• unlimited access with print and download
    $ 37 00
  • read full document, no print or download, expires after 72 hours
    $ 4 99
More info
Unlimited access including download and printing, plus availability for reading and annotating in your in your Udini library.
  • Access to this article in your Udini library for 72 hours from purchase.
  • The article will not be available for download or print.
  • Upgrade to the full version of this document at a reduced price.
  • Your trial access payment is credited when purchasing the full version.
Buy
Continue searching

Thin film solar cells: Thin film deposition, solar cell fabrication and characterization

ProQuest Dissertations and Theses, 2011
Dissertation
Author: Qiaoer Zhou
Abstract:
Thin film solar cells have the potential to lower the solar cell cost. The global market share of the thin film solar cells have been increased from 5% to 17.7% in just 6 years (2003-2009) compared to the crystal silicon solar cells. In this work, a Chemical Vapor Deposition (CVD) system was converted into an Atomic Layer Deposition system by hardware modification, including the installation of metal-organic precursor bubbler and in-line H2 S generator, and computer automation realization by Labwindows programming. ALD Cux S film, a good absorber layer, was deposited on both corning glass and nanostructured TiO2 layer. SEM image of ALD CuS coated 100nm nanostructured TiO2 showed perfect penetration of ALD film, which resulted from self-saturated surface reaction of ALD. Another absorber layer, SnS was also deposited by the same ALD system using dibutyltin diacetate (DBTDA) metal-organic precursor as the tin source. It is the first time DBTDA was used as tin source in ALD SnS deposition. ALD CuS and SnS films are proposed to be used in extremely thin absorber (ETA) solar cells. Chemical Bath Deposition (CBD) of Indium sulfide was also demonstrated in this thesis. It was the first time that CBD indium sulfide was used in P3HT:PCBM bulk heterojunction organic solar cell as the buffer layer between ITO and P3HT:PCBM to replace the problematic PEDOT:PSS. By using different front contact metal (Al and Au), it was found that In 2 S3 can transport both electrons and holes, which makes it a really good buffer layer in the tandem cells serving both hole transporter and electron transporter, where usually two buffer layers are needed. The efficiency of the un-optimized single junction device using In 2 S3 as hole transporter is 1.5%. The last part of the thesis shows thermoreflectance (TR) imaging and Electroluminescence (EL) imaging of poly-crystalline silicon solar cells. These two imaging techniques in principle can be used in thin film solar cell characterization also. The local I-V curve on a diode shunt was obtained using high spatial resolution TR imaging for the first time. From local I-V curve on the defects, we can tell if a shunt is linear or diode behavior. The shunt we looked was a weak diode; by fitting local I-V curve with the diode equation, we got the ideality factor of that particular shunt much larger than 2, which indicated that the defect is probably caused by high concentration dislocation. Thermal diffusivity and thermal conductivity were also obtained from the TR phase image and amplitude image of the hot spot. Both TR image and EL image were used to detect the cracks. Analyzing the reverse biased EL image of the solar cell the cause of the pre-breakdown was discussed.

Table of Contents List of Figures vi List of Tables xi Abstract xii Acknowledgement xiv 1

Literature review ................................................................................................... 1

1.1

Solar Cell Progress ......................................................................................... 1

1.2

Thin Film Solar Cells ..................................................................................... 4

1.3

The Scope of Thesis ....................................................................................... 6

2

ALD deposition of Cu x S and Sn x S and application in solar cells ......................... 8

2.1

Introduction .................................................................................................... 8

2.1.1

Character of Atomic Layer Deposition ................................................... 8

2.1.2

ALD application in solar cells .............................................................. 11

2.2

Precursor Consideration ............................................................................... 13

2.3

Experimental Setup ...................................................................................... 18

2.3.1

Bubblers ................................................................................................ 18

2.3.2

H 2 S Generator ....................................................................................... 19

iv

2.3.3

Reaction Chamber and Vacuum System............................................... 21

2.3.4

Computer Control and Automation ...................................................... 23

2.3.5

Substrate Preparation ............................................................................ 24

2.3.6

Characterization Method ....................................................................... 25

2.4

Results .......................................................................................................... 26

2.4.1

ALD copper sulfide by Cu(thd) 2 ........................................................... 26

2.4.2

ALD and CVD Cu 2 S by KI5 ................................................................. 30

2.4.3

ALD tin sulfide ..................................................................................... 38

