carrier transport in multilayer organic photodetectors: II. Effects of anode preparation, J. Appl. Phys. 95 (2004) 1869–1877. doi:10.1063/1.1640454.

[33] K. Cnops, B.P. Rand, D. Cheyns, B. Verreet, M.A. Empl, P. Heremans, 8.4% efficient fullerene-free organic solar cells exploiting long-range exciton energy transfer, Nat. Commun. 5 (2014). doi:10.1038/ncomms4406.

[34] V.S. Gevaerts, L.J.A. Koster, M.M. Wienk, R.A.J. Janssen, Discriminating between Bilayer and Bulk Heterojunction Polymer:Fullerene Solar Cells Using the External Quantum Efficiency, ACS Appl. Mater. Interfaces. 3 (2011) 3252–3255. doi:10.1021/am200755m.

[35] A.M.A. Leguy, P. Azarhoosh, M.I. Alonso, M. Campoy-Quiles, O.J. Weber, J. Yao, D. Bryant, M.T. Weller, J. Nelson, A. Walsh, M. van Schilfgaarde, P.R.F. Barnes, Experimental and theoretical optical properties of methylammonium lead halide perovskites, Nanoscale. 8 (2016) 6317–6327. doi:10.1039/C5NR05435D.

[36] L. a a Pettersson, L.S. Roman, O. Inganäs, Modeling photocurrent action spectra of photovoltaic devices based on organic thin films Modeling photocurrent action spectra of photovoltaic devices based on organic thin films, J. Appl. Phys. 487 (2011) 487–496. doi:10.1063/1.370757.

[37] Y. Liu, K. Zojer, B. Lassen, J. Kjelstrup-Hansen, H.G. Rubahn, M. Madsen, Role of the Charge-Transfer State in Reduced Langevin Recombination in Organic Solar Cells: A Theoretical Study, J. Phys. Chem. C. 119 (2015) 26588–26597. doi:10.1021/acs.jpcc.5b08936.

Chapter 4 Molybdenum-Oxide Interlayers for Small Molecule Organic Solar Cells

Molybdenum-Oxide Interlayers for Small Molecule Organic Solar Cells

51

In this chapter, the results of using different deposition methods to fabricate Molybdenum oxide thin films for hole transport layers in solar cells will be shown. Considerations on the effect of different deposition methods on formation of films with different compositions and crystallinity will be shown in depth for the Molybdenum oxide films, and the impact of these films in OSC devices are studied. For the films, thickness and temperature are optimized for the different molybdenum oxide compositions. At the end of this chapter, primary results regarding the influence of these hole-transport layers on air-exposed OSC under constant illumination are demonstrated.

4-1 Sputtering Molybdenum Oxide film formation

Transition Metal Oxides (TMOs) such as Molybdenum oxide, Vanadium Oxide and Tungsten Oxide has been intensively used in organic electronic devices due to their high work functions, combined with their variety in film stoichiometry and molecular arrangement, and thus also their optoelectronic properties. Amongst these metal oxides, Molybdenum Oxide shows the largest bandgap[1], [2], relatively low deposition temperatures[3] and interesting electrical[4], optical[3] and physical properties[5]. Molybdenum oxide has been vastly used in several applications including chemical sensing[3], lithium ion batteries[6], electrodes[7], [8] and displays[7], [9], and they are an important part of organic semiconductor devices for instance Organic Light Emitting Diodes[10] and Organic Solar Cells[11]–[13]. The use of Molybdenum oxide films in organic electronics encompass amongst others p-type doping of organic materials via HOMO level electron transfer,[3] and hole extraction from both polymer[14] and small molecule[15] electron donors in organic solar cells. In solar cells, Molybdenum oxide (MoOx) have been used in different organic, inorganic and hybrid technologies as hole transport layer, demonstrating important technological developments on the power conversion efficiency as well as on the stability of different generation of solar cells[11], [16]–[18]. Among several different deposition methods available to fabricate MoOx thin-films are Spray Pyrolysis[19], Electrodeposition[20], Physical Vapor Deposition[21] and Sol-Gel[22] deposition techniques. In this project, vacuum deposition of MoOx using both magnetron sputtering and thermal evaporation are conducted. Each of these techniques have their advantages and drawbacks regarding film formation and deposition rate, as mentioned in chapter 1, and as it will be discussed further in this work. To have a complete understanding of the effect of each methods, both DC-sputtered and thermal evaporated films are developed and analyzed in terms of properties of the films, both individually, and in the device stack.

4-2 Molybdenum Oxide film formation and properties

We investigate the changes in optoelectronic properties of as-deposited (at room temperature) DC-sputtered MoOx films by tuning of the oxygen partial pressure (pO2) during the MoOx growth process. DC-sputtered MoOx films were deposited from a Molybdenum target (99.95 % from Kurt Lesker) on non-crystalline Fused Silica (BK7) glass substrates for transmittance analysis, on commercial ITO (Praezisions Glas & Optik GmbH) for conductivity measurements and on Silicon substrates for composition and microstructure analysis investigations, respectively. The surface roughness was analyzed by AFM in tapping mode (Veeco Dimension 3100) and the total (specular plus diffuse) optical transmittance and reflectance spectra were obtained using a UV-VIS-NIR Lambda 900 Perkin Elmer spectrophotometer. J-V curves have been taken using a Keithley 2400 series source meter. The active area of the fabricated devices is ranged between 2mm2 and 4mm2. The samples for conductivity measurements were capped with Silver (50 nm)/Gold (30 nm) as top cathodes. Both layers were deposited immediately after the DC - reactive

Molybdenum-Oxide Interlayers for Small Molecule Organic Solar Cells

52

sputtering process by thermal evaporation (E-beam) without breaking the high vacuum in between the depositions. The transmittance and reflectance spectra were taken shortly after breaking the vacuum in order to minimize surface oxidation. The thickness of the samples was about 300nm ±20nm for all investigations. The DC-sputtering process was carried out on mainly three different oxygen partial pressures (1.00 x 10-3 mbar, 1.98 x 10-3 mbar and 2.70x 10-3 mbar) and also at four different sputtering powers (100W, 150W, 200W and 250W).[23]

The deposition rate of the films decreased with increasing oxygen partial pressure for all investigations regardless of the sputtering pressure. Since the argon pressure was kept constant at 2.44x10-3 mbar during all depositions, it is likely that the oxygen causes target poisoning (Figure 4.1) and therefore reduces the deposition rate.[4] In this regime though, MoOx molecules are sputtered off instead of metallic Mo. The film-growth rate increases directly with the applied power in the DC sputtering process, which is in agreement with observations reported in the literature.[24]

Figure 4.1. the deposition rate versus the oxygen partial pressure at 250 W[25].

An AFM topography profile of the samples as-deposited on commercial ITO is shown in Figure 4.2. The surface roughness (Rrms) was calculated for the films grown at different oxygen partial pressures, and the values are very similar: The Rrms decreases from 2.20 nm (bare ITO) to roughly 1nm for all cases. This AFM analysis was also extended to different substrates such as BK7 glass and thermally grown SiO2 (not shown), where similar results are obtained: A low surface roughness (1.0-1.5 nm) is present, which also supports the presence of an amorphous film structure as seen by HRTEM.

It is known that poly-crystalline MoOx films are formed at higher substrate temperatures during growth, or through post-annealing processes. This has for example been shown by Nirupama et al.,[26] who reported on the influence of the oxygen partial pressure (ranging from 8.00x10-5 mbar up to 8.00x10-4 mbar) on as-deposited (at 473K) DC sputtered MoOx films. In their work, polycrystalline films were observed independently of the oxygen partial pressure, and in addition, an increase in surface roughness with the oxygen partial pressure was also reported. In the case of as-deposited films formed at room temperature, Sook Oh et al.[4] also found amorphous films at similar sputtering pressures as presented here, and they in addition reported on poly-crystalline films formed at ultra-low sputtering pressures. In this work, non-crystalline films are observed independently of the oxygen partial pressures in the range of 1.00x10-3 to 3.00x10-3 mbar.

Molybdenum-Oxide Interlayers for Small Molecule Organic Solar Cells

53

Figure 4.2. AFM relative height profiles of the as-deposited MoOx films on commercial ITO under different oxygen partial pressures (red line at 1.00x10-3 mbar, green line at 1.98x10-3 mbar and blue line at 2.70x10-3 mbar). The relative

height profile of the bare ITO surface is shown as a reference (solid black line)[25].

Absorption coefficient (α) spectra of the samples grown at 250W are presented in Figure 4.3 and were calculated using the approach described elsewhere.[27] As the oxygen partial pressure increases, α decreases in the range from 0.6-3.0 eV. The sub-stoichiometric MoO2.57 grown at 1.00x10-3 mbar of oxygen (black line) holds the stronger and the broader α, and the absorption extends from NIR to the UV. As the pO2 increases to 1.37x10-3 mbar, α dramatically decreases in the spectral range from 0.6-3.0eV, which is attributed to a change in the defect band of the MoOx. At that pressure, stoichiometric amorphous MoO3 is formed according to the RBS analysis, for which similar defect band centered at 1.5 eV for amorphous and polycrystalline MoO3 was reported by Sian et. al.[28] As the pO2 increases even further, e.g., higher than 1.37x10-3 mbar, the absorption coefficient up to 3.0 eV decreases further, and it is almost vanished at 2.70x10-3 mbar. We attribute this phenomenon to a modification of the d-band occupancy,[29] i.e., electrons initially at the defect states are released to the valence band as the oxygen content increases, and empty out the defect band. The more oxygen within the film, the less light absorption in the NIR-VIS range occurs. The inset in Figure 4.3 shows in more detail the changes on α for 1.98, 2.28 and 2.70x10-3 mbar (the curves were multiplied by a factor of 10 for a clear view). Due to the changes on absorption, the transmittance of the films has been modified from very low to semi-transparent films for photon energies between 0.6 eV and 3.0 eV.[23] We thus demonstrate, that small changes in the oxygen partial pressure (an increase of about 3.70x 10-4 mbar) presents remarkable changes on the optical properties of the as-deposited MoOx films, i.e., without altering the crystallization of the films through post-annealing methods. As seen in table I, the sub-stoichiometric MoOx as-deposited at an oxygen partial pressure of 1.00x10-3 mbar shows an O/Mo ratio of about 2.57 and yet a similar amorphous structure as seen in the films formed under higher pO2, as indicated in Figure 4.1(d) and also by HRTEM.[23] The optical band gap was calculated for all as-deposited samples using Tauc plot for non-crystalline semiconductors.[30] The optical band gap increases about 230 me.V from the sub-stoichiometric MoO2.57 phase (3.06 eV) to the nearly stoichiometric one MoO3.00 (3.29 eV). The optical band gap remains constant at 3.29 eV for the samples grown with oxygen partial pressure higher than 1.37x10-3 mbar.

Molybdenum-Oxide Interlayers for Small Molecule Organic Solar Cells

54

Figure 4.3. (left) Absorption coefficient (α) of 310 nm MoOx as-deposited films on BK7 substrate as a function of different oxygen partial pressures (1.00 x 10-3 mbar, 1.37 x 10-3, 1.98 x 10-3 mbar, 2.28 x 10-3 mbar and 2.70x 10-3 mbar) at 250 W. The Inset shows in more detail the absorption coefficient from 0.6 eV to 2.5 eV, and here, the 1.98, 2.28 and

2.70 x10-3 mbar samples were multiplied by a factor of 10. The discontinuity at 1.40 eV is an artifact of the measurement[25].(right) J-V curves of the devices under study as a function of the oxygen partial pressure. The inset

shows the device schematic used to measure the conductivity at sputtering power of 250W[25].

