failure analysis and long term stability of thin film ... · i want to thank felix hoga, urs...

i

Failure Analysis and Long Term Stability of

Thin Film Solar Cells and Modules

Fehleranalyse und Langzeitstabilitätsmessungen von Dünnschichtsolarzellen und Modulen

Der Technischen Fakultät

der Friedrich-Alexander-Universität

Erlangen-Nürnberg

zur

Erlangung des Doktorgrades Dr. Ing.

vorgelegt von

Dipl.-Phys. Jens Adams

aus Berlin

Als Dissertation genehmigt

von der Technischen Fakultät

der Friedrich-Alexander-Universität Erlangen-Nürnberg

Tag der mündlichen Prüfung: 21. Oktober 2015

Vorsitzender des Promotionsorgans: Prof. Dr. Peter Greil

1.Gutachter: Prof. Dr. Christoph J. Brabec

2.Gutachter: Prof. Dr. Vladimir Dyakonov

Zusammenfassung

i

Zusammenfassung

Fehleranalyse und Langzeitstabilitätsmessungen von Dünnschichtsolarzellen

und Modulen

Dünnschichtsolarzellen und –module, basierend auf organischen und

anorganischen Halbleitermaterialen, wurden innerhalb der letzten 10 Jahre zu

hoch effizienten Leistungsgeneratoren entwickelt. Diese Zellen bestehen aus

einem Stapel verschiedener dünner Zwischenschichten, welche von einigen

Nanometern bis hin zu einigen Mikrometern variieren können. Durch ungenaue

Zellherstellung als auch durch umweltbedingte Alterungsprozesse können sich

Defekte bilden, welche die Leistungsfähigkeit und Lebensdauer der Solarzellen

beeinträchtigen. Aus diesem Grund ist ein genaues Verständnis des

thermischen und elektrischen Einflusses von unterschiedlichen

Verlustmechanismen entscheidend für eine Verbesserung der Zuverlässigkeit

von Solarzellen und -module

Diese Arbeit wurde aus der Motivation heraus initiiert, das derzeitige

Verständnis bezüglich des thermischen und elektrischen Einflusses von

unterschiedlichen Defekten in Dünnschichtsolarzellen zu verbessern. Das Ziel

dieser Arbeit ist es, mittels bildgebender Messverfahren unterschiedliche

Verlustprozesse in Solarzellen zu charakterisieren und zu quantifizieren. Hierzu

werden verschiedene Alterungsexperimente von organischen Solarzellen in

Verbindung mit bildgebenden Messverfahren und elektrischen

Charakterisierungsmethoden durchgeführt. Besonders die Kombination von

elektrischen Charakterisierungen mit bildgebenden Messverfahren ermöglicht

eine detaillierte Untersuchung verschiedener lokaler Rekombinations- und

Verlustprozesses von licht- und strominduzierten Ladungsträgern.

Zusammenfassung

ii

Neben einer kurzen Einführung, welche auf die derzeitige Problematik

bezüglich leistungsminimierender Verlustmechanismen in

Dünnschichtsolarzellen hinweist, wird in Kapitel 2 der theoretische Hintergrund

zur Beschreibung von organischen und anorganischen Dünnschichtsolarzellen

präsentiert. Kapitel 3 geht anschließend auf die unterschiedlichen

Verlustmechanismen von Ladungsträgern in Solarzellen ein und präsentiert

zudem eine detaillierte Beschreibung des technischen Standes von

bildgebenden Messverfahren, mit Hauptaugenmerk auf Lock-In Thermographie.

Eine Zusammenfassung der unterschiedlichen Degradationsmechanismen von

organischen Solarzellen wird in Kapitel 4 präsentiert. Die Bescheinigung der

experimentellen Aufbauten sowie deren Messprinzipen, welche für diese Arbeit

verwendet werden, ist in Kapitel 5 zu finden.

Im Detail wird im Kapitel 6 die Langzeitstabilität von organischen

Tandemsolarzellen unter kontinuierlicher Beleuchtung mit sichtbarem Licht

untersucht. Die untersuchten Testzellen zeigen eine Verlustleistung von 11%

innerhalb der ersten 2000 Betriebsstunden. Die hohe Stabilität wird in erster

Linie durch eine invertierte Zellenstruktur, der Verwendung von MoOx anstelle

von PEDOT:PSS als Loch-Extraktions-Schicht, sowie der Modifikation der

ZnO/Halbleitergrenzschicht mit Ba(OH)2 erzielt. An Hand von verschiedenen

Alterungsexperimenten wird der Einfluss und die Bedeutung von ultravioletter

(UV) Strahlung, für Zellen welche eine ZnO Zwischenschicht beinhalten,

untersucht. Experimente mit unterschiedlichen Lichtbedingungen zeigen zu

dem die Kinetik der S-Verformung der IV-Kennlinie. Des Weiteren kann der

Ursprung der S-Verformung in der Tandemzelle genau lokalisiert werden.

Kapitel 7 bezieht sich auf die Charakterisierung der temperatur- und

feuchtigkeitsbedingten Alterung von invertierten und glasverkapselten

organischen Solarzellen. An Hand der Kombination unterschiedlicher

bildgebender Infrarot- und Lumineszenzmessungen, sowie der elektrischen

Charakterisierung der Testzellen können wertvolle Erkenntnisse von

feuchtigkeitsbedingten Alterungspfaden gewonnen werden. Für die präsentierte

Studie werden glasverkapselte Testzellen bei unterschiedlichen kontrollierten

Umweltbedingungen gelagert. Neben Lebenszeitdauern von ca. 20.000

Zusammenfassung

iii

Stunden zeigen die Untersuchungen, dass die Diffusion von Feuchtigkeit einen

Hauptgrund für eine beschleunigte Alterung bei hohen Umgebungs-

temperaturen darstellt. Die Quantifizierung der temperaturabhängigen

Beschleunigungsfaktoren, bestimmt nach dem Arrheniusmodel, zeigt zu dem

eine Aktivierungsenergie der feuchtigkeitsbedingten Alterung von ~450 meV.

Die Messung der ortsaufgelösten Photo- und Elektrolumineszenz offenbart,

dass Wasser in erster Linie die Elektroden/Halbleitergrenzschicht der Testzellen

beeinträchtigt und nicht das Halbleitermaterial selbst. Mittels unterschiedlicher

Elektrodengeometrien wird gezeigt, dass Feuchtigkeit primär durch einer der

Zellschichten diffundiert.

Neben der Charakterisierung unterschiedlicher Degradationspfade in

organischen Solarzellen konzentriert sich Kapitel 8 auf den thermischen und

elektrischen Einfluss von produktionsbedingten Defekten in

Dünnschichtsolarmodulen. Dazu wurden 15 Testsolarmodule aus einer

großindustriellen Produktionslinie für CIGS Module entnommen und für die

Charakterisierung zur Verfügung gestellt. Jedes der einzelnen Testmodule

enthält mehrere produktionsbedingte Fehlerstellen, welche die maximale

Leistung eines Modules beeinträchtigten. Mit „illuminated lock-in

thermography“ (ILIT) kann jeder einzelne Defekt im Modul lokalisiert und

dessen elektrischen Einfluss auf die ihm umgebene Zelle charakterisiert und

quantifiziert werden. Um dies zu erreichen, wurde eine neue Methode zur

Spannungsbestimmung einzelner Zellen im Module entwickelt. Das ermittelte

Verhältnis aus IR-Emission eines Defektes und dessen verursachten

Spannungseinbruch der Zellspannung zeigt, dass über 95% der untersuchten

Defekte lediglich eine Minimierung der Zellspannung von weniger als 20%

verursachten. Es wird gezeigt, dass besonders starke Defekte eine schwache

IR Emission aufweisen und gleichzeitig einen starken Spannungseinbruch

verursachen. Mittels Computersimulationen kann das Phänomen bestätigt und

auf eine Sättigung des Defektstromes zurückgeführt werden.

Acknowledgement

iv

Acknowledgement

The first person I would like to thank is Professor Christoph J. Brabec who

gave me the opportunity to make this thesis under his supervision. Thanks for

fruit full discussions and for the right words at the right time. Furthermore, I

would like to thank Professor Vladimir Dyakonov, who had agreed to act as a

second supervisor for this thesis.

I would like to thank Dr. Monika M. Voigt, Dr. Claudia Buerhop-Lutz, Dr.

Michael Salvador, and Dr. Hans-Joachim Egelhaaf for their guidance of the last

five years. Special thanks go to Dr. Michael Salvador for sharing his

experiences and knowledge of OPV as well as for asking questions which

sometimes drove me crazy. Furthermore, I would like to thank Dr. Monika M.

Voigt who introduced me to the world of OPV and who gave me the opportunity

to see the world.

I would like to thank all members of the ZAE, I-Meet, and EnCN for their

support and for a nice atmosphere. Special thanks go to Frank, Luca, George,

Stefan, Andi, Simon, and Sebastian for supporting this thesis by making cells,

simulations, measurements, and discussions. I thank George, Frank, Thomas,

Fei, and Anastasia for an awesome time in the USA.

I want to thank Felix Hoga, Urs Bogner, and Karl Borutta for the technical

support of this thesis and for good times in Nürnberg and Franken.

I want to thank my lovely wife Rena Hamabata, my family on both sides of

the world and all my friends in Germany and Japan for their support during this

time.

Publications

v

Publications

2015

J. Adams, M. Salvador, Luca Lucera, George Spyropoulos, F.W.

Fecher, M.M. Voigt, S. A. Dowland, A. Osvet, H. Egelhaaf, C.J

Brabec “Water ingress in encapsulated inverted organic solar

cells: correlating infrared imaging and photovoltaic performance”,

Advanced Energy Materials, aenm.201501065 (submitted)

2014

J. Adams, A. Vetter, F. Hoga, F. Fecher, J.P. Theisen, C.J.

Brabec, C. Buerhop-Lutz, “The influence of defects on the cellular

open circuit voltage in CuInGaSe2 thin film solar modules—An

illuminated lock-in thermography study”, Solar Energy Materials &

Solar Cells 123 (2014) 159–165

J. Adams, G. D. Spyropoulos, M. Salvador, N. Li, S. Strohm, L.

Lucera, S. Langner, F. Machui, H. Zhang, T. Ameri, M. M. Voigt, F.