2.4.4

3D solar cell structure with ALD SnS x ................................................. 41

2.5

Conclusion .................................................................................................... 44

3

Chemical Bath Deposition of In 2 S 3 and In 2 S 3 /P3HT :PCBM solar cell [105] ... 46

3.1

Introduction .................................................................................................. 46

3.2

Experimental Setup ...................................................................................... 47

3.3

Results .......................................................................................................... 49

3.3.1

In 2 S 3 Film Results ................................................................................. 49

3.3.2

Solar Cell Device Results ..................................................................... 53

3.4

Conclusion .................................................................................................... 59

v

4

Defect characterization of poly-silicon solar cells by spatially resolved imaging 60

4.1

Introduction .................................................................................................. 60

4.2

High spatial resolution characterization of poly-silicon solar cells using thermoreflectance imaging [106] ............................................................................ 63

4.2.1

Introduction ........................................................................................... 63

4.2.2

Experimental Setup ............................................................................... 65

4.2.3

Results ................................................................................................... 69

4.2.4

Conclusions ........................................................................................... 81

4.3

Electroluminiscence imaging from forward and reverse biased Poly-Silicon Solar Cells ............................................................................................................... 82

4.3.1

Introduction ........................................................................................... 82

4.3.2

Experimental Setup ............................................................................... 84

4.3.3

Results ................................................................................................... 85

4.3.4

Conclusion ............................................................................................ 88

5

Conclusions ......................................................................................................... 89

6

References ........................................................................................................... 91

7

Appendix: Labwindows/CVI programming of ALD process ........................... 101

vi

List of Figures Figure 1.1 Best research cell efficiencies from 1975-2009[61]. ................................... 2

Figure 1.2 Global cumulative installed PV capacity by interconnection status[2]. ...... 3

Figure 1.3 LCOE for residential PV systems in several U.S. cities in 2008, with and without the federal investment tax credit [99]. ............................................................. 4

Figure 2.1 Schematic of one ALD cycle. ...................................................................... 9

Figure 2.2 The temperature dependence of the growth rate. ...................................... 10

Figure 2.3 Reported ALD films[102]. ........................................................................ 11

Figure 2.4 Cu(thd) 2 formula. ....................................................................................... 14

Figure 2.5 KI5 formula. .............................................................................................. 15

Figure 2.6 DBTDA formula. ....................................................................................... 16

Figure 2.7 Schematics of bubbler system. .................................................................. 19

Figure 2.8 Schematic of the in-situ H 2 S generator ..................................................... 21

Figure 2.9 Reaction Chamber and chamber vacuum system ...................................... 22

Figure 2.10 Graphic User Interface of ALD system. .................................................. 24

Figure 2.11 AFM of ALD Cu x S films and the effect of annealing ............................. 27

Figure 2.12 EXAFS spectrum on as-deposit ALD CuS film (solid line) and CuS bulk film reference (dashed line). ....................................................................................... 28

Figure 2.13 h vs photon Energy E. ....................................................................... 29

vii

Figure 2.14 a) SEM image of TiO2 film, b) SEM image of TiO2/CuS matrix. ........ 30

Figure 2.15 Growth rate versus substrate temperature using ALD process. .............. 31

Figure 2.16 Radial distribution function simulations for the CuS, Cu 1.7 S and Cu 1.94 S phases (top). Radial Distribution patterns measured with EXAFS for the CuS and Cu 1.7 S (middle). Simulated and measured radial distribution patterns for the Chalcocite Cu 2 S phase. ............................................................................................... 33

Figure 2.17 SEM cross section of a mesoporous TiO2 film before (top) and after Cu 2 S depositon. The Porous film is successfully filled with ALD Cu 2 S.................... 35

Figure 2.18 Optical absorption and thickness for a series of CVD deposited Cu 2 S films showing metallic behavior for films less than 100nm thick and semiconducting Cu 2 S for films greater than 100nm. ............................................................................ 37

Figure 2.19 Sheet resistance and thickness for CVD deposited Cu x S films showing the presence of a metallic layer at the ZnO/ Cu x S interface. ...................................... 38

Figure 2.20 ALD tin sulfide deposition rate versus temperature. ............................... 39

Figure 2.21 AFM of SnxS deposited a) at 340C, RMS=10nm, thickness 70nm, b) at 200C, RMS=1 nm, thickness 30nm with the same scale as a), c) the same image as b) but with optimized scale to see the grains. ................................................................. 40