Conductivity measurements were performed in order to study the transport mechanisms at room temperature of the as-deposited films on glass/ITO (100 nm) followed by Ag (50 nm) and Au (30 nm) using the MoOx deposition series at 250W. The results from the devices investigated under direct bias (+ITO and -Au) are shown in Figure 4.3 (righ), which displays the J-V curves of the devices fabricated at 1.37x10-3 mbar (red open square), 1.98x10-3 mbar (green solid line), 2.28x10-3 mbar (black dashed line) and 2.70x10-3 mbar (dark blue solid line) of oxygen partial pressure, respectively. A schematic of the devices is shown as an inset in Figure 4.3. The J-V curve for the film with the lowest oxygen partial pressure of about 1.00x10-

3 mbar is not shown, since the resistance of the ITO layer dominates the resistance of the MoOx layer in this case. The film has been measured from 4-point probe immediately after venting the chamber, leading to a conductivity of about 3.22 S/cm.

The stoichiometric MoO3 (represented by red squares) exhibits an Ohmic current and the overall conductivity was estimated to be about 1.6x10-5 S/cm from the J~V1 (and confirmed also by 4-point probe method), which is in agreement with previous studies.[31], [32] In order to compare the resistivity of the whole set of samples, we have chosen to extract the conductivity under low bias (up to 200 mV). The conductivity of the sample grown under 1.98-2.28 x10-3 mbar of oxygen is about 1.50 x10-9 S/cm, while the one grown at 2.70x10-3 mbar of oxygen is 3.80x10-10 S/cm, indicating that both films are insulators. Conductivities in the range from 104 S/cm[4] down to 10-12 S/cm[33] have already been reported in the past years by reactive sputtering. As explained by Sian et al.[28] and Vasilopoulou et al.[34]. the electronic structure of sub-stoichiometric amorphous MoOx presents a defect band that arises due to the growth in an oxygen deficient environment, and electrons are trapped, occupying partly the d-d band located deep in the band gap. Thus, an upshift of the Fermi level occurs, leading to an increase in conductivity as the oxygen partial pressure is decreased during the growth. On the other hand, as the pO2 increases, the conductivity decreases as one modifies the occupancy of the defect states within the energy gap. Here, the results from both optical and electrical studies point out that the d-band occupancy changes systematically

Molybdenum-Oxide Interlayers for Small Molecule Organic Solar Cells

55

from a high density of defects (MoO2.57- high conductivity and absorption), an intermediate defect density state (MoO3.00) and reaching a little defect state (MoO3.16- low absorption, low conductivity, insulator character). We are here reporting on systematic changes of the optoelectronic properties of MoOx films by slightly tuning the oxygen partial pressures. For the films grown under 2.28 x10-3 mbar, hysteresis loops that are so far unseen for MoOx films appear.

4-3 Molybdenum oxide as interlayer for organic solar cells

Transition metal oxides’ energy alignment is related to their electron-chemical potential equilibrium for which they are used for efficient charge injection/collection from organic molecules to electrodes. The main property that makes transition metal oxides so interesting is their ability to exchange charge with condensed molecules. For high work function metal oxides, this ability requires that both a close alignment of their donor HOMO level with the metal-oxide work function level.

When metal-oxides face organic molecules, they thus have the ability to decrease the hole injection barrier at the interface, which result in a decrease in the contact resistance between the organic molecule and the respective electrode. Metal oxides can both affect the HOMO/LUMO level of the organic molecules and result in a better connection between organic layer and the respective electrode by either decreasing HOMO or increasing the LUMO. Due to their tenability in their physical, chemical and electronic properties, they have been widely used in those applications where electron injection/extraction properties required.[29]

The thickness of Molybdenum oxide transport films in an organic solar cell has been investigated by many groups, and many different thicknesses for optimum output characteristics have been reported.[15] The thickness of the transporting and active layers materials are also depending on each other, and due to many different materials combinations, people have reported MoOx thicknesses from 3nm to 50nm to provide the best performing solar cell performance.

As we have already determined regrind optimized thicknesses for our active and hole transport layer throughout chapter 3, we have here kept the thicknesses (f) our organic layers constant, in order to be able to only analyze the effects from the MoOx layer. We use the bilayer architecture due to:

• The Power Conversion Efficiency (PCE) for DBP/C70 system is reasonably high (4%), with relatively high solar cell parameters;

• There is no co-evaporation materials involved. As co-evaporation relates to the ratio between two materials in the final layer, 10% variation in ratio could dramatically change the performance of the final device making difficult to get reach conclusions on the MoOx layer.

• It is more simple and fast to fabricate bilayer compared to PIN (NIP) or BHJ.

4-3-1 Thermal evaporation of MoOx layers

We first investigate the effect of thickness variation in the performance of the OSC. Thermally evaporated MoOx has been used as hole-transporting material and we use 3 main thicknesses (10, 30 and 50nm) to find the thickness that provides the best overall performance focusing on the PCE. MoOx films have for this study been prepared using the cluster deposition system, and the transfer of samples for depositing metal

Molybdenum-Oxide Interlayers for Small Molecule Organic Solar Cells

56

oxides and organic layers was done through the transfer chamber (base pressure ~10-9mbar). Therefore, the MoOx layers have not been exposed to neither ambient nor glovebox environment before active layers are deposited on top of them.

Figure 4.4 shows the average JV characteristics and Table 4.1, their respective parameters as a function of different thicknesses for the thermally evaporated MoOx is shown. By varying the thickness of the MoOx layer, it is clear that the current density is the main parameter that is changed, followed by small Fill Factor changes. This change is due to the fact that by changing the transport layer thickness, mainly the series resistances of the whole stack change, in combination with low transmittance changes, resulting in the modified performance observed.

0.0 0.2 0.4 0.6 0.8 1.0-7

-6

-5

-4

-3

-2

-1

0

1

Cur

rent

Den

sity

(mA

/cm

2 )

Voltage (V)

10nm 30nm 50nm

Figure 4.4. (left) Organic solar cell layer thicknesses to investigate effects of MoOx film thickness variation: ITO/MoOx(x nm)/DBP(20nm)/C70(30)/BCP(10nm)/Ag(100nm). (right) Average PCE (black curve) variation due to the change in MoOx thickness change. The parameter which changes the most due to thickness change of MoOx is the current

density which is also shown in this plot (blue curve).

Table 4.1 includes the average performance of thermally evaporated MoOx with different thicknesses. It also shows the average current density which is the parameter that changes the most due to variation of MoOx thickness.

MoOx thickness (nm) Voc (V) Jsc (mA/cm2) FF (%) PCE (%) 10 0.88±0.29 6.00±0.52 59.2±5.7 3.14±0.54 30 0.87±0.28 6.20±0.52 59.1±6.0 3.21±0.52 50 0.86±0.35 5.60±0.35 56.3±5.7 2.71±0.44

Table 4.1. Average parameters of OSC where all layers have consistent thickness except MoOx.

Figure 4.5 shows the transmittance of the as-deposited thermally evaporated MoOx at the different thicknesses. MoOx has been deposited on plane ITO samples and has been encapsulated inside Nitrogen glovebox before measuring the transmittance. The effects from substrate and glass encapsulation have been deducted from the measurement and the data display only the MoOx film. Although the MoOx layer is very transparent in the visible regime, its absorption starts to increase as the thickness increases, which I will further elaborate on throughout this chapter.

10 20 30 40 500.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Mean PCE Mean Jsc

Thickness (nm)

Mea

n P

CE

(%)

4.0

4.5

5.0

5.5

6.0

6.5

7.0

7.5

8.0

Mea

n Js

c (m

A/c

m2 )

Molybdenum-Oxide Interlayers for Small Molecule Organic Solar Cells

57

350 400 450 500 550 600 650 700 7500

10

20

30

40

50

60

70

80

90

100

Tran

smitt

ance

(%)

Wavelength (nm)

10nm 30nm 50nm

Figure 4.5. Transmittance of thermal MoOx in different thicknesses. Hole-transport materials are usually broadband materials which are mostly transparent in the visible regime and MoOx shows reasonable transparency even for a

thicker film of 50nm.

4-3-2 Sputtered deposited MoOx layers

Sputtered films are prepared on pre-patterned ITO and plane ITO samples for OSC fabrication and for opto-electronic properties investigations. Molybdenum target (99.99%) from Kurt Lesker has been used in the DC magnetron sputtering chamber where Argon uses as the carrier gas and Oxygen was introduced during the deposition as mentioned at the beginning of this chamber. As mentioned earlier, we use 3 main compositions of sputtered Oxide with different oxygen concentration (sub-stoichiometric MoO2.7, stoichiometric MoO3.0 and super-oxide MoO3.2). The deposition power has been kept constant at 250W for the MoOx which has been used for OSC fabrication, as we have already shown that major composition changes have been occurring at this power. During sputtering the substrate temperature has been kept at 20oC, and during organic evaporation, the substrate temperature was kept under careful monitoring and it did not exceeded 22oC. As the DC sputtering chamber is placed in the MCI cleanroom, all sputtering samples were transferred in ambient air before placed inside a Nitrogen glovebox prior to Organic depositions.

As in the case of thermally evaporated MoOx, we first investigate the effects of thickness changes, here for all 3 compositions of the Molybdenum oxide films. We have chosen 3 main thicknesses for each composition and fabricated OSC with the optimized active layer thickness from chapter 3. The results of the PCE of each composition in comparison with thermal molybdenum oxide in different thickness are shown in figure 4.6. Table 4.2 shows all the parameters of all compositions using 3 different thicknesses including thermal as reference.

Molybdenum-Oxide Interlayers for Small Molecule Organic Solar Cells

58

MoO2.7 MoO3.0 MoO3.2 MoOTh

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

PC

E (%

)

10nm 30nm 50nm

MoO2.7 MoO3.0 MoO3.2 MoOTh

0

10

20

30

40

50

60

70

FF (%

)

10nm 30nm 50nm

MoO2.7 MoO3.0 MoO3.2 MoOTh

0

1

2

3

4

5

6

7

8

9

Jsc

(mA

/cm

2 )

10nm 30nm 50nm

MoO2.7 MoO3.0 MoO3.2 MoOTh

0.0

0.2

0.4

0.6

0.8

1.0

Voc

(V)

10nm 30nm 50nm

Figure 4.6. OSC performance parameters using different molybdenum oxide compositions, for 3 different film thicknesses. The 10nm shows almost no performance dependency on composition, while for thicker films, the film

performance dramatically changes in different compositions.

Molybdenum-Oxide Interlayers for Small Molecule Organic Solar Cells

59

Thickness 10 30 50 10 30 50 10 30 50 10 30 50

Composition PCE(%) Jsc(mA/cm2) Voc(V) FF(%)

MoO2.7 2.98±0.94 2.37±0.25 2.43±0.24 6.73±0.28 5.09±0.18 4.91±0.36 0.80±0.08 0.87±0.01 0.85±0.02 54.1±13.5 53.9±4.9 58.4±3.2

MoO3.0 2.89±0.56 2.70±0.86 2.62±0.43 6.72±1.43 6.65±1.68 5.75±0.48 0.86±0.03 0.80±0.05 0.84±0.05 51.4±10.5 50.5±4.9 53.8±4.6

MoO3.2 3.09±0.95 0.09±0.11 0.21±0.62 7.04±1.67 1.28±0.91 1.38±0.37 0.83±0.06 0.53±0.22 0.80±0.05 52.6±6.2 15.0±4.1 19.0±0.7

MoOTh 3.14±0.54 3.21±0.52 2.71±0.44 6.00±0.52 6.20±0.52 5.60±0.35 0.88±0.03 0.87±0.03 0.86±0.04 59.2±5.7 59.1±6.0 56.2±5.7

Table 4.2. All performance parameters for all MoOx thicknesses of 3 different sputtering compositions, in comparison with thermal molybdenum oxide. The OSC structure is: ITO/MoOx (Xnm)/DBP (30nm)/C70 (30nm)/ BCP (10nm)/Ag.