C. Krebs and C. J. Brabec, “Air-processed organic tandem solar

cells on glass: toward competitive operating lifetimes”, Energy

Environ. Sci., 2015,8, 169-176

J. Adams; F. W. Fecher ; F. Hoga ; A. Vetter; C. Buerhop, C.J.

Brabec "IR-imaging and non-destructive loss analysis on thin film

solar modules and cells", Proc. SPIE 9177, Thin Films for Solar

and Energy Technology VI, 917703 (October 3, 2014);

doi:10.1117/12.2061724; http://dx.doi.org/10.1117/12.2061724

http://profiles.spiedigitallibrary.org/summary.aspx?DOI=10.1117%2f12.2061724&Name=Jens+Adams


http://profiles.spiedigitallibrary.org/summary.aspx?DOI=10.1117%2f12.2061724&Name=Felix+Hoga

http://profiles.spiedigitallibrary.org/summary.aspx?DOI=10.1117%2f12.2061724&Name=Andreas+Vetter


http://dx.doi.org/10.1117/12.2061724

Publications

vi

T. R. Andersen, H. F. Dam, M. Hösel, M. Helgesen, J. E. Carlé,

Thue T. Larsen-Olsen, S. A. Gevorgyan, J. W. Andreasen,

J. Adams, N. Li, F. Machui, G. D. Spyropoulos, T. Ameri,

N. Lemaître,. M. Legros, A. Scheel, D. Gaiser, K. Kreul, S. Berny,

O. R. Lozman, S. Nordman, M. Välimäki, M. Vilkman, R. R.

Søndergaard, M. Jørgensen, C. J. Brabec and F. C. Krebs,

“Scalable, ambient atmosphere roll-to-roll manufacture of

encapsulated large area, flexible organic tandem solar cell

modules”, Energy Environ. Sci., 2014, 7, 2925–2933

F. Livi, Dr. R. R. Søndergaard, Dr. T. R. Andersen, B. Roth,

Dr. S. Gevorgyan, Dr. H. F. Dam, Dr. J. E. Carlé, Dr. M. Helgesen,

G. D. Spyropoulos, J. Adams, Dr. T. Ameri, Prof. C. J. Brabec,

Dr. M. Legros, Dr. N. Lemaitre, Dr. S. Berny, Dr. O. R. Lozman,

Dr. S. Schumann, Dr. A. Scheel, P. Apilo, Dr. M. Vilkman,

Dr. E. Bundgaard, and Prof. F. C. Krebs, “Round-Robin Studies

on Roll-Processed ITO-free Organic Tandem Solar Cells

Combined with Inter-Laboratory Stability Studies”, Energy

Technology. doi:10.1002/ente.201402095

F. W. Fecher, J. Adams, A. Vetter, C. Buerhop-Lutz, C. J. Brabec,

“Loss analysis on CIGS-modules by using contactless, imaging

illuminated lock-in thermography and 2D electrical

simulations“ Photovoltaic Specialist Conference (PVSC), 2014

IEEE 40th, 3331 – 3334, doi: 10.1109/PVSC.2014.6925648

2013

Vetter, F. Fecher, J. Adams, R Schaeffler, J. P. Theisen,

C. J. Brabec & C. Buerhop, “Lock-in thermography as a tool for

quality control of photovoltaic modules“, Energy Science and

Engineering 2013; 1(1): 12–17

Oral Presentation

vii

2012

Cl. Buerhop, J. Adams, F. Fecher, C. J. Brabec, Lock-in

Thermographie an Dünnschichtmodulen, ep Photovoltaik aktuell,

no. 7/8 (2012) pp. 37-41

Oral Presentation

2015

J. Adams, M Salvador, F. W. Fecher, L. Lucera, S. Langner, A.

Osvet, M. M. Voigt, H. J. Egelhaaf, C. J. Brabec, LOPEC 2015,

Munich, Germany, “Characterization of moisture induced

degradation of organic solar cells using non-destructive lock-in

NIR/IR imaging techniques”

2014

J. Adams; F. W. Fecher ; F. Hoga ; A. Vetter; C. Buerhop, C.J.

Brabec, SPIE 14, San Diego, USA, August, " IR-imaging and non-

destructive loss analysis on thin film solar modules and cells"

J. Adams, F. W. Fecher, F. Hoga, S. Besold A. Vetter, C.

Buerhop, M. M. Voigt, C. J. Brabec, ISFOE14, Thessaloniki,

Greece, „ Loss analyses on organic solar cells and modules by

using contactless and non-destructive infrared lock-in imaging

techniques”

J. Adams.,1st Joint Workshop on Organic Electronics, May 2014,

ZAE Bayern, Erlangen, Germany “Long term stability of OPV

single junction and tandem cells“

F. W. Fecher, J. Adams, A. Vetter, C. Buerhop-Lutz, C. J. Brabec,

IEEE 2014, Denver, USA, June, “Loss analysis on CIGS-modules

by using contactless, imaging illuminated lock-in thermography

and 2D electrical simulations“, given by A. Vetter



http://profiles.spiedigitallibrary.org/summary.aspx?DOI=10.1117%2f12.2061724&Name=Felix+Hoga



Poster Presentation

viii

H.-J. Egelhaaf, J. Adams, F. Fecher, M. Salvador, B. van der Wiel,

T. Sauermann, A. Distler, U. Dettinger,H. Peisert, T. Chassé, X.

Wang, D. Zhang, and C.J. Brabec, ISOS 7, Chambery, France,

„Lifetime of Printed OPV – Aspects of Intrinsic Stability and

Packaging”, given by M. Salvador

Poster Presentation

2013

J. Adams, L. Lucera, M. M. Voigt, C. J. Brabec, ISOS 6,

Chambery, France, “ELLI-Imaging on degraded and inverted

P3HT-PCBM solar cells”

2011

J. Adams, F. Fecher, R. Schäffler, R. Auer, C. J. Brabec, C.

Buerhop-Lutz, 26th European Photovoltaic Solar Energy

Conference and Exhibition, Hamburg, Germany „Influence of

Macroscopic Defects at Different Position in CIGS Modules on

Optical Imaging Techniques“

Academic Work

2014

Chairman, SPIE14, San Diego, USA, Session 5: “Light trapping

and anti-reflection”

Table of Figures

ix

Table of Figures

Figure 2.1: Crystalline lattice structure of a CIGS chalcopyrite unit cell. 6

Figure 2.2: a) Optical band structure of an irradiated CIGS solar cell

under short circuit condition. 7

Figure 2.3: 2 carbon atoms and 4 hydrogen (H) atom from an ethane

molecule. 9

Figure 2.4: Chemical structure of two common used donor and acceptor

materials used to form a BHJ absorber layer of an organic

solar cell. 10

Figure 2.5: Illustration of the different physical processes which take

place inside an organic and bulk hetero junction solar cell

under irradiation. 11

Figure 2.6: a) Illustration of cell stack configurations used for single

junction OPV devices. 12

Figure 2.7: Illustration of cell stacks used for multifunction OPV Devices

with top and bottom cells connected in series. 14

Figure 2.8: Schematic module layout with monolithic interconnection of

cells illustrated as top view (a) and cross section (b). 16

Figure 2.9: Replacement circuit of the one diode model 17

Figure 2.10: IV characteristics of a dark and irradiated organic solar cell

solar cell 18

Figure 3.1: Generation and recombination processes of electron hole

pairs in thin film solar cells under the Voc condition 22

Figure 3.2: Simplified illustrations of the different basic defect classes. 23

Figure 3.3: (a) Simulated distribution of the local electrical potential of

the electrodes layers of a shunted CIGS solar cell under the

Voc condition. 24

Figure 3.4: Simulated local electrical power density of a CIGS cell with a

defect as a function of cell position. 26

Figure 3.5: DLIT images of a multicrystalline solar cell measured at (a) -

0.5 V, (b) +0.5 V Image a and b are scaled to 0 mK (black)

Table of Figures

x

to 5 mK (white). Image is reproduced with permission from

[111]. 29

Figure 3.6: DLIT images of the local distributed current density 30

Figure 3.7: a) Photograph of organic thin film solar modules with three

in series connected cells. 32

Figure 3.8: DLIT measurement of a commercial available a-Si thin film

solar module with randomly distributed defects. 33

Figure 3.9: DLIT image (left) and ILIT image (right) of a silicon solar cell. 34

Figure 3.10: Comparison between DLIT and Voc-ILIT images excited

with different light spectra. 36

Figure 3.11: Relation between module maximum peak power Pmpp and

ILIT defect value X for 103 thin film CIGS solar modules. 37

Figure 3.12: (left) Voc-ILIT measurement of flexible OPV modules with

10 cells connected in series. 38

Figure 3.13: Image of the local Rs distribution of a multicrystaline Si cell

calculated using Eq 8. Image is reproduced with permission

from [148]. 40

Figure 3.14: (a) Photograph of the investigated module. 42

Figure 3.15: (left) EL images (top) and PL images (bottom) of a

P3HT:PCBM solar cell measured after different dark

storage times. 43

Figure 4.1: Different oxidation mechanism of P3HT polymer. 47

Figure 4.2: LBIC images of an inverted P3HT:PCBM solar cell with

PEDOT:PSS as HTL at different aging stages. 50

Figure 4.3: Absorption and desorption of oxygen at the ZnO interface 52

Figure 5.1: Typical degradation characteristic of an OPV. 56

Figure 5.2: Photograph of a representative sample holder used for

degradation studies, developed at ZAE-Bayern. 58

Figure 5.3: LED solar simulator used for illuminated JV characteristics 60

Figure 5.4: Irradiation spectra from the sun (AM1.5) and the LED solar

simulator. 60

Figure 5.5: Emission spectrum of the LEDs used for photo aging. 62

Table of Figures

xi

Figure 5.6: Circuit diagram showing the open circuit voltage evaluation

of a single thin film solar cell within a module with solar cells

each connected in series. 63

Figure 5.7: Local cell voltages of a CIGS-module with 36 connected

cells in series as measured by the above mentioned mask

method (at 30 W/m² illumination power). 65

Figure 5.8: Schematic image of lock-in calculation. 67

Figure 5.9: Schematically sketch of Lock-in setup for DLIT, ELLI and

ILIT applications 69

Figure 6.1: Operational stability of organic tandem solar modules 72

Figure 6.2: Absorption spectra of the active materials used for light

absorption in the tandem cell. 73

Figure 6.3: Schematic device representation of the tandem and single

cells investigated in the present photo-degradation study.

Reproduced with permission from [198] 74

Figure 6.4: JV-characteristics of a representative Tandem and

respective sub-cells measured at t=0 h. Reproduced with

permission from [198] 76

Figure 6.5: Lifetime of the UV light soaking state under continuous

photo-aging. 77

Figure 6.6: S-shape formation in the JV characteristic of a

representative OPV tandem cell. 78

Figure 6.7: Long-term decay of the UV light soaking (LS) state in the

dark. 79

Figure 6.8: Photo-aging of single and tandem OPV cells. 80

Figure 6.9: Relative change of device performance after 2000 h of

continuous white light illumination. 81

Figure 6.10: Comparison of operating lifetime of P3HT:PC60BM cells

with and without and additional Ba(OH)x interlayer. 82

Figure 6.11: Life time extrapolation from tandem PCE data presented in

the experiment shown in Figure 6.7. 83

Figure 7.1: a) Cross section of the inverted P3HT:PCBM devices used

in this study. 88

Table of Figures

xii

Figure 7.2: Impact of relative humidity at 65 °C on the temporal

evolution of Voc, Jsc, FF, and PCE for encapsulated

P3HT:PCBM solar cells of inverted architecture. 90

Figure 7.3: Periodically measured JV characteristics of an inverted

P3HT:PCBM solar cell stored in a climate chamber at

65°C/85%RH. 91

Figure 7.4: Long-term behavior of P3HT:PCBM solar cells stored at

different temperatures (7 °C/51%RH, 20 °C/63%RH,

50 °C/20%RH, 65 °C/20%RH) in the dark. 92

Figure 7.5: Temperature dependence of the acceleration factor K for Jsc,

FF, and PCE as extracted from the experimental data

shown in Figure 7.4. 94

Figure 7.6: ELLI images measured at different storage times using a

pulsed voltage of 1 V (forward bias). 96

Figure 7.7: Top to bottom: ELLI, DLIT, and ILIT images of the same cell

measured at different storage times. 97

Figure 7.8: Comparison of the electroluminescence (EL) and

photoluminescence (PL) signal of a fresh and a degraded

cell. 98

Figure 7.9: a) Water permeation model for inverted and encapsulated

organic solar cells based on ITO/Al-

ZnO/P3HT:PCBM/PEDOT:PSS/Ag. 102

Figure 8.1: a) Photograph of a 28 cm x 28 cm CIGS test module with 67

cells connected in series. 107

Figure 8.2: A defect (bright spot) identified by an ILIT-Voc measurement

(illumination power of 30 W/m², 1Hz lock-in frequency,

measuring time of 10 min): a) 0°-image, b) -90°-image, c)