Figure 2.22 absorption spectrum of SnS films deposited at 200 o C (yellow curve) and 350 o C (blue curve). ..................................................................................................... 41

Figure 2.23 A 3D structured solar cell using TiO2 nanoparticle matrix and ALD SnSx thin film. ...................................................................................................................... 42

viii

Figure 2.24 Preliminary IV results of solar cell device with ALD SnS x film. ........... 43

Figure 2.25 AFM image of P3HT on ALD SnS x film. ............................................... 44

Figure 3.1 Growth curve of In2S3 film on TiO 2 sol-gel and ITO .............................. 50

Figure 3.2 AFM images of In2S3 film on a) ITO after 45-min deposition time b) on TiO 2 after 3 hour deposition time ............................................................................... 51

Figure 3.3 absorption data from standard optical absorption (linear scale) and photothermal deflection spectroscopy (logarithmic scale). ........................................ 53

Figure 3.4 semilog J- V curves. The dashed lines are dark curves while the solid lines are light curves. The bias polarity is defined as positive when the positive bias is applied to ITO. ............................................................................................................ 53

Figure 3.5 J-V curve for different blend annealing timing. ........................................ 54

Figure 3.6 Semilog J-V curves taken in the light (with symbols) and the dark (no symbols) for P3HT:PCBM bulk-heterojunction devices with Al electrode (solid line) and Au electrode (dot-dash line) as well as P3HT only (dotted line) with Al electrode. ..................................................................................................................................... 55

Figure 3.7 Liner J-V curves for the devices presented above. .................................... 56

Figure 3.8 A proposed energy diagram for bulk heterojunction devices containing an In 2 S 3 charge transport layer. ....................................................................................... 57

Figure 3.9 The external quantum efficiency compared to the absorption of the P3HT:PCBM bulk heterojunction solar cell. .............................................................. 59

ix

Figure 4.1 Experimental setup for a) widefield and b) microscopic thermoreflectance imaging of the solar cell c) and d) are amplitude(∆/) and phase image of a shunt in an Evergreen cell .................................................................................................... 68

Figure 4.2 False color, uncalibrated thermal images of a textured Evergreen solar cell operated at -13V with no illumination. ....................................................................... 71

Figure 4.3 a) Thermoreflectance amplitude map of an NREL test cell (square wave modulated 0-1V at 5Hz) shows a large diffuse shunt; b) Local IV curve measured at the position of the shunt, contrasted with whole cell IV curve. Measured data are fit with the Shockley diode equation. .............................................................................. 73

Figure 4.4 (a) Thermoreflectance phase image of shunts in a reverse biased (-12V) Evergreen solar cell at various square wave modulation frequencies (b) Open circles show measured phase lag versus distance from the heat source for the center shunt in the 1Hz image; solid line shows best fit of Equation 4, using α as a fitting parameter. ..................................................................................................................................... 76

Figure 4.5 (a) Thermoreflectance magnitude image of shunts in a reverse biased (- 12V) Evergreen solar cell at various square wave modulation frequencies (b) Open circles show measured temperature versus distance from the heat source for the center shunt in the 1Hz image ............................................................................................... 77

Figure 4.6 a) Schematic of TE (thermoelectric heater/cooler driven solar cell b) solar cell top surface temperature swing with TE applied current modulated -2A to +2A at 0.2 Hz (square wave). ................................................................................................. 78

x

Figure 4.7 Crack detection suing thermoreflectance imaging .................................... 81

Figure 4.8 Diagrams of AB and IFE and resulting current-voltage characteristic of AB and IFE processes. ................................................................................................ 83

Figure 4.9 a) EL of forward biased (0.6V) NREL sample A; b) EL of reverse biased (- 13V) NREL sample A; c) EL of forward biased (0.6V) Evergreen sample B; d)EL of reversed biased (-20V) Evergreen sample B .............................................................. 85

Figure 4.10 forward-based EL of a block-cast mc-silicon solar cell. The framed dark area is due to the broken contact ................................................................................. 86

Figure 4.11 reverse biased EL of Evergreen silicon solar cell at voltage a)-5V, b)-14V, c)-16V d) EL intensity versus applied voltage for spot 1 (blue circle), spot 2 (red diamond) and spot 3 (black triangle). ......................................................................... 88

xi

List of Tables Table 2-1 Thioacetamide and Acetonitrile phase transition temperature. .................. 17