Molybdenum-Oxide Interlayers for Small Molecule Organic Solar Cells

60

Generally, TMOs with higher conductivity have higher optical absorption, which is due to the fact that by reducing the TMO, oxygen vacancies that are responsible for the optoelectronic properties are formed (Figure 4.3). The Transmittance of the as-deposited sputtered films is shown in Figure 4.7. It is seen that as the film thickness increases, the transmittance shows a reduction in transmittance, as expected. It is also seen that the thermal evaporated films absorb more in wavelengths above 500nm for thicker films, which potentially impact OSC performance.

350 400 450 500 550 600 650 700 7500

20

40

60

80

100

Tran

smitt

ance

(%)

Wavelength (nm)

MoOTh (10nm) MoOTh (30nm) MoOTh (50nm) MoO3.2 (10nm) MoO3.2 (30nm) MoO3.2 (50nm)

Figure 4.7. To avoid confusion in the plot, we only show the extreme case in sputtering (maximum oxygen concentration) in comparison to the thermal transmittance of the molybdenum oxide film.

4-3-3 Post-Annealing of the Molybdenum Oxide Films

The effect of film annealing on the film crystallinity, energy level and optoelectrical properties of the thermal evaporated and sputtered films (mainly MoO3.2), has been investigated Figure 4.3 [25] to shed light on the advantages of using annealed molybdenum oxide films in the fabrication of OSC. Later in this chapter, we show the positive impact of this method on the stability of non-encapsulated OSC to further emphasize on the advantages of these films.

Thermal and sputtered samples have been prepared in different compositions and thicknesses on pre-patterned and plane ITO samples. Samples have then been transferred to the Nitrogen filled glovebox in connection to the ultra-high cluster deposition system, to be used either in annealed form or as-deposited form. The samples afterwards were placed in a sample holder for annealing at different temperature under ultra-high vacuum (~10-8mbar). The annealing temperature was 200, 350 and 500oC. Different steps were employed to reach the set temperature, in order to avoid sudden thermal shock of the samples and the MoOx films, these steps are:

Molybdenum-Oxide Interlayers for Small Molecule Organic Solar Cells

61

• 30min kept in ultra-high vacuum • Ramp up to 90oC in 10min • Stay at 90oC for 30min • Ramp up to 150oC in 10min • Stay at 150oC for 30min • Ramp to the set temperature during 15min • Stay at the set temperature for 10min • Cool down for 8hours in ultra-high vacuum (~10-8mbar)

Organic solar cells have been made with these MoOx films to study the best performing temperature for the film. Figure 4.8 shows the performance of 3 different thicknesses, and 4 different compositions of molybdenum oxide films, including both thermal and sputter deposited films. At 500oC which is close to the glass transition temperature of our substrates, plus the roughness of the films has excessively grow at this temperature as formerly reported by our group[35] we see that the samples are deformed and most samples show short circuit. Although other groups have reported on 500oC being reasonable temperatures to form crystalline molybdenum oxide from[36], [37], they anneal ambient conditions, whereas these films are annealed under ultra-high vacuum.

0 50 100 150 200 250 300 350 400 450 500 5500.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

Temperature (oC)

10nm MoOx

Mea

n P

CE

(%)

MoO2.7

MoO3.0

MoO3.2

MoOTh

Molybdenum-Oxide Interlayers for Small Molecule Organic Solar Cells

62

Figure 4.8. PCE of DBP/C70 solar cell with 10, 30 and 50nm MoOx thicknesses for 3 different compositions for sputter deposited films, plus thermally evaporated films. For all films, post-annealing at 3 different elevated temperatures are

investigated (as-deposited, 200, 350 and 500oC)

As outlined above, Figure 4.8 shows the performance of the DBP/C70 based OSC devices employing hole transport layers from 30nm thermally deposited MoOx (MoOTh) and 30nm reactive sputtered MoOx (MoOSp) with and without vacuum post-annealing. Besides the MoO3.2 devices, all MoOx films lead to reasonably good performance for the non-annealed samples. As for 10nm films, only the sputtered stoichiometric films show good performance at elevated temperature, and in general, the thermal deposited MoOx show a drop in performance with temperature. Interestingly, considering the 30nm and 50nm thicknesses, devices using as-deposited MoO3.2 shows very poor device performance, however, the annealed MoO3.2 devices shows promising characteristics that compares to that of the as-deposited MoOTh films. To clarify this, we investigate these films further in terms of electrical properties of the films, the

0 50 100 150 200 250 300 350 400 450 500 5500.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Mea

n P

CE

(%)

Temperature (oC)

MoO2.7

MoO3.0

MoO3.2

MoOTh

MoOx 30nm

0 50 100 150 200 250 300 350 400 450 500 5500.00.30.60.91.21.51.82.12.42.73.03.33.6

Mea

n P

CE

(%)

Temperature (oC)

MoO2.7

MoO3.0

MoO3.2

MoOTh

MoOx 50nm

Molybdenum-Oxide Interlayers for Small Molecule Organic Solar Cells

63

microstructure, morphology and optical properties of these films have already shown in Fig. 4.10, 4.11 and 4.12 and they will be called MoOSp from now on.

Regarding crystallinity investigations of the films, thermal MoOx and super-oxide MoO3.2 has been employed from X-Ray Diffraction investigation (XRD), conducted at the collaborator Prof. R. Resel, at Technical University of Graz. XRD has been performed on 30nm thick films, which have been deposited on ITO. Figure 4.9 shows the XRD spectra of the aforementioned films.

XRD of plane ITO samples display the peaks related to the ITO layer and substrate and it has been (Figure 4.9, brown) which has been reproduced for all other samples. For thermal film which has been annealed at 2 temperatures, there is no peak other than substrates’ one (Figure 4.9, pink and blue) whereas sputter film has started to show tiny extra peak around 200oC (Figure 4.9, green) and at 350oC the film clearly has a crystalline orientation as marked in Figure 4.9. There are similar reports on desirable crystalline orientation of Molybdenum oxide films using different methods where they show that crystalline orthorhombic α-phase MoO3 will form which exhibits reflection peaks at (0 2 1) in XRD profile [21], [38], [37].

Figure 4.9. XRD of 30nm thermal vs. super-oxide sputtered MoOx as-deposited and annealed at 350oC where it shows distinctive peaks for sputtering one when annealed. Measurements conducted by Prof. R. Resel, at Technical University

of Graz.

MoOx peaks that appear for sputtered films annealed at 350oC

ITO ITO/MoOTh @200oC ITO/MoOTh @350oC ITO/MoOSp @200oC ITO/MoOSp @350oC

Molybdenum-Oxide Interlayers for Small Molecule Organic Solar Cells

64

Tapping mode AFM images provides information about the surface morphology of the films that are shown in Figure 4.10. Clearly, morphological differences between the super-oxidized sputtered MoO3.2 film and thermally evaporated MoOx are seen (Figure 4.10), both for samples deposited at 200oC (Fig. 4.10A, C) and at 350oC (Fig. 4.10 B, D). For the 350 annealed films, the grain sizes of the sputtered films are larger, also indicating the (poly) crystalline nature as seen from the XRD data.

Figure 4.10. Surface topography of 30 nm Molybdenum oxide films on ITO at different post-growth temperatures: A) MoOTh at 200oC B) MoOTh at 350oC C) MoOSp at 200oC D) MoOSp at 350oC. It becomes evident that no substantial

effects on the film morphology occur for post-growth treatments in vacuum at 200oC. Scale bar is 1 μm.

Transmittance spectra of the films at different annealing temperatures also show a change in the optical absorption of the films. From figure 4.3 we recognize that among all the films compositions, the most dramatic changes occur at wavelength around 830nm. For the transmittance analysis we study the optical properties at this wavelength, and at a wavelength equal to 600nm, where the involved active layers show maximum absorption. Figure 4.11 shows the transmittance of the films in these two wavelengths. Clearly for higher annealing temperatures, the transmittance around 600nm in the two films is almost similar, however, the thermal evaporated films made at room temperature show larger transmittance, which could

Molybdenum-Oxide Interlayers for Small Molecule Organic Solar Cells

65

benefit the OSC. At 830nm, sputtered films are more transparent than the thermal evaporated ones, indicating a low defect density in these films.

0 50 100 150 200 250 300 350 400

74

76

78

80

82

84

86

88

90

92

Tranmittance of 30nm MoOx at 600nm

Tran

smitt

ance

(%)

Temperature (oC)

MoOTh

MoO3.2

0 50 100 150 200 250 300 350 40080

82

84

86

88

90

92

94

96

98

100Transmittance of 30nm MoOx at 830nm

Tran

smitt

ance

(%)

Temperature (oC)

MoOTh

MoO3.2

Figure 4.11. Transmittance of Thermal and super-oxide Molybdenum oxide at 2 different annealing temperatures at 2 wavelengths, namely 600nm where the organic active layers have maximum absorption, and at 830nm where

Molybdenum oxide absorbs the most, according to Figure 4.3[25]

In table 4.3, the results from the sputtered MoO3.2 are presented. Clearly the maximum performance for the thermal evaporated films is observed at room temperature, whereas it is at 350oC for the sputtered films. It is also clear that the performance for the thermal evaporated room temperature MoOx device is very similar to that seen for the sputtered MoOx (at 350oC), where the device from the sputtered MoOx shows slightly higher Jsc and FF, but slightly lower Voc. In a separate study, the work functions of these films deposited on Silicon have been analyzed [cite Cauduro]. From that study, it was clear that high work function values in these films are obtained when the films are crystalized. In this work, we have also shown that the films crystalize on ITO surfaces, see Fig. 4.10, and it could be expected that such high work function could be obtained for these crystalline films as well, which in fact could explain the small FF and Jsc rise, compared to thermally evaporated films.

Figure 4.12 emphasize on the equal current density of 30nm MoOx for both as-deposited thermal and super-oxide @350oC through their EQE. The spectral spectrum of both OSC’s using these HTLs show almost same level where the peaks/valise match the absorption spectrum of the active layer (no generation from MoOx).

Composition Temp. (oC) PCE (%) Jsc (mA/cm2) FF (%) Voc (V)

MoOTh

30 3.21 ± 0.52 6.20 ± 0.52 59.1 ± 6.0 0.87 ± 0.03 200 2.45 ± 0.17 4.71 ± 0.18 60.4 ± 3.4 0.86 ± 0.01 350 2.20 ± 0.01 5.83 ± 0.02 57.5 ± 0.1 0.66 ± 0.01 500 0.23 ± 0.09 4.20 ± 0.10 38.5 ± 8.3 0.14 ± 0.03

MoOSp

30 0.09 ± 0.11 1.28 ± 0.91 15.0 ± 4.1 0.53 ± 0.22 200 2.71 ± 0.44 5.05 ± 0.82 61.7 ± 0.1 0.87 ± 0.00 350 3.25 ± 0.26 6.70 ± 0.44 61.1 ± 4.2 0.80 ± 0.06 500 0.04 ± 0.04 4.86 ± 0.45 21.4 ± 16.6 0.04 ± 0.01

Table 4.3 OSC parameters regarding the use of MoOx as HTL with 30nm of both Sputter super-oxides annealed at 350

oC and Thermal as deposited.