amplitude image and d) phase image. Reproduced with

permission from [129] 109

Figure 8.3: a) Line scan (Sdiff) of a defective cell (amplitude image) in a

CIGS solar module (see Figure 8.2). 110

Figure 8.4: Line scans of a defective cell (amplitude image) in CIGS

solar module irradiated with different light intensity. 111

Figure 8.5: Simulated Voc,cell, current through the defect (defect

current), and power dissipation of the defect depending on

the defect resistance provided a defect size of 0.001 cm². 113

Table of Figures

xiii

Figure 8.6: a) Normalized open circuit voltage of different cells vs. the

defect IR-emission. 114

Abbreviations

xiv

Abbreviations

Ag silver

Al aluminum

AL active layer

AZO aluminum doped zinc oxide

BHJ bulk hetero junction

C carbon

Ca calcium

CCD charge coupled device

CdS cadmium sulfide

CIGS copper indium gallium selenide

CO2 carbon dioxide

CuGaSe copper gallium selenide

CuInSe copper indium selenide

DLIT dark lock-in thermography

EL electroluminescence

ETL electron transport layer

EQE external quantum efficiency

HOMO highest occupied molecular orbital

H2O water

HP high power

HTL hole transport layer

ILIT illuminated lock-in thermography

InGaAs indium gallium arsenide

InSb indium antimonide

IR Infrared

ISOS International Summit on Organic Photovoltaic

Stability

ITO indium tin oxide

LED light emitting diode

LiF lithium fluoride

Abbreviations

xv

LIT lock-in thermography

LS light soak

LUMO lowest unoccupied molecular orbital

MgF magnesium fluoride

Mo molybdenum

MoOx molybdenum oxide

MoSe molybdenum selenide

mpp maximum power point

NETD noise equivalent temperature difference

NIR near infrared

O2 oxygen

OC open circuit

OPV organic photovoltaic

P3HT Poly(3-hexylthiophen-2,5-diyl)

pDPP5T-2 diketopyrrolopyrrolequinquethio-phene

PCBM phenyl-C60-butyric acid methyl ester

PEDOT poly-3,4-ethylendioxythiophen

PL photoluminescence

PSS polystyrolsulfonat

QNR quasi neutral region

S camera signal

SC short circuit

SCR space charge region

Si silicon

TCO transparent conductive oxide

TFSC thin film solar cell

TOF-SIM time of flight secondary ion mess spectroscopy

TWI thermal wave imaging

WOx tungsten or wolfram oxide

ZnO zinc oxide

Symbols

xvi

Symbols

A amplitude

c speed of light m/s

cT specific heat K

E energy eV

Ea activation energy eV

FF fill factor % or %/100

flock-in lock-in frequency Hz

h Planck constant Js

I current A or mA

J current density A/m² or mA/cm²

P power W

λ wavelength of light m or nm

λT heat conductivity W/K

Λ heat diffusion length m

PCE photo conversion efficiency %

R resistance Ω

ρ material density Kg/m³

r distance m or cm

T temperature K or °C

t time h or s

V voltage V or mV

ν frequency of light Hz

ω frequency Hz

φ electric potential

Table of Content

xvii

Table of Content

Zusammenfassung .............................................................................................. i

Acknowledgement .............................................................................................. iv

Publications......................................................................................................... v

Oral Presentation .............................................................................................. vii

Poster Presentation .......................................................................................... viii

Academic Work ................................................................................................. viii

Table of Figures ................................................................................................. ix

Abbreviations ................................................................................................... xiv

Symbols ........................................................................................................... xvi

Table of Content .............................................................................................. xvii

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

2 Thin film solar cells: concept and materials ................................................. 5

2.1 CIGS thin film solar cell and module: materials and structure .................. 5

2.2 Organic solar cells .................................................................................... 8

2.3 Cell structures of thin film and organic solar cells .................................. 12

2.4 Device architecture of thin film solar modules ........................................ 15

2.5 Electrical description of a solar cell ........................................................ 17

3 Loss analysis of solar cells using different imaging techniques ................. 21

3.1 Radiative and non-radiative recombination in thin film solar cells .......... 22

3.2 Electrical influence of macroscopic defects in thin film solar cells

and modules ........................................................................................... 23

3.3 Power dissipation of a defective cell under the Voc-condition ................. 25

3.4 Defect imaging with lock-in based IR imaging techniques...................... 27

3.4.1 Defect characterization with dark lock-in thermography ................ 28

3.4.2 Illuminated lock-in thermography .................................................. 33

Table of Content

xviii

3.4.3 Loss analysis with EL imaging ...................................................... 38

4 Degradation and Stability of OPV solar cells ............................................. 45

4.1 Degradation of active layer .................................................................... 46

4.2 Degradation of electrodes ...................................................................... 48

4.2.1 Metal electrode .............................................................................. 48

4.2.2 Degradation of hole transport layer ............................................... 49

4.2.3 Degradation of electron transport layer ......................................... 52

5 Methods for device characterization .......................................................... 55

5.1 Lifetime evaluation of OPV devices ........................................................ 55

5.1.1 Sample holder for degradation cycles ........................................... 57

5.1.2 JV characterization ........................................................................ 59

5.1.3 Setup of photo-degradation ........................................................... 61

5.2 Voc-Measurement of thin film solar cells within a module ....................... 62

5.2.1 Cell voltage of modules with randomly distributed defects ............ 64

5.3 Lock-in Imaging ...................................................................................... 66

5.3.1 Theory of lock in based imaging .................................................... 66

5.3.2 Imaging setup ................................................................................ 69

5.3.3 IR camera ...................................................................................... 70

5.3.4 ELLI camera .................................................................................. 70

6 Photo-degradation of organic tandem solar cells ....................................... 71

6.1 Device fabrication and used materials ................................................... 72

6.1.1 Tandem cell layer deposition ......................................................... 74

6.1.2 Device Encapsulation .................................................................... 75

6.2 Photo-aging without UV light .................................................................. 75

6.2.1 Photo-aging of tandem cells .......................................................... 75

6.2.2 Darkaging of OPV tandem cells .................................................... 78

6.3 Photo-aging with UV lights soaking ........................................................ 79

6.4 Conclusion ............................................................................................. 84

Table of Content

xix

7 Temperature and moisture induced degradation of inverted and

organic P3HT:PCBM solar cells ............................................................... 85

7.1 Test conditions and device fabrication ................................................... 87

7.1.1 Organic solar cell preparation and characterization ...................... 87

7.1.2 Imaging parameters and settings .................................................. 88

7.2 Moisture induced degradation of inverted OPC devices ........................ 89

7.2.1 Heat- and damp heat-induced degradation ................................... 89

7.2.2 Determination of the activation energy .......................................... 93

7.2.3 IR imaging of moisture-induced degradation at 65°C/85%RH ....... 95

7.2.4 Moisture diffusion in encapsulated devices ................................... 99

7.3 Conclusion ........................................................................................... 102

8 The electrical and thermal characterization of macroscopic defects in

thin film solar modules ............................................................................ 105

8.1 Description of test modules .................................................................. 107

8.2 Influence of a defect on its surrounding cell measured by ILIT ............ 108

8.3 Relation between Voc,cell and IR- emission of a defect .......................... 112

8.4 Conclusion ........................................................................................... 116

9 Summary and Outlook ............................................................................. 119

9.1 Summary .............................................................................................. 119

9.2 Outlook ................................................................................................. 121

Appendix A ..................................................................................................... 123

Appendix B ..................................................................................................... 127

Appendix C ..................................................................................................... 138

Bibliography .................................................................................................... 141

Table of Content

xx

1. Introduction

1

1 Introduction

In the future, the major economies will face great problems in both energy

generation and supply. It is expected that current energy demand will increase

from 18 TW (2013) to 30 TW in 2050 [1] and this increase will not be covered

using only through conventional energy sources. The dominated use of fossil

fuels and nuclear elements within these sources causes critical issues which

affecting humans and our environment. The continuously increasing emission of

carbon dioxide (CO2) results in climate change while the problem of a safe

disposal of nuclear waste is still under discussion. In addition, the 2011

Fukushima disaster served as an impressive example of the very real risks to

both humans and the environment which stem from nuclear power sources.

Regarding these problems it is understandable that the call for clean and

unlimited (renewable) energy continuously rises. Amongst the currently

available renewable energy sources such as wind energy and hydropower,

solar energy bears the largest potential of future energy supply [1]. One way to

use the energy of the sun is the direct conversion of solar power into electricity

by using solar cells. Based on the photoelectric effect, firstly described 1839 by

Edmond Becquerel [2], the energy of light is absorbed by charge carriers which

are lifted to a higher energy level within the cell. By connecting the terminals of

the cell with an external load the light induced charge carriers contribute to a

photocurrent and power can be extracted from the cell.

Large efforts have been made to develop these cells into high

performance energy converters [3]. Now a day, different solar cell technologies

are available at the market while silicon (Si) based solar cells are the main

technology [4]. However, due to the need of electrical grade silicon and the

1. Introduction

2

huge material loss during cell fabrication, the production of Si solar modules is

both cost and energy intensive. To save costs and material, a new cell concept

was developed which based on the direct deposition of thin films with

thicknesses of some microns on rigid glass substrates [5–8]. During the last 15

years, the field of thin film solar cells (TFSC) has been highly active in

developing new cell concepts (e.g. single-junction and multi-junction cells) and

absorber materials [9]. Therefore, one needs to distinguish between both

inorganic and organic solar cells (OPVs). Especially inorganic solar cells based

on Copper Indium Gallium Selenide (CIGS) have been developed to highly

advanced solar cells which are processed in a sequence of different co-

evaporation and chemical bath deposition steps. These devices show among all

other technologies the currently highest values of efficiency for single cells

(21.7%) and modules (15%-17%) [10–12]. However, the use of cost-intensive

vacuum technology and the use of rare elements (e.g. indium, gallium etc.) limit

the production of low-cost solar modules.

Interestingly, in contrast to inorganic TFSCs, OPVs are made of cheap

semiconductor materials based on conjugated polymers which can be either

evaporated or processed from a solution to thin films. Due to the development

of new OPV materials and cell concepts the efficiency of these devices could be

continuously improved during the last decade. Current record efficiencies are in

a range of about 12% and prove that OPV devices are competitive to standard

thin film technologies such as amorphous silicon [9,12,13]. Furthermore, the

prospect of implementing fully solution based device fabrication makes this

solar technology uniquely suited for roll-to-roll printing applications with

potentially very low costs.

However, both thin film technologies show various problems with respect

to their fabrication and long-term stability. Due to the thin film deposition, these

cells and modules are very prone to production-related defects which limit an

optimal power extraction and life time. The electrical influence of defects mainly

depends on the size, type, position, and irradiation condition [14]. Especially,

under low light conditions (e.g. <100 W/m²) the electrical impact of a defect

increase and might cause a complete breakdown of an entire cell within the

module. The origin of defects can be manifold and is not exclusively related to

improper module fabrication. In particular, OPV devices suffer from a large

1. Introduction

3

variety of different intrinsic and extrinsic degradation phenomena which might

also result in the formation of defects. Several research groups have provided

important insight into different aspects affecting the aging process of OPV

devices [7,15–20]. As a result, multiple degradation possibilities have been

identified, which are related to the influences of ambient conditions

(temperature, moisture, oxygen), light exposure, and physical stress [7,17,21–

27].

In order to improve the efficiency and the reliability of TFSCs, it is crucial

to have a firm physical and electrical understanding of both macroscopic

defects and degradation phenomena. Besides the commonly used JV

characterization different imaging techniques such as electroluminescence (EL),

photoluminescence (PL), dark lock in thermography (DLIT), and illuminated

lock-in thermography (ILIT) have been utilized as powerful tools for delivering

information about localized power losses and recombination processes in

semiconducting films with up to 5 µm spatial resolution [28–32]. IR imaging

based on lock-in technology has been widely used as fast, non-destructive and

contactless characterization tool for all kinds of solar cell technologies [31,33–

36].