Table 2-2 Precursor duration time and purge time during ALD process. .................. 22

Table 2-3 XPS data for ALD Cu x S deposited on ZnO ............................................... 32

Table 4-1 Thermoreflectance calibration values for various cell types and illumination methods. All of the cells tabulated here are SiN-coated, ribbon-pulled, mc-Si solar cells from Evergreen Solar.......................................................................................... 68

Abstract Thin film solar cells have the potential to lower the solar cell cost. The global market share of the thin film solar cells have been increased from 5% to 17.7% in just 6 years (2003-2009) compared to the crystal silicon solar cells. In this work, a Chemical Vapor Deposition (CVD) system was converted into an Atomic Layer Deposition system by hardware modification, including the installation of metal-organic precursor bubbler and in-line H 2 S generator, and computer automation realization by Labwindows programming. ALD Cu x S film, a good absorber layer, was deposited on both corning glass and nanostructured TiO 2

layer. SEM image of ALD CuS coated 100nm nanostructured TiO 2 showed perfect penetration of ALD film, which resulted from self-saturated surface reaction of ALD. Another absorber layer, SnS was also deposited by the same ALD system using dibutyltin diacetate (DBTDA) metal-organic precursor as the tin source. It is the first time DBTDA was used as tin source in ALD SnS deposition. ALD CuS and SnS films are proposed to be used in extremely thin absorber (ETA) solar cells. Chemical Bath Deposition (CBD) of Indium sulfide was also demonstrated in this thesis. It was the first time that CBD indium sulfide was used in P3HT:PCBM bulk heterojunction organic solar cell as the buffer layer between ITO and P3HT:PCBM to replace the problematic PEDOT:PSS. By using different front

contact metal (Al and Au), it was found that In 2 S 3 can transport both electrons and holes, which makes it a really good buffer layer in the tandem cells serving both hole transporter and electron transporter, where usually two buffer layers are needed. The efficiency of the un-optimized single junction device using In 2 S 3 as hole transporter is 1.5%. The last part of the thesis shows thermoreflectance (TR) imaging and Electroluminescence (EL) imaging of poly-crystalline silicon solar cells. These two imaging techniques in principle can be used in thin film solar cell characterization also. The local I-V curve on a diode shunt was obtained using high spatial resolution TR imaging for the first time. From local I-V curve on the defects, we can tell if a shunt is linear or diode behavior. The shunt we looked was a weak diode; by fitting local I-V curve with the diode equation, we got the ideality factor of that particular shunt much larger than 2, which indicated that the defect is probably caused by high concentration dislocation. Thermal diffusivity and thermal conductivity were also obtained from the TR phase image and amplitude image of the hot spot. Both TR image and EL image were used to detect the cracks. Analyzing the reverse biased EL image of the solar cell the cause of the pre-breakdown was discussed.

xiv

Acknowledgement First and foremost, I would like to thank my advisor Glenn Alers for his incredible support during my whole PHD period of time. Without his encouragement, endless patience and guidance I would not be where I am. He is not only a great mentor for my research, but also a great friend and co-worker. It has been pleasurable to work with him and I have learned so much from him that will become my treasure throughout my life. I would also like to thank my EE advisor Ali Shakouri for his passion to the research and students. He always led me to think physics intuitively. My gratitude also goes to Sue Carter. Working in her lab is my privilege. I would like to thank my intern advisor, Professor Janice Hudgings, for giving me the opportunity to work with her on thermoreflectance imaging. I appreciate Holger Schmidt for giving suggestions to my thesis writing, and to be my committee member. I would also like to thank all the members in the above four groups, both past and present for their willingness to answer my question, helpful discussions and friendship. Tong Ju, Chris France, Rebecca Graham, Lily Yang, Jeremy Olson, Josh Ford, Guangmei Zhai, Ben, Yvonne Rodriguez, Zhixi Bian, Xi Wang, Mona, Kerry

xv

Maze, Xiaolin Hu, Kadhair and more. I would also like to thank Thomas and Len from Mount Holyoke College for their technical and hardware support. I owe my thanks to my family, my parents and parents-in-law for their great support, my daughter Duanduan for being so reasonable, and my husband, Yingcai for his encouragement, his inspirational discussion and his enthusiasm to research. Without him, I would have not started this program. And thanks to all my friends and people that I have ever met!