Molybdenum-Oxide Interlayers for Small Molecule Organic Solar Cells

66

4.12 EQE spectrum of 30nm MoOx as hole transport layer in 2 different compositions: as-deposited MoOTh and super-oxide annealed at 350oC.

In order to study further the electrical properties of the films, Figure 4.13 shows conductivity measurements of the 30nm as-deposited sputtered and thermal films at different annealing conditions. The results show that although the as-deposited sputtered films demonstrate rather poor conductivity, the conductivity raises almost 1 order of magnitude upon annealing, and higher than for as-deposited MoOTh.

0.01 0.1

1

10

Cur

rent

Den

sity

(mA

/cm

2 )

Voltage (V)

MoSp@30oC

MoSp@350oC

MoTh@30oC

Figure 4.13 Conductivity of MoOx film which shows an increase in the conductivity of sputtered upon annealing pushing it higher that as-deposited MoOTh. The structure for this plot is: ITO/MoOx 100nm/Ag

From the device studied, it is clear that Molybdenum oxide films developed from reactive sputtering can lead to high device performance in organic solar cells. The developed devices using super-oxidized MoO3.2

show performance parameters that are almost similar to those seen from standard used thermal evaporated Molybdenum oxide films. The MoO3.2 devices show slightly higher Jsc, although the films are

350 400 450 500 550 600 650 7000

5

10

15

20

25

30

35

40

45

MoOTh @ 30oC MoOSp @ 350oC

Wavelength (nm)

EQ

E (%

)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

DBP/ C70/ BCP

Abs

orpt

ion

(A.U

.)

Molybdenum-Oxide Interlayers for Small Molecule Organic Solar Cells

67

less transparent at visible wavelengths compared to thermally evaporated films, together with a slightly higher fill factors. This can be explained by both the high work function from these films [35], leading to a larger build-in field, and the improved conductivity of the films. All together this leads to improved hole contacts in the investigated DBP/C70 based devices. The large work function comes partially from the crystalline layers formed, which are seen both in AFM and XRD analysis.

4-4 Stability of OSC devices employing sputtered MoOx layers

The stability analysis focuses on comparing and understanding the stability properties of devices incorporating different Molybdenum oxide hole transport layers. Therefore, we first investigate the stability of the Molybdenum oxide layers themselves, by pre-treating the layers (deposited on ITO) in 4 different ways, before devices are finalized on top of the layers. The treatments are:

• Annealing under vacuum at 60oC; as 60oC is the temperature of cells measured under the SDU NanoSYD lifetime solar simulator

• Placement under 1 sun irradiation inside solar simulator which includes both heat (60oC) and light treatment

• Ambient exposure with 50% humidity in presence of oxygen • Freshly prepared MoOTh The result of the Molybdenum oxide pre-treatment experiments is shown in Figure 4.14. Clearly, the films placed under light soaking shows the worst performance, and the main parameter leading to a PCE drop is the Fill-Factor. The other pre-treatments also show impact on the devices, compared to the fresh MoOTh, but not as dramatic as the light-soaked sample. The drop in PCE performance under light exposure for OSC employing MoOth have been reported on before, where was shown that the formation of Mo5+ species due to Molybdenum oxide light exposure[39] are responsible for lowering the Molybdenum Oxide work function, leading to energy misalignment, and thus hole contact problems. This could explain the performance shown here for the light treated sample. Table 4.3 shows the parameters regarding the usage of as-deposited 10nm MoOTh which has been kept in different conditions to show the impact of hole-transport layer on the performance of OSC. The thicknesses of the OSC layers have been taken from chapter 3.

Molybdenum-Oxide Interlayers for Small Molecule Organic Solar Cells

68

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9-10-9-8-7-6-5-4-3-2-1012

Cur

rent

Den

sity

(mA/

cm2 )

Voltage (V)

Fresh Annealed under vacuum Light Soaked Degraded

Figure 4.14 Comparison between the conditions of hole-transporting layer on the performance of OSC based on chapter 3 optimized thicknesses.

10nm MoOTh conditions PCE(%) Jsc(mA/cm2) Voc(V) FF(%)

Fresh 3.16 5.76 0.86 63

Degraded (Constant measuring) 2.09 5.17 0.76 53

Annealed@60C Under Vacuum(5 days) 2.62 4.90 0.88 61

MoOx light soaked (5 days) 1.50 5.26 0.81 35

Table 4.3 Average OSC parameters using 10nm thermal molybdenum oxide layer which has kept under different condition.

After investigated the effect of different pre-treatments of MoOx on the device stability, the effect of using different films inside the OSC while degrading complete devices are investigated. The fact that the devices that base on the MoO3.2 films is improving with annealing temperature could have important implications on the overall device stability. For stability test of the devices, we performed ISOS3-L1 degradation under 1 sun illumination and tracked the performance parameters over time. It is clear that the devices based on MoO3.2 show an improved device stability compared to those developed from the MoOTh films, and an impressive 90% of the initial PCE for the MoO3.2 devices is maintained after more than 120 hours of light soaking under 1 sun (1000W/m2) at ~60°C figure 4.15, for non-encapsulated devices. This is in clear contrast to the MoOTh based devices, which at best (for 10 nm film thickness) is at 50% device performance after the same duration. Clearly, film changes take place in the initial light soaking phase of the MoO3.2 based devices, which results in drops mainly in Fill-Factor. However, following a recovery process, the MoO3.2 devices show superior stability compared to the MoOTh devices, which show clear drops in mainly current density over time. Since the drop in the performance of MoOth based devices can be related to the reduction of the oxide during light exposure, it’s tempting to assume that a smaller reduction of the Molybdenum oxide under light exposure is taking place in the sputtered films.

Molybdenum-Oxide Interlayers for Small Molecule Organic Solar Cells

69

0 24 48 72 96 120

20

40

60

80

100

Effi

cien

cy (%

)

Time (Hour)

10nm MoOSp @ 350oC 10nm MoOTh @ 30oC 30nm MoOSp @ 350oC 30nm MoOTh @ 30oC

Figure 4.15 Light soaking non-encapsulated OSC measure every 2minutes for 5 consecutive days using different hole-transporting layers.

Molybdenum-Oxide Interlayers for Small Molecule Organic Solar Cells

70

4-5 Reference

[1] J. Meyer, A. Shu, M. Kröger, and A. Kahn, “Effect of contamination on the electronic structure and hole-injection properties of MoO3 /organic semiconductor interfaces,” Appl. Phys. Lett., vol. 96, no. 13, pp. 1–4, 2010.

[2] V. Shrotriya, G. Li, Y. Yao, C.-W. Chu, and Y. Yang, “Transition metal oxides as the buffer layer for polymer photovoltaic cells,” Appl. Phys. Lett., vol. 88, no. 7, p. 73508, 2006.

[3] J. Meyer, S. Hamwi, M. Kröger, W. Kowalsky, T. Riedl, and A. Kahn, “Transition metal oxides for organic electronics: Energetics, device physics and applications,” Adv. Mater., vol. 24, no. 40, pp. 5408–5427, Oct. 2012.

[4] M. Sook Oh et al., “Improvement of electrical and optical properties of molybdenum oxide thin films by ultralow pressure sputtering method,” J. Vac. Sci. Technol. A Vacuum, Surfaces, Film., vol. 30, no. 3, p. 31501, May 2012.

[5] V. E. (Cambridge U. Henrich and P. A. (Cambridge U. Cox, “The Surface Science of Metal Oxides,” Adv. Mater., vol. 7, no. 1, pp. 91–92, Jan. 1995.

[6] Y. Sun et al., “Morphosynthesis of a hierarchical MoO2 nanoarchitecture as a binder-free anode for lithium-ion batteries,” Energy Environ. Sci., vol. 4, no. 8, p. 2870, 2011.

[7] H. Kim, K. T. Lee, C. Zhao, L. J. Guo, and J. Kanicki, “Top illuminated organic photodetectors with dielectric/metal/dielectric transparent anode,” Org. Electron. physics, Mater. Appl., vol. 20, pp. 103–111, 2015.

[8] M. F. Al-Kuhaili, S. H. A. Ahmad, S. M. A. Durrani, M. M. Faiz, and A. Ul-Hamid, “Energy-saving spectrally-selective coatings based on MoO3/Ag thin films,” Mater. Des., vol. 73, pp. 15–19, 2015.

[9] A. Arfaoui, S. Touihri, A. Mhamdi, A. Labidi, and T. Manoubi, “Structural, morphological, gas sensing and photocatalytic characterization of MoO3 and WO3 thin films prepared by the thermal vacuum evaporation technique,” Appl. Surf. Sci., vol. 357, pp. 1089–1096, 2015.

[10] H. You, Y. Dai, Z. Zhang, and D. Ma, “Improved performances of organic light-emitting diodes with metal oxide as anode buffer,” J. Appl. Phys., vol. 101, no. 2, pp. 2005–2008, 2007.

[11] Y. Zang, J. Huang, H. Li, J. Yu, and Y. Jiang, “Effect of molybdenum oxide anode buffer layer on the performance of inverted small molecular organic solar cells,” Energy Procedia, vol. 12, pp. 513–518, 2011.

[12] W. Zeng, K. S. Yong, Z. M. Kam, Z. Chen, and Y. Li, “Effect of MoO3 as an interlayer on the performance of organic solar cells based on ZnPc and C60,” Synthetic Metals, vol. 161, no. 23–24. pp. 2748–2752, 2012.

[13] J. Griffin, A. J. Pearson, N. W. Scarratt, T. Wang, D. G. Lidzey, and A. R. Buckley, “Organic photovoltaic devices incorporating a molybdenum oxide hole-extraction layer deposited by spray-coating from an ammonium molybdate tetrahydrate precursor,” Org. Electron., vol. 15, no. 3, pp. 692–700, Mar. 2014.

[14] Y. Sun et al., “Efficient, air-stable bulk heterojunction polymer solar cells using MoO(x) as the anode interfacial layer.,” Adv. Mater., vol. 23, no. 19, pp. 2226–30, May 2011.

Molybdenum-Oxide Interlayers for Small Molecule Organic Solar Cells

71

[15] I. Hancox et al., “The effect of a MoOx hole-extracting layer on the performance of organic photovoltaic cells based on small molecule planar heterojunctions.” 2010.

[16] J. Griffin, D. C. Watters, H. Yi, A. Iraqi, D. Lidzey, and A. R. Buckley, “The Influence of MoO x Anode Stoicheometry on the Performance of Bulk Heterojunction Polymer Solar Cells,” Adv. Energy Mater., vol. 3, no. 7, pp. 903–908, Jul. 2013.

[17] Y. Kanai, T. Matsushima, and H. Murata, “Improvement of stability for organic solar cells by using molybdenum trioxide buffer layer,” Thin Solid Films, vol. 518, no. 2, pp. 537–540, 2009.

[18] A. Soultati et al., “The role of metal/metal oxide/organic anode interfaces in efficiency and stability of bulk heterojunction organic photodetectors,” Microelectron. Eng., vol. 117, pp. 13–17, 2014.

[19] P. . Patil and P. . Patil, “Preparation of mixed oxide MoO3–WO3 thin films by spray pyrolysis technique and their characterisation,” Thin Solid Films, vol. 382, no. 1, pp. 13–22, 2001.

[20] D. Guan, J. Li, X. Gao, and C. Yuan, “Controllable synthesis of MoO3-deposited TiO2 nanotubes with enhanced lithium-ion intercalation performance,” J. Power Sources, vol. 246, pp. 305–312, 2014.