The experiments in this study are mainly focused on the characterization

and quantification of the long term stability and loss processes amongst

different thin film solar cell technologies. Using imaging techniques based on

lock-in several experiments were performed in order to investigate both aging

aspects in different OPV device systems and the electrical influence of

macroscopic defects on the cell performance in commercially available CIGS

solar modules.

1. Introduction

4

2. Thin film solar cells: concept and materials

5

2 Thin film solar cells: concept and

materials

The wide field of photovoltaic thin film technologies can be separated into

inorganic and organic solar cells. In contrast to Si solar cells, the great

advantage of these technologies is the monolithic deposition of different thin film

interlayers resulting in less use of material and energy during production as well

as the direct fabrication of large area modules with interconnected cells.

Modules based on inorganic thin film solar cells are already commercially

available while most of the OPV devices are in a pre-commercialization state. In

this chapter a theoretical background of the used materials, fabrication steps

and characterization of both CIGS and OPV device will be given.

2.1 CIGS thin film solar cell and module: materials and

structure

Among the different thin film technologies solar cells based on Copper

Indium Gallium Selenide (CIGS) semiconductor absorber layers provide with

21.7% the highest power conversion efficiencies (PCE) [9,11]. This high

performance might be explained by regarding the chemical structure of the

semiconductor absorber layer. CIGS (Cu(Inx,Ga1-x)Se2) is a p-type

semiconductor material and originate from an element compositions of CuInSe2


6

(copper indium selenide) and CuGaSe2 (copper gallium selenide)

semiconductors. Both semiconductor materials have a polycrystalline

chalcopyrite lattice structure (Figure 2.1), a diamond structure similar to the

sphalerite structure with an order substitution of Cu and In or Ga atoms on the

zinc sites of sphalerite.

Figure 2.1: Crystalline lattice structure of a CIGS chalcopyrite unit cell. Image is reproduced with permission from [37]

Due to the substitution of Ga atoms on the In site, the band gap of CuInSe2

(1,04 eV) can be shifted to 1.7 eV (CuGaSe2) [38]. At a ratio of about 30%-40%

for the Ga/(In+Ga) concentration the band gap is shifted to around 1.2 eV,

which provides the highest PCE for CIGS solar cells [39,40]. The semiconductor

material has a direct band gap and a relatively high absorption coefficient of

approximately 105 cm-1 for photons with energies higher than 1.5 eV [41]. This

allows the deposition of polycrystalline and thin film semiconductor absorber

layer with layer thicknesses around 2 µm. Due to the low film thicknesses the

CIGS absorber layer can be deposited on both rigid (glass) and flexible

substrates (e.g. metal foils, polymer foils) [8,39,42]. However, the use of rear

elements such as In and Ga, mainly limits the production low-cost CIGS

semiconductor absorber layers. Hence current investigations also focus on the

substitution of In and Ga with low-cost elements such as zinc (Zn) and tin (Sn)


7

[43]. Wang et al. showed, that cells comprising a Cu2ZnSn(S,Se) (CZTS)

semiconductor absorber layer is able reach PCEs above 12% [44].

As can be seen in Figure 2.2, high efficient CIGS solar cells are made of

a stack of several thin film interlayers fabricated by a sequence of different

processing steps such as co-evaporation and chemical bath deposition [38].

Figure 2.2: a) Optical band structure of an irradiated CIGS solar cell under short circuit condition. b) Cross section of a CIGS solar cell investigated with a scanning electron microscope. In order to reduce optical losses due to reflection a thin layer of magnesium fluoride (MgF2) was deposited in top of the ZnO window layer. Image is reproduced with permission from [39]

A standard CIGS solar cell consists of a high conductive molybdenum

(Mo) layer (~0,2Ω/) [14], sputtered on a soda-lime glass substrate. On top of

the Mo layer the photovoltaic CIGS absorber layer is deposited with common

thicknesses around ~2 µm [38]. During the CIGS deposition, a thin MoSe2

interlayer is formed at the active layer/electrode interface inducing in an

schottky-type contact formation and an improvement of the hole extraction [40].

The p-n junction is built between the p-type CIGS semiconductor absorber layer


8

and n-type cadmium sulfide (CdS) and zinc oxide (ZnO) window layers. Here,

the surface properties of the CIGS are important for a high efficient cell, since

the surface becomes the active interface of the compete device. Interestingly, a

widen of the band gap with respect to the bulk of the absorber material close to

the CIGS surface was observed [45,46]. This widening originates from a Cu-

poorness at the CIGS/CdS interface leading to a decrease of the valence band

and a reduction of interface recombination processes of charge carriers [47–49].

On top of the CIGS surface a buffer layer of CdS with a thickness of

about 50 nm is deposited. As can be seen in Figure 2.2, the wide band gap of

the CdS layer (~ΔE=2,4 eV) induces a conduction band offset ΔEc which is

needed to reduce surface recombination at the CIGS/ZnO interface and to

improve charge carrier extraction [50]. Typical offset energies are in the range

of ~0.3 eV and can slightly differ by the material composition of CdS [38].

However, Cd belongs to the toxic elements and current investigations focus on

the replacement of the CdS layer with alternative materials. Different material

compositions such as In(OH,S), Zn(OH,S) ZnSe, and ZnS in combination with

standard ZnO are tested as alternative buffer layer while cells comprising ZnS

show with 18% PCE the most promising results [51,52]. A further advantage

arises from higher band gap energies of about 3.6 eV leading to less parasitic

absorption of light and an improved transmission of light into the CIGS absorber

layer. However, the layer structure presented in Figure 2.2 can be divided into a

space charge region (SCR) and a quasi-neutral region (QNR). Due to the large

band gap of the ZnO window layer (~ΔE=3,2 eV) the majority of the incident

photons is absorbed by charge carriers in the first µm of the absorber layer. The

light induced electron-hole pairs are separated by the internal electrical field of

the p-n junction and contribute to the photocurrent of the cell.

2.2 Organic solar cells

Solar cells based on organic semiconductor materials in combination

with low cost, light weight and flexible substrates bear the potential of a low-cost

mass production. Organic absorber layers are made by semiconducting

polymers based on the conjugation of carbon (C) atoms. The electrical ground


9

state of a C atom is a 1s²2s²2p² configuration. Four electrons are in the outer

electronic energy level and determine the chemical reaction. The two s

electrons are paired while the two p electrons are unpaired. In organic

semiconductors the s and p orbitals form three sp² orbitals (sp² hybridization)

the so called σ-bonds. As can be seen in Figure 2.3 three sp² orbitals are

coplanar and directed of about 120° apart from each other. The remaining

fourth orbital, pz, stays unaltered and is perpendicular to the plane of the other

sp² orbitals. If the space between the carbon atoms is close enough, the overlap

of the pz orbitals lead to a delocalization of the charges. Due to the Peierls

instability two delocalized energy bands are created, the bonding and anti-

bonding π and π* orbital, also called as highest occupied molecular orbital

(HOMO) and lowest occupied molecular orbital (LUMO), respectively. These

orbitals can be seen as both charge and conduction band which are separated

by an gap energy in a range of several eV [53].

Figure 2.3: 2 carbon atoms and 4 hydrogen (H) atom from an ethane molecule. The σ bonds (green) are formed due to the overlap of the sp² orbitals of the C atoms and the s-orbital of the H atom. The remaining π bounds are formed by the remaining pz-orbitals of the C atom. Image is reproduced and rearranged from [54]

In contrast to crystalline Si semiconductors, OPV semiconductors show a

high degree of disorder of electric states resulting in no formation of a classic

energy band structure [55]. The transition between the LUMO and HOMO is


10

induced by light and the absorption of its energy by charge carriers in the

ground state.

State of the art organic solar cell absorber layer consists of an intermixed

structure of donor and acceptor polymers which can be described as a bulk

heterojunction (BHJ). Due to the intermixed structure a large donor and

acceptor interface is formed. Typical domain sizes range from 10 nm up to and

30 nm [56]. Therefore the used acceptor and donor materials need to fulfill

different requirements:

a simple processing condition in terms of solubility

a “good” alignment of the acceptor material’s LUMO level with respect to

the donor material

the donor material should provide a low ionization potential

the acceptor material should provide a high electron affinity

the ability of intermixing both materials to form a large interface nano-

structure

a) b)

Figure 2.4: Chemical structure of two common used donor and acceptor materials used to form a BHJ absorber layer of an organic solar cell. a) Structure of P3HT, a p doped semiconductor polymer with Δ Eg between 1.8 eV-2.0 eV and a hole mobility between µh=0.1 cm²/Vs -0.3 cm²/Vs in a regular undoped state. b) Chemical structure of PCBM, a fullerene derivate consisting of a C60 Buckminsterfullerene (Bucky-ball) and an additional side chain to enable solution processing. The optical band gap is about 1.8 eV.

Two widely studied representatives who fulfill these requirements are P3HT

(Poly(3-hexylthiophen-2,5-diyl)) as donor material and PCBM (phenyl-C60-

butyric acid methyl ester) as acceptor material. The chemical structure can be


11

seen in Figure 2.4. In general the two materials are mixed together to a blend

which is deposited either on rigid or flexible substrates via different coating

techniques. After different drying and post processing steps an interpenetrating

network of both donor and acceptor material is formed. The large interface of

the BHJ allows an charge carrier separation while the interconnected domains

enable an efficient charge carrier transport to the respective electrodes (Figure

2.5).

Figure 2.5: Illustration of the different physical processes which take place inside an organic and bulk hetero junction solar cell under irradiation. (left) the device kinetics from the charge carrier generation (I) till the charge carrier extraction (VI) is illustrated. (right) Illustration of a simplified band diagram related to the device kinetics (left).

The charge carrier generation and extraction can be separated in six

single steps and is illustrated in Figure 2.5. The energy of incident light is

absorbed by charge carriers under the formation of a strongly bound Franklin

excitons or singlet excitons in the donor material (I). Typical binding energies

are between 0.3 eV and 1.4 eV [53,55]. After creation, the excitons diffuse

within the donor material towards donor acceptor interface where the

dissociation of the excitons takes place (II-III). The diffusion length of the

excitons is rather short (~10nm) [57] and recombination may occur if the

excitons are generated too far away from the donor acceptor interface. From

the band diagram it can be seen that the acceptor material has a large electron

affinity while the donor material has the larger ionization potential. The


12

difference in electron affinities induces a driving force at the donor acceptor

interface which is strong enough to split the excitons (IV). The electrons are

transferred from the LUMO of the donor to the more electro negative LUMO of

the acceptor molecule. The positive charge carriers remain at the donor

material. The final steps refer to charge carrier transport inside the bulk material

(V) and charge extraction at the electrodes (VI).

2.3 Cell structures of thin film and organic solar cells

Organic solar cells are made of a stack of different interlayers,

sandwiched between two electrodes. Most of the interlayer materials base on

soluble components and can be therefore deposited by doctor blading, spin

coating, and slot-die coating. Two different device geometries are commonly

used which are referred to as “normal” and “inverted” and can be seen in Figure

2.6.