1

1 Literature review 1.1 Solar Cell Progress Solar energy in one form or another is the source of nearly all energy on the earth. PV devices (solar cells) are unique in that they directly convert the incident solar radiation into electricity, with no noise, pollution or moving parts, making them robust, reliable and long lasting. Since the discovery of a p–n junction Si photovoltaic (PV) device[29] reported in 1954, the science and technology of PV devices (solar cells) and systems have undergone revolutionary developments. The efficiencies of all PV cell types have improved over the past several decades, as illustrated in Figure 1.1, which shows the best research-cell efficiencies from 1975 to 2008. The highest- efficiency research cell shown is a multijunction concentrator at 41.6% efficiency. Other research-cell efficiencies illustrated in the figure range from 20% to almost 28% for crystalline silicon cells, 12% to almost 20% for thin film, and about 5% and to 11% for the emerging PV technologies organic cells and dye-sensitized cells, respectively. The installation capacity also increased over the years. Figure 1.2 shows the global cumulative installed PV capacity from 1992 to 2008. From 2005 to 2008, the PV installation increased over 250%, up to year of 2008, the total installed PV was 13.9 GW, and the total revenue of about 20 billion US dollars.

2

Figure 1.1 Best research cell efficiencies from 1975-2009[61]. The application fields of the solar cell have increased as well, started from only used in the spacecraft, to today, it has been used as domestic electric power by roof top installation. The motivations that have stimulated high demand for solar cells include: 1. The need for low maintenance, long lasting sources of electricity suitable for places remote from both the main electricity grid and from people; e. g. satellites, remote site water pumping, outback telecommunications stations and lighthouses. 2. The need for cost effective power supplies for people remote from the main electricity grid; e. g. Aboriginal settlements, outback sheep and cattle stations, and some home sites in grid connected areas.

3

3. The need for non-polluting and silent sources of electricity; e. g. tourist sites, caravans and campers . 4. The need for a convenient and flexible source of small amounts of power; e. g. calculators, watches, light meters and cameras. 5. The need for renewable and sustainable power, as a means of reducing global warming, such as residential roof top solar panels, solar electricity plants.

Figure 1.2 Global cumulative installed PV capacity by interconnection status[2]. The main type of the solar cell in the market now is silicon solar cell, which has more than 80% of market. In the United States, the average levelized cost of energy with ITC is 0.20$/KWh in Phoenix in 2008, that is still high than the grid energy cost. For the silicon solar cell, silicon substrate is cut from ingot or block, and half of the silicon material is wasted as saw dust. The average silicon utilization rate

4

is 8.7 g/W and the polysilicon contract price in 2009 was about $70/kg[115], which makes the material cost about 0.6$/W. The cost limitation of silicon solar cells has been realized at the very beginning and varies thin film solar cells has been proposed[37]. After decades of efforts, the thin film solar cells have occupied the market share of 17.7% in 2009[39], compared to 5% in 2003, and is expecting to play more and more important roles in the PV market.

Figure 1.3 LCOE for residential PV systems in several U.S. cities in 2008, with and without the federal investment tax credit [99].

1.2 Thin Film Solar Cells Thin film solar cells aim to reduce the cost of crystal silicon technology, as they use less than 1% of the raw material (silicon) compared to wafer based solar

5

cells. Thin film technology basically consists in coating a glass or ceramic substrate with a thin film of semiconductor, such as cadmium telluride (CdTe), copper indium gallium selenide (CIGS), amorphous silicon (a-Si) or nanocrystalline silicon. These materials are all direct band-gap, strong light absorbers and only need to be about 1 micron thick, so materials costs are significantly reduced. Other advantages of the thin film solar cells are low-cost thin film deposition techniques and flexible substrates. The earliest thin-film cells were based on Cu 2 S/CdS[37] and suffered from poor stability owing to the high diffusivity of Cu. Amorphous hydrogenated silicon (a-Si:H) cells entered the PV market in the 1980s and today are increasingly challenged by CdTe and Cu(In,Ga)Se 2 based cells. One of the problems with thin-film materials other than a-Si:H is that they are not used elsewhere in the electronics industry. Therefore, there is comparatively little expertise about them. Nevertheless, the presently most efficient (single-junction) thin-film solar cells are made with polycrystalline CuIn 1−x Ga x Se 2 (commonly abbreviated as Cu(In,Ga)Se 2 or CIGS) or CdTe absorbers. CIGS solar cells with efficiencies greater than 20% have been claimed by both the National Renewable Energy Laboratory (NREL)[108] and the Zentrum für Sonnenenergie und Wasserstoff Forschung (ZSW), which is the record to date for any thin film solar cell[1]