[21] R. K. Sharma and G. B. Reddy, “Effect of substrate temperature on the characteristics of α-MoO3 hierarchical 3D microspheres prepared by facile PVD process,” J. Alloys Compd., vol. 598, pp. 177–183, 2014.

[22] W. Dong, A. N. Mansour, and B. Dunn, “Structural and electrochemical properties of amorphous and crystalline molybdenum oxide aerogels,” Solid State Ionics, vol. 144, no. 1–2, pp. 31–40, 2001.

[23] See supplementary material at [to be inserted by AIP] for further information about the sputtering process and also about the microstructure of the MoOx grown in oxygen defficient environment. .

[24] M. Ohring, Materials Science of Thin Films: deposition and structure, 2nd ed., vol. 40, no. 6. San Diego: Academic Press, 2002.

[25] A. L. Fernandes Cauduro et al., “Tuning the optoelectronic properties of amorphous MoO x films by reactive sputtering,” Appl. Phys. Lett., vol. 106, no. 20, p. 202101, May 2015.

[26] V. Nirupama, K. R. Gunasekhar, B. Sreedhar, and S. Uthanna, “Effect of oxygen partial pressure on the structural and optical properties of dc reactive magnetron sputtered molybdenum oxide films,” Curr. Appl. Phys., vol. 10, no. 1, pp. 272–278, Jan. 2010.

[27] M. Cesaria, a P. Caricato, and M. Martino, “Realistic absorption coefficient of ultrathin films,” J. Opt., vol. 14, no. 10, p. 105701, 2012.

[28] T. S. Sian and G. B. Reddy, “Optical, structural and photoelectron spectroscopic studies on amorphous and crystalline molybdenum oxide thin films,” Sol. Energy Mater. Sol. Cells, vol. 82, no. 3, pp. 375–386, May 2004.

[29] M. T. Greiner, M. G. Helander, W.-M. Tang, Z.-B. Wang, J. Qiu, and Z.-H. Lu, “Universal energy-level alignment of molecules on metal oxides,” Nat. Mater., vol. 11, no. 1, pp. 76–81, Nov. 2011.

[30] J. Tauc and A. Menth, “States in the gap,” J. Non. Cryst. Solids, vol. 8, pp. 569–585, 1972.

[31] T. Matsushima, Y. Kinoshita, and H. Murata, “Formation of Ohmic hole injection by inserting an ultrathin layer of molybdenum trioxide between indium tin oxide and organic hole-transporting layers,” Appl. Phys. Lett., vol. 91, no. 25, p. 253504, 2007.

Molybdenum-Oxide Interlayers for Small Molecule Organic Solar Cells

72

[32] T. Matsushima and C. Adachi, “Enhanced hole injection and transport in molybdenum-dioxide-doped organic hole-transporting layers,” J. Appl. Phys., vol. 103, no. 3, p. 34501, 2008.

[33] N. Miyata and S. Akiyoshi, “Preparation and electrochromic properties of rf-sputtered molybdenum oxide films,” J. Appl. Phys., vol. 58, no. 4, p. 1651, 1985.

[34] M. Vasilopoulou et al., “The influence of hydrogenation and oxygen vacancies on molybdenum oxides work function and gap states for application in organic optoelectronics.,” J. Am. Chem. Soc., vol. 134, no. 39, pp. 16178–87, Oct. 2012.

[35] A. L. F. Cauduro et al., “Crystalline Molybdenum Oxide Thin-Films for Application as Interfacial Layers in Optoelectronic Devices,” ACS Appl. Mater. Interfaces, vol. 9, no. 8, pp. 7717–7724, Mar. 2017.

[36] V. Nirupama and S. Uthanna, “Investigations on morphological and electrical studies of sputtered MoO3 films,” J. Mater. Sci. Mater. Electron., vol. 27, no. 4, pp. 3668–3674, Apr. 2016.

[37] A. Arfaoui, B. Ouni, S. Touihri, A. Mhamdi, A. Labidi, and T. Manoubi, “Effect of annealing in a various oxygen atmosphere on structural, optical, electrical and gas sensing properties of MoxOy thin films,” Opt. Mater. (Amst)., vol. 45, pp. 109–120, 2015.

[38] S. Uthanna, V. Nirupama, and J. F. Pierson, “Substrate temperature influenced structural, electrical and optical properties of dc magnetron sputtered MoO3 films,” Appl. Surf. Sci., vol. 256, no. 10, pp. 3133–3137, Mar. 2010.

[39] H. Zhang et al., “Photochemical Transformations in Fullerene and Molybdenum Oxide Affect the Stability of Bilayer Organic Solar Cells,” Adv. Energy Mater., vol. 5, no. 2, p. 1400734, Jan. 2015.

Chapter 5 Transport Layers for Perovskite Solar Cell

Transport Layers for Perovskite Solar Cells

75

A new promising hybrid solar cell belonging to the class of third-generation solar cells (Thin-film solar cells) is Perovskite Solar Cells (PSCs), which are built on the basic concept of Dye-Sensitized solar cells (DSSC)[1], however, adopting solid films as in the case of OSCs. In PSC, the DSSC dye has been replaced by a Perovskite crystalline layer, and the transport layer, which in DSSC is a mesostructured metal-oxide to increase the interaction area between dye and active material, has been replaced by planar inorganic or organic transport layers[2].

Perovskite is a crystalline structure, which is formed by the absorber (active material) of PSCs. This crystal can be formed by a simple solution process technique such as from spin coating[3]. In most cases, the perovskite structure if formed by using a lead halogen compound, and an organic part, which is an amine derivative. The most used materials to form PSCs is methyl ammonium iodide (Ch3NH3I or in short MAI) as the organic part and PbI2 as lead salt[4–7]. The final structure of the PSC is Ch3NH3PbI3 or as it is mostly referred to MAPI[8].

5-1 Perovskite solar cells

Both Lead-salt and the organic precursor can be dissolved in a polar solvent (usually Dimethylfomimide DMF), which produces a yellow film when deposited as a thin-film. The film has to be annealed in order to form a shiny dark brown perovskite structure. This method is called 1-step process, which gives a poor crystallinity and low performance[9]. Because of the low PCE of such solar cells, the 2-step process was introduced in which the lead salt is deposited separately from the organic part:

1. Lead-salt (PbI) dissolved in a DMF base solvent is deposit on a pre-patterned TCO substrates 2. Subsequently, the organic part is deposited on the substrate. In this step, the perovskite crystalline

layer is formed[8] 3. The sample is then annealed at a moderate temperature in order to remove the excessive solvent, and

additionally help forming the Perovskite crystal[10]. Modification is possible in any of these steps, which can result in a different crystalline structure and at the end different properties and performance of PSC, which will be further elaborated throughout this chapter. Modifications could be categorized into these few families:

• Replacing Iodine with other halogens (like bromide)[11] • Solvent engineering which modifies the evaporation temperature and allows the film to form larger

crystals • Modifying deposition recipes creating thicker or thinner PSC films • Replacing the organic precursor and anti-solvent

In this chapter, results related to experiments conducted in collaboration with Abengoa Research, Seville, Spain and IMEC, Leuven Belgium is presented.

The first step towards understanding charge generation inside PSCs is to understand its optical properties. Figure 5.1 compares the absorption of three typical Perovskite absorbers. The smaller the halide atomic number gets the larger the shift towards UV in the absorption spectrum, as it can be seen in figure 5.1. Since MAPI absorbs almost the whole visible spectrum, where the sun intensity is strongest, it generally leads to higher performance in comparison with the other two perovskite types, mainly due to larger short-

Transport Layers for Perovskite Solar Cells

76

circuit currents, whereas the other two types typically have higher output voltage due to a wider bandgap[12]. It has been shown that MAPI Perovskite is a direct bandgap semiconductor with a bandgap of 1.6 eV, which can be calculated from absorption spectrum[13].

Figure 5.1 Absorption coefficient of three typically used Perovskite absorbers (a)MAPbI3. (b)MAPbBr3. (c)MAPbCl3[12]

From an architecture perspective, there are two main families of PSCs: NIP architecture and PIN. Both perovskite architectures are consisting of a TCO bottom electrode. For NIP, a thin layer of TiO2 as electron transport layer is formed on top, which will act as the electron transport layer, which usually has a thickness less than 100nm. TiO2 is usually deposited in form of small nanocrystals, which is a porous layer that in DSSC cells is filled with the absorber, which will accelerate the extraction of generated charges. The next layer, as already indicated, is the absorber formed from the 1 or 2-step processing. The hole-extraction layer in this structure is usually 2,2',7,7'-Tetrakis[N,N-di(4-methoxyphenyl)amino]-9,9'-spirobifluorene or in short Spiro-OMeTAD and the favorite electrode for this structure is gold. The two architecture types are shown in Figure 5.2.

Transport Layers for Perovskite Solar Cells

77

Figure 5.2. (left)NIP configured Perovskite Solar with typically used Spiro-OMeTAD/gold anode contacts. (right) PIN configuration of Perovskite solar cell having different HTL, ETL and metallic electrodes.

The PIN structure shown in Figure 5.2 does not need a gold electrode as in the case of the NIP structure, which makes it economically favorable. In this architecture, the hole-transport layer could vary from poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS)[14,15], which is a well-known polymeric hole-transport material used in OSCs, to metal-oxide layers[16], as also have been demonstrated for use in OSC in chapter 3 and 4 of this thesis, which also can be used as hole-transporting material. The absorber layer is the same in both methods, namely the perovskite crystal. The electron transport layer also has a wide variation going from polymers to small molecule organic molecules[17,18] plus n-type metal-oxides, such as ZnO[19,20], which can also be used as electron transport layer. The top electrode should be chosen according to the device stability and fabrication method, with consideration of cost, which will be elaborated further on in this chapter.

5-2 Hole Transport Layers in Perovskite Solar Cells

5-2-1 DBP

In chapter 3, the use of DBP as an efficient p-type organic electron donor molecule for use in efficient OSC has been demonstrated. The good optical and electrical properties along with its relatively deep lying HOMO level potentially makes it useful in other solar cells as well, such as in PSCs. PSCs are quite sensitive in terms of energy alignment as they are fabricated from one common absorber layer, the perovskite crystal, which is sandwiched between two transport layers, where the energy levels of the ETL and HTL is going to generate the built-in potential forcing the carriers towards the correct electrodes. Misalignment in energy level can result in carrier losses resulting in extremely low performing devices.

Transport Layers for Perovskite Solar Cells

78

10nm DBP+10nm MoO 3

20nm DBP+10nm MoO 3

30nm DBP+10nm MoO 3

20nm DBP+20nm MoO 3

30nm DBP+20nm MoO 3

40nm DBP+20nm MoO 3

50nm DBP+20nm MoO 3

35nm DBP+20nm MoO 3

0

2

4

6

8

10

Pow

er C

onve

rsio

n Ef

ficie

ncy

(%)

Figure 5.3 shows the result of using DBP/MoO3 hole transport layer in mixed-perovskite.

From literature it is known that the HOMO level of DBP is lying around 5.5 e.V [21]. As already mentioned, band-alignment tenability is one of the key parameter of perovskite absorbers, and this tenability come from adding extra cations, changing the halide, replacing the lead or by simply adding a salt to the solution. For this application, an additional salt was applied to the mixed perovskite to match the HOMO level of DBP, so that hole transport occurs with minimum losses[22].

Hole transporting occur through the HOMO of the DBP level and the Molybdenum oxide assist in generating a high work function contact, that will increase the build-in field. Of course, recombination effects at the hole contact from excitons or electrons are not prevented with the LUMO level offered by DBP, and as the Molybdenum oxide is a pure n-type metal-oxide. If these losses should be prevented, an organic layer with a higher lying LUMO, or a wide bandgap p-type metal-oxide should be employed.