Figure 2.6: a) Illustration of cell stack configurations used for single junction OPV devices. (a) “normal” structure (b) “inverted” structure

On top of the substrate a transparent conductive oxide (TCO) is deposit

which allows the transmission of visible light into the semiconductor absorber

layer where the charge carrier generation takes place. A common used window

layer material is indium tin oxide (ITO), which base on a wide band gap

semiconductor with a band gap larger 3 eV [58]. However, ITO is rather brittle

and tends to crack formation at thicker film thicknesses when it is deposited on


13

flexible substrates [59]. To overcome this problem several material systems are

reported which have the potential to replace ITO as window layer [60–62]. An

interesting approach is the use of high conductive silver nanowire firstly

presented by Krantz and co-workers. [63]. For an efficient charge extraction

different buffer layers are used to improve the charge carrier transport with in

the device [64]. Therefore the buffer layers have to ensure the alignment of the

quasi-Fermi level of the absorber layer materials with the quasi-Fermi level of

the respective electrodes. Depending on the position in the cell stack these

buffer layer allow either the transport of positive or negative charge carriers. In

case of a normal device geometry a hole transport layer (HTL) is introduced

between the active layer and the ITO electrode (see Figure 2.6). A common

used material is PEDOT:PSS (poly(3,4-ethylendioxythiophene):poly(styrene-

sulfonate)) which can be processed in a water based solution and deposited on

top of the ITO surface. PEDOT:PSS is a hole conducting metal like polymer

which lowers the intrinsic barrier between the ITO work function and the HOMO

level of the active layer. Furthermore it hinders the extraction of electrons on the

ITO electrode and allows smoothing of the ITO surface resulting in a reduction

of shunting and interface recombination [65,66]. However, PEDOT:PSS is

mostly dissolved in a water based solution which might lead to water induced

degradation [67,68]. After cell fabrication rest water might diffuse within the cell

and react with the other layer materials. To overcome this limitation several new

and water free materials are deloped such as WOx (wolfram oxide) and MoOx

(molybdenum oxide) which can be used to replace PEDOT:PSS as HLT [69,70].

As active layer an organic semiconductor material is used in which light

is absorbed and the electron hole pair generation takes place. In case of a

normal device geometry a low work function metal is commonly used as top

electrode. This electrode serves as electron injection and extraction layer and

mostly consists of a two layer configuration made of LiF/Ca (lithium

fluoride / calcium) or Ca/Al (calcium / aluminum). However Ca is highly reactive

with oxygen and water and tends to corrosion resulting in strong reduction of

cell efficiency [53,71]. Improvement, in terms of stability, is achieved with the

development of the inverted device geometry and the use of more nobel metal

electrodes such as silver (Ag) and gold. In the inverted device geometry the ITO

electrode serves as electron injection layer while the metal electrode serves as


14

hole injection layer. The HTL is located between the metal electrode and the

active layer. To improve the electron extraction from the ITO side an electron

transport layer (ETL) is introduced between the ITO electrode and active layer.

Similar to the HTL, the ETL serves as barrier potential which suppresses

unwanted recombination processes of positive charge carrier at the ITO contact.

Typical used materials are zinc oxide (ZnO) and aluminum doped ZnO (AZO)

[72,73].

Figure 2.7: Illustration of cell stacks used for multifunction OPV Devices with top and bottom cells connected in series. (a) “normal” structure (b) “inverted” structure

Besides commonly used single junction solar cells, double junction

(tandem) or multi junction solar cells gain more scientific interest [74–76]. Here,

two or more single sub-cells with complementary absorption spectra are

stacked together and connected by either semi-transparent recombination

layers (series connection) or semi-transparent electrodes (parallel connection)

[76,77]. In particular, the combination of two or more active layers allows the

enlargement of the spectral sensitivity of a single cell resulting in an

improvement of the absorption of light. Since a wider spectral range of the

incoming light can be converted into free charge carriers, tandem cells show a

higher efficiency as their single junction counterparts. According to the

efficiency prediction of in series connected organic tandem cells presented by

Dennler and co-workers, an maximal improvement by 40% is achievable

compared to the optimized single junction solar cells [78]. Similar improvement

may also be achieved for parallel connected tandem cells. However, due to the


15

difficulties in realizing highly conductive semi-transparent electrodes, the current

efficiencies of tandem solar cells with parallel-connected sub-cells legs far

behind both state-of–the-art single junction and series-connected tandem solar

cells [76].

Until now, the most efficient tandem solar cells are based on the series-

connection mode, in which the different sub-cells are connected by a

recombination layer (intermediate layer). The highest reported PCE of an

organic tandem solar cell is 12%. In this case the active layers of the tandem

cell was fabricated by thermal evaporation of small molecules [13]. For solution-

processed tandem devices an efficiency of about 10.6% is achieved [75]. It is

important to note that the performance of a tandem solar cell depends on the

active materials of the sub-cells. Along with the continuous development and

design of efficient low band gab materials during the last three years, the PCE

of organic solar cells were rapidly improved from 8% to 10% [75,79–83].

2.4 Device architecture of thin film solar modules

Since the power generation of thin film solar cells is limited by sheet

resistances, several cells need to be connected in series to form highly efficient

power generator. Hence, the top electrode of one cell is connected to the

bottom electrode of an adjacent cell. An illustration of a schematic thin film solar

module with a cell-to-cell series connection can be seen in Figure 2.8.


16

Figure 2.8: Schematic module layout with monolithic interconnection of cells illustrated as top view (a) and cross section (b). Dimensions are not in scale. The monolithic interconnection is realized by three patterning lines labeled as P1, P2, and P3.

Similar to the preparation of individual cells, the production of a thin film

module consists of a sequence of different deposition and patterning processes.

The interconnection of two adjacent cells is realized with three patterning lines

labeled with P1, P2, and P3. After the deposition of the back contact material,

the P1 line guarantees that the back contacts of the individual cells are

electrically isolated from each other. A second line (P2) is introduced after the

deposition of the absorber material and allows the interconnection of two

adjacent cells. Finally, the P3 line enables the electrical isolation of the front

contacts.

Depending on the cell technology different pattering techniques are used.

In case of CIGS modules the interconnection is realized either with laser

patterning or with a sequence of laser patterning and mechanical scribing

[38,84–86]. In case of organic solar modules different patterning strategies have

been proposed. Commonly used techniques are shifted layer deposition or

screen printing [87], spatial resolved ink-jet printing [88], photolithography [89],

soft lithography [90], and advanced laser pattering [91,92]. In particular, laser

patterning enables a strong reduction of the dead area (see Figure 2.8),

resulting in improved geometric fill factors of about 95% [93]. A further and also


17

interesting approach for the interconnection of individual cells originates by the

combination of laser patterning and ink-jet printing [94]. In contrast to

conventional structuring steps, this combination allows the series connection

with only one patterning step. Here, the different cell layers are deposited on a

large-area substrate. After deposition the three pattering lines are introduced

with a laser. The individual cells are electrically isolated by printing a dielectric

ink into the P1 Line. A highly conductive ink is printed over the P1 line and into

the P2 line and enables across the dielectric ink the series connection of the

adjacent cells.

2.5 Electrical description of a solar cell

All processes which contribute to the energy conversion within the solar

cells can be incorporated into a simple one diode model. This model is based

on the equivalent circuit illustrated in Figure 2.9 and contains a diode, a current

source, a series resistor with the resistance Rs, and a parallel resistor with

resistance Rp.

Figure 2.9: Replacement circuit of the one diode model

From the equivalent circuit the JV characteristics of a cell can be

expressed by the Shockley equation [95,96] and is given as

𝐽 = 𝐽0 (𝑒𝑥𝑝 (𝑞(𝑉+𝐽𝑅𝑆)

𝑛𝑘𝐵𝑇) − 1) − (

𝑉+𝐽𝑅𝑠

𝑅𝑝) − 𝐽𝑝ℎ Eq. 1

where J is the current density flowing through the cell, q the elementary charge,

V the applied voltage, kB the Boltzmann constant, n the ideality factor, T the


18

temperature, Jph the photocurrent, and I0 the saturation current of the device.

The term JRs accounts to the voltage losses due to current flow through all

series resistances while the term (V+JRs)/Rp is attributed to the current flow

through parallel resistance of the cell. Rs is primarily governed by all ohmic

resistances of the cell such as sheet resistances, contact resistances, and bulk

resistances. The series resistance is demanded to be as minimal as possible to

obtain high efficient solar cells and to avoid losses of the photocurrent. The

parallel resistance induces unwanted leakage currents of the diode and needs

to be as high as possible. Due to light absorption Jph is genarted and shifts the

dark JV-characteristics to negative values for lower voltages.

A dark and irradiated JV characteristic of an OPV device can be seen in

Figure 2.10. In general solar cells are operated between OC (open circuit) and

SC (short circuit) condition.

Figure 2.10: IV characteristics of a dark and irradiated organic solar cell solar cell

From the irradiated IV characteristic different cell parameters can be

extracted. The overall efficiency is represented by the PCE (power conversion

efficiency) and given by:

𝑃𝐶𝐸 =𝑉𝑜𝑐∙𝐽𝑠𝑐∙𝐹𝐹

𝑃𝑙𝑖𝑔ℎ𝑡 Eq. 2


19

with Voc as the voltage at open circuit conditions (J(V = Voc)=Jdark – Jph = 0), Jsc

as the current at short circuit conditions (Jsc=J(V=0)), and Plight as the power of

the incident light. The Fill Factor (FF) is then given by:

𝐹𝐹 =𝑃𝑚𝑝𝑝

𝑉𝑜𝑐∙𝐽𝑠𝑐 Eq. 3

where Pmpp represents the power at the maximum power point (mpp) which is

defined as the product of the current and voltage at the mpp. The JV-

characteristic can be separated into three parts. The first part (shunt regime)

lies between -0.3 V and 0.3 V and is mainly dominated by Rp. The second

regime (diode regime) between 0.3 V and 0.8 V can be attributed to the diode

with a proportional slope to the diodes ideality factor n. The third regime for

voltages >0.6 V is the current injection regime. This part is mainly limited by

series resistances. The values of Rp and Rs can be determined from the dark IV

characteristic by calculating the inverse slope at V = 0 for Rp and >0.8V for Rs.


20

3. Loss analysis of solar cells using different imaging techniques

21

3 Loss analysis of solar cells using

different imaging techniques

Thin film solar cells are made of large area semiconductors (typically

several square centimeters) which are optimized for light absorption and charge

carrier collection. However, these devices suffer from different loss processes

which limit the cell with respect to both device performance and stability. An

advanced characterization of spatial resolved loss mechanisms is therefore

essential to understand recombination processes of charge carriers and

degradation phenomena. Interestingly, if the operation mode of a cell is inverted

(e.g. injection of current) different loss processes can be triggered and the

injected charge carriers recombine either by the emission of photons (light) or

by the transmission of phonons to the lattice (heat). With the development of bi-

dimensional and focal plane array (FPA) detectors, this signals can be detected

and converted into a spatially resolved image [97].

Infrared (IR) imaging techniques and near IR (NIR) imaging techniques

such as dark lock-in thermography (DLIT), illuminated lock-in thermography

(ILIT), electroluminescence lock-in (ELLI), and photoluminescence (PL) imaging

are particularly suitable to provide a complementary insight of both localized

power losses and recombination mechanisms. Especially IR imaging based on

the lock-in technology has been widely explored for silicon cells as a fast, non-

destructive, and contactless characterization tool for macroscopic defects (e.g.,

dark spots, electrode defects, shunts), changes in sheet resistance and for

visualizing short circuit current distributions.


22

3.1 Radiative and non-radiative recombination in thin

film solar cells

Thin film solar cells are made of micro-crystalline, semi-crystalline or

amorphous semiconductor materials with a high degree of disorder of electric

states. This leads to different loss processes within the cell which limit an

optimal formation of Voc, Jsc, and PCE. During the operation mode of the cell,

light is absorbed by charge carriers (e.g. electrons) which are lifted into a higher

energy level. If no charge carriers are extracted by an external load, the excess

energy can be released either by the emission of photons (radiative) or the

transmission of phonons (non-radiative). Regarding the literature different loss

processes are discussed [98–106] in thin film solar cells and can be classified

into three basic recombination processes (see Figure 3.1).