6

CdTe, CIGS and a-Si:H based solar cells have been commercialized. Among these three types of solar cell, CdTe solar cells has the biggest market share, produced mainly by First Solar. First Solar, the largest U.S. PV cell/module manufacturer of thin-film (CdTe) modules with the production capacity of 1.4GW in 2010 and is expecting 2.7 GW in 2012 based on current line run rates. First Solar’s manufacturing cost per watt reached $1.23 in 2007 and $1.08 in 2008. On February 24, 2009, the cost/watt ratio broke the $1 barrier, reaching $0.98 per watt. In the third quarter of 2010, its production cost had fallen to $0.77 per watt[96]. Other thin film solar cells including dye sensitized solar cell (DSSC), Extremely Thin Absorber (ETA) solar cell and organic solar cell. DSSC and organic solar cells have been attractive for their low cost and flexibility. Champion DSSC reaches 11% efficiency with module efficiency of 5-7%. Organic champion cell efficiency is 5%. The application of those two cells is mainly for portable charging. 1.3 The Scope of Thesis This thesis includes two thin film deposition techniques for thin film solar cells as well as the poly-silicon solar cell defect imaging study. It has 4 chapters with the first chapter as this introduction. A custom build atomic layer deposition is demonstrated in Chapter 2; and the uniform and conformal ALD CuxS and SnxS films were deposited using this system. Those thin films are p-type semiconductor

7

with high absorption coefficient. They were proposed to be used in nano/micro structured solar cells as extremely thin absorber layer. Chapter 3 introduces the chemical bath deposition of In 2 S 3 . In 2 S 3 can be used in CIGS solar cell as a window layer to replace toxic CdS. In our case, it was used as a carrier transporter/barrier layer in heterojunction organic solar cell. Chapter 4 covers two defect characterization techniques: thermoreflectance imaging and Electroluminescence imaging. The grain boundaries, shunts, cracks and pre-breakdown sites of poly-silicon solar cells are detected using these two methods. In principle, both of them can be used in thin film solar cells as well.

8

2 ALD deposition of Cu x S and Sn x S and application in solar cells 2.1 Introduction 2.1.1 Character of Atomic Layer Deposition Atomic Layer Deposition (ALD), Also known as atomic layer epitaxy is a particularly suitable technique for making highly uniform and conformal layers[79]. ALD is similar to chemical vapor deposition (CVD), the difference is that two or more vapor reactants come to the reaction chamber in sequences. Usually one ALD reaction cycle has 4 steps as shown by Figure 2.1. First, a dose of vapour from one precursor is brought to the reaction chamber and got chemisorbed by substrate surface where a film is to be deposited; then any excess unreacted vapour or any byproducts is pumped away by inert gas purge step. Next, a vapour dose of the second reactant is brought to the surface and reacts with the previous chemisorbed layer to form the desired film and the excess is pumped away. This one cycle normally deposit a monolayer or less of material due to the steric hindrance, however multilayer deposition of one ALD cycle has been reported also[45]. In some cases, plasma is used to assist the reaction, which is called Plasma Enhanced ALD (PEALD). ALD process is complementary and self-limiting[79]. Be complementary means that each of the two reactants prepares the surface for its reaction with the other vapor reactant, so that the deposition cycles can be repeated to get thicker (more than one monolayer)

9

film. Be self-limiting means that the amount of material that can be deposited on to the surface is limited or saturating automatically by the surface deposited on to the surface is limited or saturating automatically by the surface chemistry. Provided that enough reactants and enough reaction time are present, there are two implications of self-limiting: (1) it will produce film with very uniform thickness over the whole area even if the vapour flux is delivered non-uniformly over the surface or there are very high aspect ratio holes and trenches on the surface[70] (that is, the film is highly conformal); (2) Every cycle produces same thickness so that the total film thickness has linear relation with the cycle numbers, which makes the easy thickness control.