5-2-2 MoOx

In this section, results focusing on the use of MoOx in PSCs are presented. MoOx have shown to be problematic in perovskites as reported by Khan et al.. The problems they report on are mainly divided into two categories. Firstly, the interface between the perovskite and MoO3 (thermally evaporated) results in reduction of the molybdenum oxide forming a MoO2 layer close to the absorber layer that will dramatically increase the serial resistance inside the cell, which reduces the Fill-Factor and widen the hysteresis gap. Secondly they encounter a problem with band-bending of the perovskite absorber, which will reduce the carrier lifetime inside the junction plus energy misalignment due to band-bending and MoO2 formation.

I this work, problems in using the different compositions of MoOx, as reported on in chapter 4, as hole transport layer also leads to challenges, especially when employing the as-deposited layers. For the annealed films, the roughness of the film is generating challenges in the integration, as the annealing leads to large roughness increase. If the roughness becomes too large, i.e. larger than the average thickness of the Perovskite absorber layer, a connection between the bottom HTL and top ETL layer will result in mostly short or very large hysteresis in the forward and reverse sweep in the JV characteristics of the PSC cells, as

Transport Layers for Perovskite Solar Cells

79

shown in [23]. Taking into consideration that the PSC absorber is typically in the order of few hundred nanometers, whereas for OSCs it is typically few tens of nanometers, the choice was to use thicker MoOx films for the PSC investigations, which thus in this case chosen to be 30, 50 and 70nm for the MoOx layer.

Figure 5.4 shows all results on using the MoOx layers as HTL in MAPI Perovskite solar cells. Reference cells employing Nickel Oxide doped with Copper ions (NiO:Cu) as HTL were continuously employed as to compare up against the MoOx based cells. Different thicknesses, temperatures and structures of the MoOx layers were investigated to test if the reactive sputtered MoOx could be used as HTL in PSCs. Clearly, as seen in Figure 5.4, the use of reactive sputtered MoOx in perovskite cells is leading to large drops in performance compared to the reference cells. This is the case for the different thicknesses investigated, and for the different annealing temperatures. The problems could be similar as those reported on by Kahn et. al., i.e. that an unfavorable interface layer between the MoOx and perovskite layer is formed, however, further analysis on especially surface characterization would be needed to conclude on that.

In general, MoOx has a very deep lying fermi level which can dramatically increase the built-in potential inside solar cells, resulting in very efficient charge collection at the interface. With this in mind, we have tried a double HTL configuration using MoOx as a potential generator, and NiO:Cu as an efficient electron blocking layer, yielding in some case to promising results.(see Figure 5.4).

The results in Figure 5.4 shows promising results for multilayer hole transport consisting of super-oxided MoOx annealed at 500oC layer followed by NiO:Cu. Although the PCE is still low due to leakage inside the cell but the current the overall performance of the best performing devices are almost comparable to the reference devices.

Transport Layers for Perovskite Solar Cells

80

NiO(30

00)_r

ef

30nm

MoO

3@

200o C

30nm

MoO

3@

350o C

30nm

MoO

3@

500o C

30nm

MoO

3.16@

200o C

30nm

MoO

3.16@

350o C

30nm

MoO

3.16@

500o C

NiO(30

00)_r

ef

NiO(30

00)_r

ef

30nm

MoO

3@

200o C -N

iO(30

00)

30nm

MoO

3@

350o C -N

iO30

00)

30nm

MoO

3@

500o C -N

iO(30

00)

30nm

MoO

3.16@

200o C -N

iO(30

00)

30nm

MoO

3.16@

350o C -N

iO(30

00)

30nm

MoO

3.16@

500o C -N

iO(30

00)

NiO(30

00)_r

ef

NiO(30

00)_r

ef

50nm

MoO

3.16 @

200o C -N

iO(30

00)

50nm

MoO

3.16@

350o C -N

iO(30

00)

50nm

MoO

3.16@

500o C -N

iO(30

00)

70nm

MoO

3.16@

200o C -N

iO(30

00)

70nm

MoO

3.16@

350o C -N

iO(30

00)

70nm

MoO

3.16@

500o C -N

iO(30

00)

-202468

1012141618

Pow

er C

onve

rsio

n E

ffici

ency

(%)

Ave. PCE Max. PCE

Figure 5.4 Performance of using MoOx as hole transport with and without NiO:Cu as HTL in MAPI.

Transport Layers for Perovskite Solar Cells

81

In addition to interface contact problems, the thicknesses seem to need additional optimization in order to be used in PSCs. In general, however, although there seem to form a non-favorable interface in these films, the deep lying fermi level (high work function) of the MoOx layer could still assist in increasing the current density to the level which is comparable to the reference devices (Figure 5.5). With more optimization of the layers, and probably reducing surface roughness and therefore coverage from the NiO:Cu layer on the Molybdenum oxide layer, a further increase in current density and thus device performance could be expected.

NiO(3000)-PL-PCBM-ZnO-BCP(5000)-A

l

MoO 3.16@200

o C-NiO 2-PL-PCBM-ZnO-BCP(5000)-A

l

MoO 3.16@350

o C-NiO 2-PL-PCBM-ZnO-BCP(5000)-A

l

MoO 3.16@500

o C-NiO 2-PL-PCBM-ZnO-BCP(5000)-A

l

MoO 3.16@200

o C-NiO 2-PL-PCBM-ZnO-BCP(5000)-A

l

MoO 3.16@350

o C-NiO 2-PL-PCBM-ZnO-BCP(5000)-A

l

MoO 3.16@500

o C-NiO 2-PL-PCBM-ZnO-BCP(5000)-A

l0

2

4

6

8

10

12

14

16

18

20C

urre

nt D

ensi

ty (m

A/c

m2 )

Ave. Jsc

Max. Jsc

Figure 5.5 The current density of a cell with annealed super oxide Molybdenum Oxide used as hole-extraction layer combined with NiO:Cu as hole-transport layer in MAPI.

5-3 Development of efficient PIN perovskite solar cells

As already demonstrated, several factors can be used to tune the performance in perovskite based solar cells. In terms of the active layer, the available methods encompass solvent and anti-solvent modifications, which are investigated in this work. In addition to this, contact layer plays a crucial role also in PIN perovskite solar cells, which is also investigated. Due to challenges using MoOx, as presented in the previous section, alternative contact layers are here investigated, both as HTL and ETL layers.

Active layer optimization in PIN Perovskite solar cells

In order to optimize the active layer of the investigated PIN perovskite solar cells, investigation on both solvent engineering, and anti-solvent process times were conducted, To investigate the impact on solvent engineering for perovskite solar cells, a well-known triple cation Perovskite (Cs0.05(MA0.17FA0.83)0.95Pb(I2.7Br03)) is used as the absorber layer. In this work, a combination of mainly four solvents is used in order to form smooth films of the perovskite layer. The solvents used are DMF, having a

Transport Layers for Perovskite Solar Cells

82

boiling point of 153oC this is the main solvent for any type of PSCs. Dimethyl sulfoxide (DMSO), having a boiling point of 180oC, and Gamma-Butyrolactone (GBL), having a boiling point of 204oC, was here added to the DMF, in order to increase the boiling point of the whole solution. In addition to these three main solvents, also NMF was added, which is a very viscous solvent compared to the other three that was employed in order to assist in the formation of a low surface roughness film. Figure 5.6 shows the results of solvent engineering employed in this work, where the solvent combination (DMF:GBL:DMSO:NMF) has been changed in order to fabricate the perovskite solar cells. The power conversion efficiencies are shown. Toluene was used as an anti-solvent for all the samples, and the number at the end of each material combination shows the time is seconds when the toluene anti-solvent flushed on the samples. HP is short for the samples that were annealed at 100oC prior to spin-coating.

9:0:1:0_5000@30

1:7:1:1@30

1:8:1:0@30

7:2:1:0_60@15HP

8:1:1:0_45@25HP

7:1:1:1_60@25HP

6.5:1:1:1.5_60@20HP11.0

11.5

12.0

12.5

13.0

13.5

14.0

14.5

15.0

15.5

PC

E (%

)

Ave. PCE Max. PCE

Figure 5.6. Solvent engineering on triple cation Perovskite solar cell using DMF:GBL:DMSO:NMF as solvent and Toluene as Anti-Solvent

The solvent combination of (GBL:DMSO:DMF:NMF 7:2:1:0) shows the best forming cells with over 15% efficiency for the best performing device. Hot-casting before starting the spin-coating also has an impact on the performance but it has negative influence the film properties because the solvent dry-out faster; to prevent fast drying film, anti-solvent has to be dropped sooner to help forming the perovskite crystal.

It worth mentioning that the best performing triple cation perovskite has been made using the NIP configuration with Spiro/Au, where more than 21% efficiency[24] can be obtained. The purpose here is to develop efficient PIN configuration cells, which potentially are much more cost-effective, and for the results shown in Figure 5.6, Aluminum has been applied as the metal electrode thus removing unwanted Spiro hole transport layers and gold electrodes.

Transport Layers for Perovskite Solar Cells

83

It is well-known that anti-solvents have a major impact on film roughness and formation of Perovskite crystals, which is strongly affected by the timing of the process or by changing the solvent itself [25–28], which of course also possess potential challenges for large scale manufacturing of the technology in the future. Figure 5.7 shows the sensitivity of the dual cation perovskite solar cells (Cs0.15FA0.85Pb(I2.8Br0.2)) to the timing in the anti-solvent processing. Similar combinations of solvents for solution preparation were made and toluene used as the anti-solvent at different times. The acceleration speed in the spin-coating process was changed in order to reach a better quality film, meanwhile the anti-solvent timing was the main factor investigated. The total time of the spin-coating process including the acceleration, coating and deceleration times were kept constant for all the samples. The best results are obtained when using a longer process delay time for the anti-solvent using DMF:DMSO as the solvent, however, a lower yield are obtained here lowering the average performance.

Figure 5.7. Power conversion effiency of perovskite solar cells as a function of the anti-solvent timing process. Small change in anti-solvent flushing on the sample can alter the performance of the final device regardless of the main

solvent or spin-coating parameters.

5-4 HTL and ETL optimization in PIN Perovskite solar cells

In order to investigate the effect of modification in the deposition process of the HTL and ETL layers, studies on the change in the spin rotation speed of the NiO:Cu (HTL) and the multi-layer ETL including ZnO, BCP and PCBM is conducted. These tests are all made on the MAPI Perovskite in PIN configuration, as previously mentioned. In this configuration, there is a variety of HTL that can be used including metal-oxides such as NiO and SnO, in addition to aforementioned MoOx. Since the intrinsic electrical properties are poor (for many stoichiometric oxides), they are typically doped with metals in order to improve their transporting properties, and also to fine-tune their energy level towards improved extraction properties[29] inside the cell. For NiO and SnO, doping with Cu and Sb can be used to obtain NiO:Cu[30,31] or SnO:Sb[32].