Figure 3.1: Generation and recombination processes of electron hole pairs in thin film solar cells under the Voc condition

Radiative recombination represents the inverse process of optical absorption.

Here, the electron-hole pair recombines directly by the emission of a photon

with the energy equal to Eg. Depending on the excitation, this recombination

process can be trigger either with the injection of current (electroluminescence)

or with the irradiation of light (photoluminescence). For light energies higher

than Eg, the charge carrier thermalizes rapidly by energy transfer in form of

lattice vibration (e.g transmission of phonons) to a metastable state (Ec). In case

of Auger recombination, a third charge carrier might be accelerated by the

absorption of a photon which rapidly thermalize to the charge band. The third


23

process refers to trap assisted recombination. In contrast to band-to-band

recombination (radiative), the charge carriers recombine via defect related

energy states (trap) within the band gap. Similar to thermalization, the excess

energy is also transmitted by phonons, resulting in an increasing cell

temperature and the emission of heat. Trap assisted recombination can take

place at different sites and can occur as interface recombination, space charge

recombination, neutral bulk recombination and back contact recombination

[46,107,108]. In particular, thin film solar cells based on organic semiconductor

materials show due to their semi-crystallinity a high degree of disorder of

electric states. This results in a high density of trap states from the beginning

which can increase if the cell is exposed to oxygen and water (see chapter 4.1)

[18,26,109].

3.2 Electrical influence of macroscopic defects in thin

film solar cells and modules

Besides intrinsic losses due to unwanted recombination processes

(section 3.1), macroscopic defects have the largest impact on cell performance

and stability [35,110,111]. The origin of these defects can be manifold but is

mainly related to both improper layer deposition and device fabrication or

intrinsic material diffusions [112–115]. During the last decade, different types of

defects in solar cells such as bulk/shunt defects [111,113,115,116], interface

defects [26,117,118], and interconnection defects [116,119] have been

identified and were intensively discussed in the literature. The different defect

classes can be seen in Figure 3.2.

Figure 3.2: Simplified illustrations of the different basic defect classes.


24

Interestingly, the influence on the surrounding cell is strongly related to

defect type, position, and irradiation condition. Due to the low defect resistance,

a potential gradient between the two electrodes is induced which results in an

inhomogeneous Voc distribution of the cell [120]. To gain a better understanding,

several electrical simulations provide a complementary insight into local cell

parameters and discuss the influence of a defect on the cell [14,121–124]. A

simulated local voltage distribution of a single cell in a thin film solar module

(CIGS) with an embedded low ohmic defect can be seen in Figure 3.3. The

defect was placed in the center of the cell and is represented by a low ohmic

resistor (see Figure 3.3 b). Regarding more details of the simulation model

please refer to Fecher et al. [14]

Figure 3.3: (a) Simulated distribution of the local electrical potential of the electrodes layers of a shunted CIGS solar cell under the Voc condition. (b) Line scan through the electrical potential (black line in (a)). The equivalent circuit indicates the current (black arrow) and the drop of the electrical potential due to the resistive electrode. Image is reproduced with permission from [14]

The local voltage distribution Vi of an irradiated cell under the Voc condition

(Figure 3.3) can be described as

𝑉𝑖(𝑥, 𝑦) = (Φ𝑇𝐶𝑂(𝑥, 𝑦) − Φ𝑀𝑜(𝑥, 𝑦)) Eq. 4

with Vi as junction voltage, Φ as the electrical potential distribution of the TCO

electrode and of the Mo back electrode, and x,y as position parameters. Due to

the low defect resistance, Vi is minimal in the center of the cell. In this case the

defect acts as an internal load and attracts light induced charge carriers from

the surrounding cell. This in combination with the local sheet resistances of the


25

electrodes induces a potential gradient leading to an inhomogeneous voltage

distribution of Vi. Cell parts around the defect show a lower Vi (0<Vi<Voc) than

the local Vi of non-disturbed cell parts. As can be seen in Figure 3.3, these cell

parts act as a current sources and contribute to a vertical cell current Iperp(x,y,Vi)

which is injected into the Mo electrode. Due to the inhomogeneous potential of

the electrodes, a good portion of the cell current flows towards the defect,

passes it, and returns to its origin. Since the potential gradient in the electrodes

is strongly related to the sheet resistances of the electrodes, its deformation is

limited and depends on the irradiation conditions [125]. Due to joule heating the

defect temperature rises in contrast to its surrounding, resulting in detectable

heat dissipation. Thermal investigations and simulations of defects in thin film

solar modules reveal a temperature contrast of several degree Celsius

[126,127]. Depending on the heat development the encapsulation might be

damaged leading to a reduced cell or module stability.

3.3 Power dissipation of a defective cell under the Voc-

condition

The explanations given in section 3.1 and 3.2 can be converted into a thermal

model of power dissipation. For a detailed description of the different thermal

loss processes within a cell, see references [123,128]. Breitenstein,

Langenkamp and Warte [123] define the power dissipation of an ideal solar cell

(perfectly homogeneous material) under irradiation as:

𝑃𝑖𝑙𝑙 = 𝐽𝑑𝑖𝑜𝑑𝑒𝑉 +𝐽𝑝ℎ

𝑒ℎ𝜈 − 𝐽𝑝ℎ𝑉 Eq. 5

where Jdiode represents the diode current density, Jph the photocurrent density, e

the elementary charge, h Planck’s constant, ν the frequency of the absorbed

incident light, and V the applied voltage. Equation 5 consists of three terms: the

first term describes the power dissipation of the cell in darkness, the second the

energy induced by the absorption of the irradiated photons, and the third the

power induced by the photocurrent. It is assumed that all irradiated photons are

absorbed and contribute to a photocurrent. In extreme cases, given either the

*Section adopted with permission from ref [129] (Copyright Elsevier)


26

Jsc-condition where V=0 and J=Jph or the Voc-condition where V=Voc and

J=Jdiode-Jph=0, the energy of the absorbed photons is mostly converted into heat

and the dissipated power is due to thermalization (second term). Further details

can be found in [110,123,128]. Based on the simulation model used for Figure

3.3, equation 5, and the energy conversion (Pill=Pelectric) the thermal dissipated

power can be expressed as the local electrical power density (Figure 3.4).

Figure 3.4: Simulated local electrical power density of a CIGS cell with a defect as a function of cell position. Image is reproduced from [129] (with permission of Elsevier)

For the simulations it should be mentioned that the second term of

equation 5 mainly represents heat dissipation due to thermalization and was set

to zero. Negative power densities represent current flow towards the defect

(power extraction of non-defective parts) while positive values represent power

dissipation [121,123]. As expected, a distinct maximum of the electrical power

density can be found around the defect [35]. Charge carriers of the non-

defective cell parts do not recombine and become attracted by the lower

resistance found in the defect. Hence, cell parts close to the defect provide

more photocurrent to the defect than those further away. The result is a

negative local electrical power density of non-defective cell parts. As shown in

Figure 3.4, at the left and right cell edges the local electrical power density

flattens out to a constant negative local value. This is due to the boundary

condition that the cell edges are isolated and no current density is assumed,


27

resulting in a voltage drop and a negative local electrical power density at the

cell edge.

3.4 Defect imaging with lock-in based IR imaging

techniques

The characterization of PV devices by measuring the JV characteristics

belongs to the fastest characterization method. Valuable information about the

electrical properties such as Voc, Jsc, FF, and PCE can be both extracted and

used to verify the quality of the investigated device. However, the mentioned

parameters only obtain global information of the device performance while local

variations are neglected. In order to gain a better understanding of the

microscopic processes limiting PV device performance with the aim of

improving the operating lifetime or the characterization of macroscopic defects,

it is vital to acquire local information on solar cell specific parameters. Infrared

(IR) imaging techniques and near IR (NIR) imaging techniques such as DLIT,

ILIT, and EL are particularly suitable for delivering information about localized

power losses and recombination processes in semiconducting films with up to

4 m spatial resolution [33,35,36,110,117,130–132]. Both imaging techniques

use high sensitive NIR/IR cameras equipped with focal plan array (FPA)

quantum detectors. Depending on the atmospheric window the spectral

response of the detectors can range between 0.8 µm-1.7 µm (indium gallium

arsenide, InGaAs) and 1.5 µm-5 µm (indium antimonite InSb). Common noise

equivalent temperature differences (NETD) of these detectors are in the range

of 20 mK with relatively high frame rates between 100 Hz and 300 Hz. With the

application of different pulsed trigger signals and detection schemes using

highly sensitive NIR and IR cameras, these techniques can provide

complementary insight into local aging mechanisms. IR imaging based on lock-

in technology has been widely explored in the case of inorganic semiconductor

technologies such as crystalline silicon cells and CIGS solar modules as a fast,

non-invasive, and contactless characterization tool [31,33–36]. Recently, these

techniques have been successfully applied to organic solar cells and modules in

order to characterize macroscopic defects (e.g., dark spots, electrode defects,


28

shunts), changes in sheet resistance and visualizing short circuit current

distributions [32,113,114,122,133–135].

3.4.1 Defect characterization with dark lock-in thermography

Dark lock-in thermography (DLIT) is mainly used to detect local heat

distributions which are related to the dark JV characteristic within electronic

circuits. For a measurement a sample is periodically exited while the emitted IR

radiation is detected with a highly advanced IR camera and digitally processed

according to the lock-in principle. Pioneers of DLIT were Kuo and co-workers

who combined as one of the first groups the principles of thermal wave imaging

(TWI) with a IR video camera in order to investigate thermal losses in liquid

crystal displays [136]. The IR camera was equipped with a bi-dimensional FPA

detector which provided a certain frame rate. Each pixel of the measured IR

images was digitally processed according to the lock-in principle, resulting in a

strong noise reduction and an improvement thermal sensitivity.

Since the introduction of DLIT imaging, this technique has been used not

only for the characterization of integrated circuits, but also for the

characterization of solids materials such as ceramics, polymers, and coatings

[137–139]. As one of the first groups, Breitenstein et al. as well as Padinger et

al. used DLIT measurements, independently from each other, to characterize

hot spots in Si solar cells and organic solar cells, respectively [140,141].

However, Breitenstein and co-workers showed in several publications different

applications of DLIT for the characterization of local power losses in Si solar

cells.

One of the main applications of DLIT imaging is the characterization of

unwanted leakage currents within a cell. Therefore, Figure 3.5 shows two DLIT

images of a multicrystalline Si solar cell, measured at different bias (-0.5 V and

+0.5 V). The relatively narrow voltage range was chosen to neglect implication

due to series resistance. Joule heating resulting in leakage current causes a hot

spot (defect) and represents an alternative current path in the cell. With respect

to the appearance of hot spots at both DLIT measurements, one can distinguish

between both ohmic and non-ohmic leakage current.


29

Figure 3.5: DLIT images of a multicrystalline solar cell measured at (a) -0.5 V, (b) +0.5 V Image a and b are scaled to 0 mK (black) to 5 mK (white). Image is reproduced with permission from [111].

From the presented DLIT images (Figure 3.5) it can be seen that 5 hot

spots (white arrows) show an ohmic behavior while 4 hot spot (red arrows) do

show a non-ohmic characteristic. Especially non-ohmic hot spots can be

interpreted as recombination sites where the pn-junction of the cell crosses the

surface [30,142]. Interestingly, in an intensive study by Breitenstein et al., it

could be proved that the majority of hot spots are mainly due to improper cell

fabrication and do show an non-ohmic behavior [111,143]. The formation of

intrinsic hot spots in Si solar cells can be seen as a rare case [111]. However,

the majority of the injection current passes under dark condition through hot

spots. This represents one of the main reasons of a reduced device efficiency

[30,110,128].