Figure 2.1 Schematic of one ALD cycle. Unlike CVD, in which the substrate temperature and the gas delivery speed are important factors to its deposition rate and the film thickness, ALD deposition rate can be easily controlled by counting the cycle number. For a certain temperature

10

window, the thickness/cycle is constant due to its self-limiting characteristic as shown in Figure 2.2. However, for a give precursor delivery speed and exposure time, too low and too high temperature should be avoided: if the temperature is too low, the precursor will condensate on the surface (Case L1 in Figure 2.2) or the reaction is so slow that never reaches saturation so that the deposition rate is depending on temperature (Case L2); if the temperature is too high, thermal decomposition of the precursor will result in deposition in normal CVD fashion (Case H1) or the deposited films re-evaporate (Case H2).

Figure 2.2 The temperature dependence of the growth rate. As discussed above, ALD film has the advantage of high uniformity and conformity. However, due to its comparably slow deposition rate (~monolayer per cycle), it is usually used in the applications where only <100nm film is required. Different metal films, metal nitrides , metal sulfides, metal oxides and dielectric materials etc, as shown in Figure 2.3, have been deposited by ALD and It has been

11

used in microelectronics, including high-k gate dielectric[89], dynamic random access memories (DRAM)[78], transition metal nitride as metal barriers and as gate metals[58], metal thin film as electrical plating seed layer (Cu[79], W[67] etc.) or as metal barrier layer (Ru[3], Ta and Ti[109] etc.). In the following, we will introduce ALD metal sulfide film used in solar cell application.

Figure 2.3 Reported ALD films[102]. 2.1.2 ALD application in solar cells Varies ALD thin films have been studied to use in different kinds of solar cells. ALD Al 2 O 3 film was proposed to use in c-Si solar cells as a surface passivation layer[48]. Yousfi et al [124]studied ALD indium sulfide and ZnO as buffer layer and window layer respectively in CIGS solar cell and achieved efficiency of 13.5%. The

12

high value of open circuit voltage (V oc ) and fill factor (FF) suggested the good interface quality. With only replacing CdS with ALD indium sulfide film, the highest efficiency of CIGS solar cell reported is 16.4%[92]. ALD films are also reported to be used in Dye Sensitized Solar Cells (DSSCs)[76, 86]. ALD TiO2 anatase blocking layer was deposited as a shell of ZnO nanowire core based DSSCs[76]. The function of TiO2 shell is to suppress the recombination rate. 2.5% efficiency was achieved by this approach. The extremely thin absorber (ETA) cell is a solar cell concept that utilizes a highly-structured heterojunction interface to increase absorption while reducing the transport path for excited charge carriers[38]. The tolerance of the film quality is increased due to the decreased carrier transport path. The challenge in the development of solid state ETA solar cells is finding an effective technique for depositing semiconducting materials within nanoporous structures to form an inter- penetrating junction. As introduced in the previous section, ALD is a superior method to fulfill this requirement. Cu 2 S is an interesting absorber and also a p-type semiconductor with its indirect band gap near 1.2eV [93] and absorption coefficient of 10 5 cm -1 at 750 nm[46]. It does not contain poisonous materials and also it is relatively abundant. Cu 2 S has been studied to use in Cu 2 S /CdS devices and an efficiency of 9.15%[20] has been achieved. However, it was dropped because this PV cell is not stable due to

13

the reason of copper diffusion. On the other hand, Cu 2 S with n-type TiO 2 film and/or nanoporous structure is very stable and the combination makes good ETA solar cells. Previous ALD work, using Cu(thd) 2 (Copper bis(2,2,6,6-tetramethyl-3,5- heptanedionate)) as the copper precursor, yielded Cu 1.8 S and CuS with the phase of the deposited material determined by substrate temperature during the deposition[56, 107]. Other work successfully produced Cu 2 S using bis(N , N' -di-sec- butylacetamidinato)dicopper(I) as the copper precursor[85]. Tin sulfide compounds have recently attracted considerable attention for the optoelectronic applications because of their physical and electrical properties. Sn (II)S compound in thin film form is a p-type semiconductor with the band gap between 1.1 ~ 1.7 eV depending on the preparation method[7] and an absorption coefficient about 10 6 cm − 1 [35], which makes it a great absorber material for photovoltaic applications. In addition, the cost could be reduced because the materials involved are cheap, non-toxic and abundant in nature. Tin sulfide films have been deposited by various methods, e. g., spray pyrolysis[27], chemical bath deposition[119], evaporation[35], chemical vapor deposition [8, 112], etc. 2.2 Precursor Consideration For copper source, first we tried Cu(thd) 2 (thd=2,2,6,6-tetramethyl-3,5-heptane dionate) (see formula in Figure 2.4)[103] from Strem Chemicals since it has been