A:4000,27"(GBL:DMF:DMSO)

A:4000,32"(GBL:DMF:DMSO)

A:2000,16"(GBL:DMF:DMSO)

A:2000,21"(GBL:DMF:DMSO)

A:4000,27"(DMF:DMSO)

A:4000,32"(DMF:DMSO)

A:2000,16"(DMF:DMSO)

A:2000,21"(DMF:DMSO)

6

7

8

9

10

11

12

13

PC

E (%

)

Ave. PCE Max. PCE

Transport Layers for Perovskite Solar Cells

84

As for ETL layer, ZnO typically show high device efficiencies in combination with MAPI perovskite active layers due to the favorable energy level alignment. Doping ZnO with Aluminum (ZnO:Al) improves this ETL layer further, mainly due to the improved electron transport properties. Although ZnO is an efficient ETL, it is known to reduce the stability of a PSC. In order to increase the stability of the ETL, we have tested deposition of BCP using solution processing, by dissolving BCP in IPA to have a capping layer on top of ZnO, and from that avoid penetration of the metallic electrode into the ZnO layer, and further into the absorber with time. Finally, in oder to obtain ideal energy level alignment and a proper contact to the metallic electrode, we introduced a PCBM layer (dissolved in Chlorobenzene CB), which provides higher device yield and higher current densities. Figure 5.8 shows the results from the processing parameters with respect to fine tuning of the NiO:Cu and BCP HTL and ETL in the MAPbI3 perovskite solar cells, here with no further ETL layer (no PCBM).

NiO:Cu(1000)-ZnO-no BCP

NiO:Cu(1000)-BCP(1000)

NiO:Cu(1000)-BCP(3000)

NiO:Cu(3000)-BCP(3000)

NiO:Cu(3000)-BCP(4000)

NiO:Cu(3000)-BCP(5000)

NiO:Cu(4000)-BCP(3000)

NiO:Cu(4000)-BCP(4000)

NiO:Cu(4000)-BCP(5000)

NiO:Cu(5000)-BCP(3000)

NiO:Cu(5000)-BCP(4000)

NiO:Cu(5000)-BCP(5000)

NiO:Cu(3000)-BCP:ZnO(5000)

0.0

2.5

5.0

7.5

10.0

12.5

15.0

17.5

PC

E (%

)

Ave. PCE Max. PCE

ZnO(3000)/BCP(5000)_Ref

ZnO:BCP(1000)

ZnO:BCP(3000)

ZnO:BCP(5000)

ZnO:BCP (70%)_5000

ZnO:BCP (80%)_5000

ZnO:BCP (90%)_5000

ZnO:BCP (100%)_50000

2

4

6

8

10

12

14

16

PC

E (%

)

Ave. PCE Max. PCE

Figure 5.8. Optimization of ETL and HTL layers in PIN MAPI perovskite solar cells with Cu electrodes. ZnO+BCP solutions with different concentrations have been used as ETL.

Transport Layers for Perovskite Solar Cells

85

5-5 Metal electrode optimization in PIN Perovskite solar cells

One of the drawbacks of the NIP structure is the standard used Spiro as HTL and gold as the metallic electrode. It has been shown by many groups that Spiro is not that stable under the working condition required for PSC (80oC, constant illumination)[33–35], mainly due to the lack of thermal stability for Spiro, while it allows gold migration from the metallic electrode to the absorber where the gold reduce the performance of the perovskite cell severely[36]. Another obstacle is the need for a high work-function metallic electrode, in order to have a proper built-in potential inside the junction, which so far only have been provided successfully by the use of gold for maximum performance. Based on this, we here investigate viable methods for replacing gold with a cheaper material in PIN electrodes, plus avoid the use of Spiro due to the aforementioned reasons. Two candidates for the PIN structure were tested:

• Aluminum: Cheap metal with a work function of 4.08 eV, and can be used as a good reflector at the back electrode.

• Copper: is also a cheap metal possessing a work function of 5.0 e.V, which is very close to gold.

Both metals were tested in the same configuration cells. Initial results showed that connecting a Copper electrode with ZnO ETL results in short circuit device. This could be due to ion migration at the interface of ZnO and Copper electrode forming a ZnO/CuO junction at the interface, and later ion exchange between the Perovskite active layer and electrode. Following this, a BCP layer was introduced in between ZnO and the Copper electrode (sacrificial ETL), in order to avoid shortening the cells. This approach yielded promising results, as already shown in Figure 5.8. In contrast to Copper, Aluminum tends to work both with and without the BCP capping layer, delivering almost the same device performance. Figure 5.9 demonstrates the comparison between PS performances using Copper and Aluminum electrodes in the MAPI configuration (top). Fig. 5.9 (top) shows the PCE of the best performing cell under maximum power-point voltage, biased for 5 minutes providing a glance towards the more stable Copper electrode device.

Transport Layers for Perovskite Solar Cells

86

NiO:Cu-PL-PCBM-ZnO-Cu

NiO:Cu-PL-PCBM-ZnO-BCP-Cu

NiO:Cu-PL-PCBM-BCP:ZnO-Cu

NiO:Cu-PL-PCBM-ZnO-Al

NiO:Cu-PL-PCBM-ZnO-BCP-Al0

2

4

6

8

10

12

14

16

18

PC

E (%

)

Ave. PCE Max. PCE

0 50 100 150 200 250 300

98

100

102

104

106

108

110

112

114

ITO/NiO:Cu/PL/ZnO/BCP/PCBM/Electrode

Nor

mal

ized

PC

E (%

)

Time (S)

With Aluminum With Copper

Figure 5.9 (Top) PSC PCE using copper and Aluminum with and without BCP as the electron contact layer. The results clarify the crucial role of BCP when Cu uses as electrode. (Bottom) Biased PSC device measurements over time using

BCP with different electrodes, which shows a better performing Copper electrode.

Transport Layers for Perovskite Solar Cells

87

In this chapter, the use of hole contacts developed for organic solar cells were tested out for use in different configuration perovskite solar cells, where they due to their promising hole transporting mechanisms could yield good contacts. The use of combined DBP and Molybdenum oxide have been demonstrated to be challenging, as quite reduced device efficiencies are obtained compared the typically used Spiro/Au references. Clearly optimizing the thickness of the contact layers can result in some improvement, but further improvements would be needed for this to act as a replacement of Spiro/Au in this technology. One route could be to look at organic transport layers having deeper lying HOMO levels.

Other methods for replacing Spiro/Au hole contacts is looking at PIN structures with implemented reactive sputtered Molybdenum oxide, from which it is clear that a capping of the MoOx layer is needed, since contact with the perovskite layer results in an unfavorable interface, as also reported on for thermally evaporated Molybdenum oxide. However, when capped with a proper perovskite metal-oxide HTL layer such as NiO, promising results from MoOx have been obtained. In this direction, different HTL and ETL configurations in PSC PIN configurations have been investigated further, and good results using NiO:Cu at HTL and PCBM/ZnO/BCP as ETL with Copper at the metallic electrode have been obtained.

Transport Layers for Perovskite Solar Cells

88

5-6 Reference

[1] B. O’Regan, M. Grätzel, A low-cost, high-efficiency solar cell based on dye-sensitized colloidal TiO2 films, Nature. 353 (1991) 737–740. doi:10.1038/353737a0.

[2] A. Kojima, K. Teshima, Y. Shirai, T. Miyasaka, Organometal Halide Perovskites as Visible-Light Sensitizers for Photovoltaic Cells, J. Am. Chem. Soc. 131 (2009) 6050–6051. doi:10.1021/ja809598r.

[3] M.M. Lee, J. Teuscher, T. Miyasaka, T.N. Murakami, H.J. Snaith, Efficient Hybrid Solar Cells Based on Meso-Superstructured Organometal Halide Perovskites, Science (80-. ). 338 (2012) 643–647. doi:10.1126/science.1228604.

[4] M. Liu, M.B. Johnston, H.J. Snaith, Efficient planar heterojunction perovskite solar cells by vapour deposition., Nature. 501 (2013) 395–8. doi:10.1038/nature12509.

[5] P. Docampo, J.M. Ball, M. Darwich, G.E. Eperon, H.J. Snaith, Efficient organometal trihalide perovskite planar-heterojunction solar cells on flexible polymer substrates, Nat. Commun. 4 (2013). doi:10.1038/ncomms3761.

[6] O. Malinkiewicz, A. Yella, Y.H. Lee, G.M. Espallargas, M. Graetzel, M.K. Nazeeruddin, H.J. Bolink, Perovskite solar cells employing organic charge-transport layers, Nat. Photonics. 8 (2013) 128–132. doi:10.1038/nphoton.2013.341.

[7] J.-H. Im, I.-H. Jang, N. Pellet, M. Grätzel, N.-G. Park, Growth of CH3NH3PbI3 cuboids with controlled size for high-efficiency perovskite solar cells, Nat. Nanotechnol. 9 (2014) 927–932. doi:10.1038/nnano.2014.181.

[8] J. Burschka, N. Pellet, S.-J. Moon, R. Humphry-Baker, P. Gao, M.K. Nazeeruddin, M. Grätzel, Sequential deposition as a route to high-performance perovskite-sensitized solar cells, Nature. 499 (2013) 316–319. doi:10.1038/nature12340.

[9] G.E. Eperon, V.M. Burlakov, P. Docampo, A. Goriely, H.J. Snaith, Morphological Control for High Performance, Solution-Processed Planar Heterojunction Perovskite Solar Cells, Adv. Funct. Mater. 24 (2014) 151–157. doi:10.1002/adfm.201302090.

[10] Z. Xiao, C. Bi, Y. Shao, Q. Dong, Q. Wang, Y. Yuan, C. Wang, Y. Gao, J. Huang, Efficient, high yield perovskite photovoltaic devices grown by interdiffusion of solution-processed precursor stacking layers, Energy Environ. Sci. 7 (2014) 2619–2623. doi:10.1039/C4EE01138D.

[11] Y. Zhao, K. Zhu, CH 3 NH 3 Cl-Assisted One-Step Solution Growth of CH 3 NH 3 PbI 3 : Structure, Charge-Carrier Dynamics, and Photovoltaic Properties of Perovskite Solar Cells, J. Phys. Chem. C. 118 (2014) 9412–9418. doi:10.1021/jp502696w.

[12] A.M.A. Leguy, P. Azarhoosh, M.I. Alonso, M. Campoy-Quiles, O.J. Weber, J. Yao, D. Bryant, M.T. Weller, J. Nelson, A. Walsh, M. van Schilfgaarde, P.R.F. Barnes, Experimental and theoretical optical properties of methylammonium lead halide perovskites, Nanoscale. 8 (2016) 6317–6327. doi:10.1039/C5NR05435D.

[13] P. Umari, E. Mosconi, F. De Angelis, Relativistic GW calculations on CH3NH3PbI3 and CH3NH3SnI3 Perovskites for Solar Cell Applications, Sci. Rep. 4 (2015) 4467. doi:10.1038/srep04467.

[14] D. Huang, T. Goh, J. Kong, Y. Zheng, S. Zhao, Z. Xu, A.D. Taylor, Perovskite solar cells with a DMSO-treated PEDOT:PSS hole transport layer exhibit higher photovoltaic performance and enhanced durability, Nanoscale. 9 (2017) 4236–4243. doi:10.1039/C6NR08375G.

[15] Q. Wang, C.-C. Chueh, M. Eslamian, A.K.-Y. Jen, Modulation of PEDOT:PSS pH for Efficient Inverted Perovskite Solar Cells with Reduced Potential Loss and Enhanced Stability, ACS Appl. Mater. Interfaces. 8 (2016) 32068–32076. doi:10.1021/acsami.6b11757.

[16] M.A. Haque, A.D. Sheikh, X. Guan, T. Wu, Metal Oxides as Efficient Charge Transporters in Perovskite Solar Cells, Adv. Energy Mater. (2017) 1602803. doi:10.1002/aenm.201602803.