Besides the characterization of defect induced leaked currents within the

cell, also quantitative values about several diode parameters such as ideality

factor and dark saturation current can be extracted from DLIT experiments

[111,123,128,131,141]. To do so, Breitenstein and co-workers defined the local

power dissipation Ploc of a solar cell in the dark as:

𝑃𝑙𝑜𝑐(𝑥, 𝑦) =𝑆𝐷𝐿𝐼𝑇(𝑥,𝑦)𝑃𝑡𝑜𝑡

∫ 𝑆𝐷𝐿𝐼𝑇(𝑥,𝑦)𝑑𝑥𝑑𝑦 Eq. 6


30

with SDLIT as the camera signal and Ptot as the applied electrical power.

Based on Eq. 6, all important diode parameters such as local recombination

currents, drift currents, and ideality factors can be investigated (see Figure 3.6).

Figure 3.6: DLIT images of the local distributed current density (a), diffusion current density (b), recombination current density, and ideality factor n (d) of a multicristalline solar cell measured at 0.55 V. The images (a), (b), and (c) are represented in the same scale while (d) is scaled from 1.5 to 5. The image is reproduced with permission from [144]

Breitenstein showed that the diffusion current is mostly governed by the

lifetime of the charge carrier in the bulk, while the recombination current and the

ideality factor of the diode is governed by the recombination of charge carriers

in the depletion region [144]. Hence, the defects A, B, C, and D, presented in

Figure 3.6 a), could be attributed to different recombination sites in the

semiconductor. It can be seen that defect A,C, and D mainly influence the

recombination processes in the depletion region while B influences bulk

recombination. The origins of these recombination processes are mostly due to

locally extended defects leading to a high local density of recombination states

[144,145]. Interestingly, the ideality factors of the presented defects was


31

measured to be 2.5 and higher which indicates a large number of defect states

in the semiconductor material [145]. Based on the investigations of Breitenstein,

Rißland et al. revealed that the electric properties of multicrystalline Si solar

cells are mainly dominated by low-lifetime defect regions which contain

recombination-active grain boundaries and dislocations [146]. With a

combination of both EL and DLIT imaging it could be shown that these regions

increase the dark saturation current. The scientific group around Otwin

Breitenstein impressively presented different applications of DLIT imaging

[128,146,147]. Besides the characterization of several defects and intrinsic

losses in solar cells, Ramspeck and co-workers demonstrated that DLIT

imaging in combination with EL imaging can be used to visualize local series

resistance distributions [148] (see Figure 3.13). Therefore, the local Rs-

distribution was defined as described in Eq. 8 (see section 3.4.3). With the EL

measurements the local voltage distribute of a cell was recalculated while the

local injection current was measured with the method described in [144].

However, since both imaging techniques required different camera systems, the

presented method is limited in terms of resolution and measure time. A further

development was presented by Chung and co-workers. The group

demonstrated by using a thermal approach that one DLIT measurement is

sufficient to recalculate both local series and shunt resistance distributions in Si

solar cells [149,150].

Regarding OPV devices, Bachman and co-workers investigated the local

power dissipation of organic solar cells in dependency of different applied

biases. It was found that, due to the sheet resistance differences of the front

and back contact, an inhomogeneous current distribution was injected resulting

in an inhomogeneous heat distribution over the cell length [133]. A similar

observation was made by Hoppe and co-workers who investigated thermal

losses in organic thin film solar modules [114]. These modules consisted of

three cells with a normal device geometry, connected in series, and comprised

of P3HT:PCBM active layer, PEDOT:PSS, and an Al electrode.


32

Figure 3.7: a) Photograph of organic thin film solar modules with three in series connected cells. Each cell had an active area of 5.4 cm² while the dead area was 0.45 cm². b) DLIT Image measured with a pulsed voltage between 0 V and 2.4 V. Image is reproduced with permission from [114]

From the DLIT measurement presented in Figure 3.7 different thermal features

can be seen. The relatively high applied forward bias was chosen to simulate

the Jsc distribution of the module under the dark condition. Several kinds of

defects could be related to edge shunts (rectangles) and leakage currents

between the two electrodes (diamonds). Furthermore, it could be proved that

DLIT imaging is sensitive enough to detect inhomogeneous heat dissipations

due to different film thicknesses (circles) in the cells. Similar to the observation

of Bachmann et al. the temperature gradient from the right edge to the left edge

of each cell was attributed to the difference in sheet resistance between front

and back contact. Hoppe and co-workers demonstrated DLIT measurements as

imaging tool to improve the fabrication of organic solar cells and modules.

Based on these experiments Bachmann et al. as well as Rösch et al. showed

that DLIT measurement in combination with several other imaging techniques

can be integrated as quality control tool for fabrication lines of thin film solar

modules [32,113].

Regarding the influence of defects on the surrounding cell, Buerhop et al.

presented an interesting feature which was previously not reported. The group

showed on commercially available thin film modules (amorphous silicon) that

each defect induced a distinct temperature gradient in its surrounding cell (see

D1 in Figure 3.8).


33

Figure 3.8: DLIT measurement of a commercial available a-Si thin film solar module with randomly distributed defects. The module was excited with a pulsed Voltage between 0 and 29 V resulting in an injection current of 3 A The lock-in frequency was set 0.1Hz. Image is reproduced from [29]

Simulations revealed that this gradient strongly depends on the defect

resistance and increases with decreasing defect resistance [29]. It was

assumed that one part of the injected current is attracted by the low resistance

of the defects. The injected charge carriers recombine in the defect and not in

the semiconductor material, resulting in a strong reduction of the heat

dissipation of non-defective cell parts.

In general DLIT measurements reveal different applications for

characterizing large area solar cells and modules. It could be shown that DLIT

can be used to obtain quantitative values of local distributed cell parameters.

Depending on the application, different power loss processes can be detected

and analyzed.

3.4.2 Illuminated lock-in thermography

During the last decade DLIT has been intensively developed as a standard

characterization method of thermal losses within different solar cell technologies

[111,128,148,151]. However, DLIT measurements are restricted by the injection

of current and therefore strongly related to the dark JV characteristic of a device.

Cell regions which are shielded by high series resistances cannot be visualized

and do not contribute to complete cell characterization. Furthermore, the sheet


34

resistance of the cell electrodes induce an inhomogeneous current distribution

resulting in material specific gradients of local power dissipation, as described

by Bachmann et al. [133]. In general, only electric power losses are visualized

while light induced losses are completely neglected.

To overcome these limitations a further development of DLIT called “illuminated

lock-in thermography” (ILIT) was developed [110,152]. In contrast to DLIT

measurements the cell is excited by a pulsed light source which enables the

characterization of thermal losses under real operating conditions or at different

cell fabrication steps. This advantage needs to be paid by a homogenous heat

distribution of the cell which can be attributed to thermalization and Peltier

heating [110,123,152]. A comparison of both DLIT and ILIT measurements of a

Si solar cell can be seen in Figure 3.9.

Figure 3.9: DLIT image (left) and ILIT image (right) of a silicon solar cell. DLIT measurements were performed at 6 Hz and 500 mV forward bias while ILIT was measured at a lock-in frequency of 31 Hz under the Pmpp condition (Vmpp=502mV). As Light source an IR semiconductor laser with a wavelength of 914 nm was used. The white square marks striation ring due to the Czochralski process. Images are reproduced with permission from [110].

Both measurements contain similar information while the ILIT

measurement under Pmpp condition shows additional thermal features.

Quantitative analysis revealed a strong difference of power dissipation of

defects under dark and illuminated condition. At DLIT 62% of the injection

current gets lost by the defects while under illuminated conditions (ILIT) the

losses were calculated to be 2.8%. The observed difference between both DLIT

and ILIT measurement could be related to thermalization and to the difference

of dark and illuminated current paths in solar cells [110]. Keas et al showed with

different monochromatic light excitations that thermalization strongly depends


35

on the used light source and increases with increasing light energy [152]. ILIT in

combination with different load conditions (e.g. short circuit, open circuit)

revealed further light induced loss processes. By creating the Voc conditions,

ILIT is sensitive enough to identify power losses due to recombination

processes of charge carriers at the grain boundaries of multicrystalline Si cells

[35,128,152]. This shows here a clear advantage of ILIT over DLIT, as it was

not possible with DLIT to identify these losses. By creating the Jsc condition (Jsc-

ILIT), the thermal losses are mainly due to thermalization joule heating due to

series resistances within the device [28,131]. However, ILIT measurements

correlate with the charge carriers lifetime [153] and can be used to characterize

different thermal loss processes in silicon solar cells. An interesting application

of Voc-ILIT was demonstrated by Straube and co-workers [130] who

investigated randomly distributed and fabrication related defects in micromorph

tandem solar modules. In this module technology, the tandem cells consist of

an amorphous Si top layer and microcrystalline Si bottom layer which are

connected in series with each other. Due to the complementary absorption

spectra and the use of different light sources, the single sub-cells could be

investigated separately. A comparison of a DLIT measurement and Voc-ILIT

measurements with different spectral light excitation can be seen in Figure 3.10.


36

Figure 3.10: Comparison between DLIT and Voc-ILIT images excited with different light spectra. Defects appear for spectral illuminations and are marked with circles. Image is reproduced with permission from [130].

From the DLIT and ILIT (white light) eight hot spots could be identified at

different positions inside the module. With the use of different light sources,

three defects could be identified in the top layer (blue light) while four could be

identified in the bottom layer (infrared light). From quantitative analysis the

defect induced leakage current could be calculated between 66 µA and 540 µA.

The group demonstrated with simple experiments that defects cannot only be

localized in the x-y area with ILIT but also investigated in z direction.

One of the first quantitative studies in which the emission of hot spots

was correlated with the device performance was presented by Vetter and co-

workers [154]. With Voc-ILIT the group investigated randomly distributed defects

in several CIGS solar modules comprising 67 cells which were provided from a

large industrial production. Based on the theoretical background from

Breitenstein and co-workers [128], an algorithm was developed (similar to Eq.

6) in which the ILIT signal of defects were set in relation with an overall ILIT

signal of a test module. The result was a defect related ILIT parameter X which

was then set in relation with the respective module performance.


37

Figure 3.11: Relation between module maximum peak power Pmpp and ILIT defect value X for 103 thin film CIGS solar modules. The Pmpp was extracted from illuminated JV-characteristics measured under the same low light conditions (30 W/m²) as used for ILIT-Voc measurements. Image is reproduced with permission from[154].

First analyses revealed only a rough relation between X and module

performance. The data analysis was improved by the introduction of a balancing

factor g which respected the defect position inside the cell (see Figure 3.11).

This needed to be done, since defects located close to the patterning lines have

a stronger electrical influence on the cell compared to defects in the cell center

[14,155].

In a more detailed study Besold et al. combined 2D electrical simulations

with the IR emission of defects, in order to assign each defect with ohm specific

value [122]. A representative module of this investigation is shown in

Figure 3.12.


38

Figure 3.12: (left) Voc-ILIT measurement of flexible OPV modules with 10 cells connected in series. (right) Determination of shunt values by a correlation between Voc-ILIT and electrical simulations of the power losses at the defects A,B, and C at different irradiation powers. Due to the weak signal of defect D a characterization was not possible. Both images are reproduced with permission from

[122]

The labeled hot spots showed during the investigations a strong ohmic

behavior which was attributed to constant photo-shunt behavior of defective

cells at different light intensities. A direct comparison of the light dependent IR

emission and electrical power dissipation of a simulated defect showed a clear

correlation. With this method it was shown for the first time that defects in solar

modules can be assigned to a distinct ohm resistance without the need to

access the single cells separately.