14

used for ALD Cu x S by other group[56] and it is very stable even in ambient atmosphere because it doesn’t absorb water due to its bulky ligands . So it is very easy to handle. Cu(thd) 2 is a purple crystal powder, at temperature 115 o C±3 o C, the vapor pressure is around 1.5Torr[56]. It is thermally stable and the decomposition temperature was reported to be around 179 o C[97]. Next we tried KI5 (see formula in Figure 2.5) from Air Products. This is a flourine free copper precursor, previously used for ALD of Cu, A close descendant of the commonly used Cu(hfac)(tmvs) precursor, KI5 is stable at high temperatures, allowing for high vapor pressures (>2 Torr at 140°C) and deep penetration into highly structured surfaces[98]. The precursor contains no fluorine, reducing possible contamination effects[51].

Figure 2.4 Cu(thd) 2 formula.

15

Figure 2.5 KI5 formula. For ALD Sn x S, dibutyltin diacetate (DBTDA) (Figure 2.6) was used as the tin source[104], DBTDA was used to deposit SnO 2 films by spray pyrolysis[91], CVD[60], PEALD[30, 69]. It was the first to be used to deposition Sn x S by ALD. DBTDA is an oily, air-stable, colorless liquid compound having a vapor pressure of 1 Torr at 120 o C which makes it convenient for transporting its vapor in CVD work. The compound does not pose severe safety hazards and is much cheaper and more readily available than any of the other compounds we considered including bis(bis(trimethylsilyl)amino)Tin(II), which is not volatile enough due to the existence of Si, and trimethyl tin, which is too flammable and highly toxic.

Full document contains 146 pages
Abstract: Thin film solar cells have the potential to lower the solar cell cost. The global market share of the thin film solar cells have been increased from 5% to 17.7% in just 6 years (2003-2009) compared to the crystal silicon solar cells. In this work, a Chemical Vapor Deposition (CVD) system was converted into an Atomic Layer Deposition system by hardware modification, including the installation of metal-organic precursor bubbler and in-line H2 S generator, and computer automation realization by Labwindows programming. ALD Cux S film, a good absorber layer, was deposited on both corning glass and nanostructured TiO2 layer. SEM image of ALD CuS coated 100nm nanostructured TiO2 showed perfect penetration of ALD film, which resulted from self-saturated surface reaction of ALD. Another absorber layer, SnS was also deposited by the same ALD system using dibutyltin diacetate (DBTDA) metal-organic precursor as the tin source. It is the first time DBTDA was used as tin source in ALD SnS deposition. ALD CuS and SnS films are proposed to be used in extremely thin absorber (ETA) solar cells. Chemical Bath Deposition (CBD) of Indium sulfide was also demonstrated in this thesis. It was the first time that CBD indium sulfide was used in P3HT:PCBM bulk heterojunction organic solar cell as the buffer layer between ITO and P3HT:PCBM to replace the problematic PEDOT:PSS. By using different front contact metal (Al and Au), it was found that In 2 S3 can transport both electrons and holes, which makes it a really good buffer layer in the tandem cells serving both hole transporter and electron transporter, where usually two buffer layers are needed. The efficiency of the un-optimized single junction device using In 2 S3 as hole transporter is 1.5%. The last part of the thesis shows thermoreflectance (TR) imaging and Electroluminescence (EL) imaging of poly-crystalline silicon solar cells. These two imaging techniques in principle can be used in thin film solar cell characterization also. The local I-V curve on a diode shunt was obtained using high spatial resolution TR imaging for the first time. From local I-V curve on the defects, we can tell if a shunt is linear or diode behavior. The shunt we looked was a weak diode; by fitting local I-V curve with the diode equation, we got the ideality factor of that particular shunt much larger than 2, which indicated that the defect is probably caused by high concentration dislocation. Thermal diffusivity and thermal conductivity were also obtained from the TR phase image and amplitude image of the hot spot. Both TR image and EL image were used to detect the cracks. Analyzing the reverse biased EL image of the solar cell the cause of the pre-breakdown was discussed.