[17] P.-Y. Gu, N. Wang, C. Wang, Y. Zhou, G. Long, M. Tian, W. Chen, X.W. Sun, M.G. Kanatzidis, Q. Zhang, Pushing up the efficiency of planar perovskite solar cells to 18.2% with organic small molecules as the electron transport layer, J. Mater. Chem. A. 5 (2017) 7339–7344. doi:10.1039/C7TA01764B.

Transport Layers for Perovskite Solar Cells

89

[18] J. Lee, J. Kim, C.-L. Lee, G. Kim, T.K. Kim, H. Back, S. Jung, K. Yu, S. Hong, S. Lee, S. Kim, S. Jeong, H. Kang, K. Lee, A Printable Organic Electron Transport Layer for Low-Temperature-Processed, Hysteresis-Free, and Stable Planar Perovskite Solar Cells, Adv. Energy Mater. 7 (2017) 1700226. doi:10.1002/aenm.201700226.

[19] L.-C. Chen, Z.-L. Tseng, ZnO-Based Electron Transporting Layer for Perovskite Solar Cells, in: Nanostructured Sol. Cells, InTech, 2017. doi:10.5772/65056.

[20] W. Hadouchi, J. Rousset, D. Tondelier, B. Geffroy, Y. Bonnassieux, Zinc oxide as a hole blocking layer for perovskite solar cells deposited in atmospheric conditions, RSC Adv. 6 (2016) 67715–67723. doi:10.1039/C6RA16865E.

[21] J.C. Bernède, L. Cattin, S.O. Djobo, M. Morsli, S.R.B. Kanth, S. Patil, P. Leriche, J. Roncali, A. Godoy, F.R. Diaz, M.A. del Valle, Influence of the highest occupied molecular orbital energy level of the donor material on the effectiveness of the anode buffer layer in organic solar cells, Phys. Status Solidi. 208 (2011) 1989–1994. doi:10.1002/pssa.201127047.

[22] T. Jesper Jacobsson, J.-P. Correa-Baena, M. Pazoki, M. Saliba, K. Schenk, M. Grätzel, A. Hagfeldt, Exploration of the compositional space for mixed lead halogen perovskites for high efficiency solar cells, Energy Environ. Sci. 9 (2016) 1706–1724. doi:10.1039/C6EE00030D.

[23] P. Schulz, J.O. Tiepelt, J.A. Christians, I. Levine, E. Edri, E.M. Sanehira, G. Hodes, D. Cahen, A. Kahn, High-Work-Function Molybdenum Oxide Hole Extraction Contacts in Hybrid Organic–Inorganic Perovskite Solar Cells, ACS Appl. Mater. Interfaces. 8 (2016) 31491–31499. doi:10.1021/acsami.6b10898.

[24] M. Saliba, T. Matsui, J.-Y. Seo, K. Domanski, J.-P. Correa-Baena, M.K. Nazeeruddin, S.M. Zakeeruddin, W. Tress, A. Abate, A. Hagfeldt, M. Grätzel, Cesium-containing triple cation perovskite solar cells: improved stability, reproducibility and high efficiency, Energy Environ. Sci. 9 (2016) 1989–1997. doi:10.1039/C5EE03874J.

[25] Y. Yu, S. Yang, L. Lei, Q. Cao, J. Shao, S. Zhang, Y. Liu, Ultrasmooth Perovskite Film via Mixed Anti-Solvent Strategy with Improved Efficiency, ACS Appl. Mater. Interfaces. 9 (2017) 3667–3676. doi:10.1021/acsami.6b14270.

[26] S. Paek, P. Schouwink, E.N. Athanasopoulou, K.T. Cho, G. Grancini, Y. Lee, Y. Zhang, F. Stellacci, M.K. Nazeeruddin, P. Gao, From Nano- to Micrometer Scale: The Role of Antisolvent Treatment on High Performance Perovskite Solar Cells, Chem. Mater. 29 (2017) 3490–3498. doi:10.1021/acs.chemmater.6b05353.

[27] M. Li, X. Yan, Z. Kang, X. Liao, Y. Li, X. Zheng, P. Lin, J. Meng, Y. Zhang, Enhanced Efficiency and Stability of Perovskite Solar Cells via Anti-Solvent Treatment in Two-Step Deposition Method, ACS Appl. Mater. Interfaces. 9 (2017) 7224–7231. doi:10.1021/acsami.7b01136.

[28] Y. Li, J. Wang, Y. Yuan, X. Dong, P. Wang, Anti-solvent dependent device performance in CH 3 NH 3 PbI 3 solar cells: the role of intermediate phase content in the as-prepared thin films, Sustain. Energy Fuels. 1 (2017) 1041–1048. doi:10.1039/C7SE00125H.

[29] S. Shanthi, C. Subramanian, P. Ramasamy, Growth and characterization of antimony doped tin oxide thin films, J. Cryst. Growth. 197 (1999) 858–864. doi:10.1016/S0022-0248(98)01066-5.

[30] J.W. Jung, C.-C. Chueh, A.K.-Y. Jen, A Low-Temperature, Solution-Processable, Cu-Doped Nickel Oxide Hole-Transporting Layer via the Combustion Method for High-Performance Thin-Film Perovskite Solar Cells, Adv. Mater. 27 (2015) 7874–7880. doi:10.1002/adma.201503298.

[31] K. Yao, F. Li, Q. He, X. Wang, Y. Jiang, H. Huang, A.K.-Y. Jen, A copper-doped nickel oxide bilayer for enhancing efficiency and stability of hysteresis-free inverted mesoporous perovskite solar cells, Nano Energy. 40 (2017) 155–162. doi:10.1016/j.nanoen.2017.08.014.

[32] Y. Sun, W.D. Chemelewski, S.P. Berglund, C. Li, H. He, G. Shi, C.B. Mullins, Antimony-Doped Tin Oxide Nanorods as a Transparent Conducting Electrode for Enhancing Photoelectrochemical Oxidation of Water by Hematite, ACS Appl. Mater. Interfaces. 6 (2014) 5494–5499. doi:10.1021/am405628r.

[33] S. Kim, S. Bae, S.-W. Lee, K. Cho, K.D. Lee, H. Kim, S. Park, G. Kwon, S.-W. Ahn, H.-M. Lee, Y. Kang, H.-S. Lee, D. Kim, Relationship between ion migration and interfacial degradation of CH3NH3PbI3

Transport Layers for Perovskite Solar Cells

90

perovskite solar cells under thermal conditions, Sci. Rep. 7 (2017) 1200. doi:10.1038/s41598-017-00866-6.

[34] E.M. Sanehira, B.J. Tremolet de Villers, P. Schulz, M.O. Reese, S. Ferrere, K. Zhu, L.Y. Lin, J.J. Berry, J.M. Luther, Influence of Electrode Interfaces on the Stability of Perovskite Solar Cells: Reduced Degradation Using MoO x /Al for Hole Collection, ACS Energy Lett. 1 (2016) 38–45. doi:10.1021/acsenergylett.6b00013.

[35] R.T. Ginting, M.-K. Jeon, K.-J. Lee, W.-Y. Jin, T.-W. Kim, J.-W. Kang, Degradation mechanism of planar-perovskite solar cells: correlating evolution of iodine distribution and photocurrent hysteresis, J. Mater. Chem. A. 5 (2017) 4527–4534. doi:10.1039/C6TA09202K.

[36] K. Domanski, J.-P. Correa-Baena, N. Mine, M.K. Nazeeruddin, A. Abate, M. Saliba, W. Tress, A. Hagfeldt, M. Grätzel, Not All That Glitters Is Gold: Metal-Migration-Induced Degradation in Perovskite Solar Cells, ACS Nano. 10 (2016) 6306–6314. doi:10.1021/acsnano.6b02613.

Chapter 6 Conclusion and Outlook

Conclusions and Outlook

93

6-1 Conclusions

Thin-film photovoltaics have attracted a lot of attention during the last few decades as OSC power conversion efficiency reaches more than 13% and for PSC, the power conversion efficiency has reached more than 22%. There has been a shift in interest towards Perovskite since 2012, mainly due to its higher efficiency reaching values close to Silicon PV, but offers properties close to those of OPV.

In this thesis both device types have been developed. For OSC development, vacuum evaporation processing we employed, and for perovskites, wet processing was conducted. The focus in both cases, however, has mainly been on the transport layers and specifically on the utilization of the tunable properties of DC-sputtered Molybdenum oxide in photovoltaic applications.

Firstly, results from using Tetraphenyldibenzoperiflanthene (DBP) and Fullerene (C70) in single and multilayer bilayer configuration stacks were investigated. Simulations have been applied to optimize a five-stack cell, and the efficiency of the final stack was increased by 50% due to current matching among each individual sub-cell.

Realizing the tenability of DC-sputtered molybdenum oxide, the layers have been implemented inside thin-film solar cells to be used as hole-transport layer. Depending on the layer thickness and composition the performance of the final devices vary a lot, which is due to change in energy alignment inside the device along with the change in the opto-electrical properties of the films. Annealing the films at certain temperatures affects the physical property of the films, introducing some levels of crystallinity in the films. This crystal formation comes along with an increase in work function values, making is suitable as hole transport layer even for organic solar cells with deep lying HOMO levels. It is shown that the OSC performance from 30nm annealed super-oxided DC-sputtered Molybdenum oxide films match that of 30nm as-deposited thermally evaporated Molybdenum oxide, typically used in organic solar cells, however, with strongly improved device stability. This large increase in device stability could hold very interesting technological and even commercial potential that should be investigated further in future work.

In chapter 5 different hole-transport layers for perovskite solar cells have been investigated as a replacement for SpiroOmeTAD for standard configuration Perovskite cells. For DBP, vacuum evaporation and spin-coating have been, and although some promising trends have been seen, DBP did not reveal to be a good candidate for hole transport in perovskite solar cells. As the evaporated metal electrode deposited directly on DBP can generate pin-holes inside the soft organic layer, evaporated molybdenum oxide layers was placed on top of DBP, both to form a stronger build-in field, and to protect the organic layer from the metal evaporation. In inverted perovskite solar cells, the use of DC-sputtered Molybdenum oxide as HTL, similar to the case of OSCs, were also investigated, and it has been shown that annealed super-oxide layers may be used as possible HTL, mostly for their positive effects on the built-in potential of the final device.

HTL and ETL have been optimized for inverted configuration PSCs to avoid the Spiro/Gold combination, firstly to reduce cost, but also to increase the thermal stability of the final cell. It is shown that a NiO:Cu at speed rotation of 3000rpm gives the best results. To simplify the fabrication process that can lead to production of PSC, we investigate a multi-stack ETL including PCBM, ZnO and BCP. Different configurations have been tested in order to reduce the complexity of the final ETL, where it was realized that Copper electrodes can be applied with single layer PCBM as the ETL.

Conclusions and Outlook

94

6-2 Outlook

It has been demonstrated that Molybdenum oxide can be used to increase the stability of OSC devices without affecting their performance. The results here can be a path for optimizing other metal oxides especially transition metals that has similar properties. This may also be interesting for other PV technologies such as in Silicon PV, where Molybdenum oxide hole contact layer have been well investigated in recent years due to promising results in this field.

With doping, it is possible to modify the electrical, morphological and optical properties of both the organic layers and metal oxide layers. Using different methods such as wet (solution processing) and dry (sputtered/co-sputter) processes, introduction of different annealing temperatures in different ambient conditions can be future approaches for further optimizing both the performance and stability of the developed solar cells.

The organic device also can be optimize further making use of the Bulk-heterojunction properties by co-evaporating DBP and C70 to reach higher efficiencies. For the multi-stack structure, donors and acceptors with different absorption spectrums can be employed for this, to basically fabricate tandem cells that results in even further increase the efficiencies, due to the larger absorption range.