3.4.3 Loss analysis with EL imaging

Electroluminescence (EL) measurements prove the reverse functionality of a

solar cell. Charge carriers are injected and the solar cell works as a light

emitting diode (LED). The injected current passes through the electrode/active

layer interface and causes radiative recombination of charge carriers within the

active layer. As side effect the quality of the electrode/active layer interfaces are

tested simultaneously. However, solar cells are designed for an efficiency

energy conversion from light into electricity. Therefore, the spectral maximum of

the EL emission is shifted in the near IR (0.8 µm ~ 2 µm) and strongly reduced

by few orders of magnitudes. Depending on the cell technology the maximum of

the spectral emissions might range between 0.9 µm till 1.6 µm.


39

Since high advanced and two dimensional charge coupled device (CCD)

detectors are available, this radiation can be detected and converted into a

spatial resolved EL image. Based on the detailed balance theory from

Shockley-Queisser [96], Rau et al. developed two basic reciprocities between

photovoltaic solar cells and LEDs. The first reciprocity relates the local Voc of a

cell to the EL quantum efficiency while the second relates the spectral and

angular dependencies of the EL emission of solar cells and LEDs to the spectral

and angular photovoltaic external quantum efficiency [156]. Therefore, the

detected EL signal can be written as:

𝑆𝑐𝑎𝑚(𝑥, 𝑦) = ∫𝑄𝑐𝑎𝑚(𝐸)𝑄𝑒(𝐸, 𝑥, 𝑦)𝜑𝐵𝐵 (𝐸)𝑑𝐸𝑒𝑥𝑝 (𝑞𝑉(𝑥,𝑦)

𝑘𝑇) Eq. 7

with Qcam(E) as the energy sensitivity of the detector, Qe(E,x,y) as external

quantum efficiency, and φBB(E) as the spectral photon density of a black body.

Large effort has been done to develop EL imaging to a standard

characterization tool for Si solar cells. As one of the first groups Fuyuki and co-

workers combined the EL emission of a polycrystalline solar cell with minority

carrier diffusion length mapping and proved a linear relation between both

signals [157]. Ramspeck and co-workers visualized locally distributed

recombination currents and series resistances (RESI-imaging) by combining

both DLIT and EL imaging [148]. Therefore the local Rs distribution was defined

as:

𝑅𝑠 =𝑉𝑒𝑥𝑡−𝑉𝑙𝑜𝑐(𝑥,𝑦)

𝐽𝑟𝑒𝑐(𝑥,𝑦) Eq. 8

where Vext represents the applied voltage at the contacts, Vloc(x,y) the local

voltage at position x,y, and Jrec(x,y) the local recombination current density. The

local recombination current density was calculated by combining the local

power dissipation (Eq. 6) with the local voltage distribution. Thus, the local

voltage can be defined as:

𝑉𝑙𝑜𝑐(𝑥, 𝑦) =𝑘𝑇

𝑒ln[𝑆𝑐𝑎𝑚(𝑥, 𝑦)] + 𝑐 Eq. 9


40

with c as calibration constant and SEL(x,y) as the local EL emission. With Eq. 8

and Eq. 9 the local Rs can be imaged as presented in Figure 3.13.

Figure 3.13: Image of the local Rs distribution of a multicrystaline Si cell calculated using Eq 8. Image is reproduced with permission from [148].

Further improvement of Rs imaging based on EL imaging was presented by

Hinke and co-workers who used the local EL emission φ(U) and its derivative

φ´(U) with respect to the applied voltage. Combing both measurements yield to

the visualization of local Rs distributions [158]. Kirchartz and co-worker proved

the two reciprocities introduced by Rau with a comparative study between both

CIGS and multicrystalline Si solar cells [36]. Comparing the spectral resolved

EL emission of different cell temperatures, two emission stages were found and

identified. At low temperatures (<140°C) the EL emission of CIGS cells was

found to be dominated by donor and acceptor recombination processes of

charge carriers, while at high temperatures (>200°C) the EL emission was

found to be dominated by band to band recombination. On the other hand, for

multicrystalline Si solar cells the EL emission was found to be governed by

transverse optical phonon assisted band to band recombination which did not

show any temperature dependency. In a further publication Kirchartz et al.

proved the applicability of the detailed balance theory on bulk hetero junction

and organic solar cells [159]. Based on the work Helbig and co-workers

presented an EL imaging method to extract local cell parameters from CIGS


41

solar modules [132]. A clear relation between the EL emission at each position

and the actual voltage drop across the junction was found. With this information

it was possible to determine the sheet resistances of both the Mo and ZnO

electrodes by analyzing the local EL emission only. Furthermore the

comparison of EL emission at different applied biases enabled a contactless

reconstruction of local and global JV characteristics of cells and modules.

Regarding the development of EL imaging at OPV devices large

improvement of understanding the optical properties has been made. Tvingstedt

and co-workers showed that the EL emission of bulk heterojunction solar cell is

not the sum of the radiative recombination processes of polymer and fullerene

material. Instead, spectral EL measurements on different active layer blends

showed that a new electric state between the donor and acceptor molecules is

formed. This charge transfer (CT) state was found to be responsible for the

emission of EL in OPVs. Further investigations showed that the CT state is

energetically lowered compared with the LUMO of the acceptor molecule.

Hence, the EL emission is red shifted (0.9 µm-1.7 µm), compared to the EL

emission to the pristine semiconductor components. Since Si based CCD

cameras have a spectral response between 500 nm and 1000 nm EL imaging

of OPV cells was limited. This changed with the development of high advanced

EL cameras equipped with InGaAs detectors which have a spectral response

between 0.9 µm and 1.6 µm. However, pioneers in this field were Hoyer and co-

workers who presented firstly EL images of organic solar modules comprising

10 cells connected in series with an P3HT:PCBM active layer and normal

device geometry [160]. The EL images were taken with a Si based CCD camera

and compared with EL images taken with an InGaAs camera. The results are

shown in Figure 3.14.


42

Figure 3.14: (a) Photograph of the investigated module. (b) EL-measurement made with a camera equipped with an InGaAs detector. (c) EL measurement made with a camera equipped with a Si CCD detector. (d) EL-measurement made with a camera equipped with a Si CCD detector and a cutoff filter at 950 nm. The white arrow marks a center of EL emission. The image is taken with permission from [160]

The EL measurements showed similar features. Both images reveal a bright

irregularity in the second cell which was attributed to a corroded electrode

resulting in no EL emission of the defective cell area. At high injection currents

the EL radiation showed a distinct emission gradient which was previously

discussed in detail in [114,132,133]. In a different work the group showed on

OPV cells comprising a PCPDTBT:PCBM active layer a strong correlation of the

EL emission and LBIC measurements. It was suggested that EL images of OPV

cells are representative for spatial EQE variations [135]. The investigations of

Hoyer and co-workers demonstrated that EL imaging on organic solar cells

reveal local features and can be easily used to investigate variations of the local

loss processes. In a later study Seeland and co-workers presented EL and

photoluminescence (PL) images of OPV devices as function of different aging

stages. The investigated cells were stored under dark and irradiated conditions

(100 mW/cm²). Under both conditions the investigated cells suffered from a


43

strong dark spot formation which mainly disturbed the local EL emission. An

example cell stored in dark is presented in Figure 3.15.

Figure 3.15: (left) EL images (top) and PL images (bottom) of a P3HT:PCBM solar cell measured after different dark storage times. (right) Time depending decay of cell parameters. The images are reproduced with permission from [117].

Interesting for cells stored in dark, the PL radiation seemed to be undisturbed

and homogeneously distributed over the cell area. This was essentially different

to the PL measurements (not shown here) where a dark spot formation, similar

to the EL images, was observed. Under both storage conditions a strong

reduction of Voc and Jsc was investigated. The dark spot formation was

attributed to both corrosion of the electrode/active layer interface and photo

oxidation of the active layer. The group concluded that the combination of EL

and PL imaging allows distinguishing between active layer photo-degradation

and corrosion of the active layer electrode interface. Since the PL emission was

measured under the Voc condition a reduction of this signal would be directly

related to the active layer quality.


44

4. Degradation and Stability of OPV solar cells

45

4 Degradation and Stability of OPV solar

cells

In the most common definition, degradation refers to a process of

continuous impairment of material describing quality parameters. In terms of

solar cells, degradation leads to a continuous reduction of specific cell

parameters such as Jsc, Voc, FF, and PCE. This continuous reduction might be

accelerated by different environmental storage conditions. Therefore, one

needs to distinguish between intrinsic and extrinsic degradation. Intrinsic

degradation mainly refers to direct changes of cell properties without being

triggered by external influences. On the other hand, extrinsic degradation refers

to changes which are directly related to environmental influences. In particular,

the diffusion of oxygen and water, temperature cycling, or the irradiation of light

might induce photo-oxidation reactions, local trap formation, as well as

morphological changes of the cell. Furthermore, adverse environmental storage

conditions can lead to defect formation at the electrodes and the

electrode/active layer interface of a solar cell resulting in an strong reduction of

cell performance. In the following different degradation aspects, known from the

literature, are summarized.


46

4.1 Degradation of active layer

The active layer (AL) of a OPV device is prone to both light and oxygen

induced degradation. Since the AL is essential for the functionality of the device,

the degradation of the AL is directly linked to a reduction of the efficiency. Under

irradiated conditions singlet oxygen or ozone can react with the side chains of

the polymers, resulting in chain scission and a reduction of life time

[17,161,162]. The exact way of photo-degradation is still under discussion but

one of the most accepted models claims that photo-degradation is originating

from singlet oxygen photosensitization [163]. The different photochemical

degradation paths can be classified in:

Photo-bleaching of the polymer part within the active layer

[163,164]

Photo-oxidation which mainly leads to an disruption of π-bonds

and chain scission [67].

formation of free radicals due to photolysis of metallic impurities

resulting in cross linking and chain scission [165]

In general it is assumed that the degradation of P3HT (Poly(3-

hexylthiophen-2,5-diyl)) originates from photosensitization of singlet oxygen.

Due to this mechanism the singlet oxygen undergoes a Diels-Adler

cycloaddition with the thienyle unit of P3HT, resulting in the formation of an

unstable endo-proxide. This endoproxide might decompose into sulfine or trans-

deketone. In both cases the π conjugation of the polymer is disturbed which

leads to a reduced stability. It was found that free radicals might attack the

thienyle ring of the polymer which directly influenced the conjugation polymer

[165]. This process was classified by Marceau et al. who defined three different

mechanisms of photo-oxidation which are illustrated in Figure 4.1.


47

Figure 4.1: Different oxidation mechanism of P3HT polymer. (i) H abstraction of alkoxy radicals form a α unsaturated alcohol, (ii) cage reaction of alkoxy radicals and the irradiation of UV light form of aromatic carboxyilc acids groups and aliphatic acids, and (iii) β-scission resulting in the formation of carboxylic acids. Image is reproduced with permission from [161]

From the scheme presented in Figure 4.1 it can be noticed that oxygen in

combination with highly energetic light (UV) mainly damages the polymer side

chains. A study by Reese and co-workers demonstrated that P3HT films under

irradiation rapidly photo-oxidize, followed by a complete bleaching of the films

after 700 h [164]. Blending P3HT with PCBM (phenyl-C60-butyric acid methyl

ester) slowed down the above mentioned process. It was concluded that the

oxygen reacts with PCBM by the creation of species with up to eight oxygen

atoms. The result was new energetic state which acts as a trap state for the

electron transport and leads to decreasing electron mobility [164]. A similar

observation was made by Manceau et al. who concluded, that the photo-

degradation rate of P3HT is strongly attenuated and nearly suppressed when

P3HT is blended with a fullerene such as PCBM [20]. Schafferhans and co-

worker investigated the effect of oxygen doping at P3HT:PCBM solar cells. The

investigations showed that cells which were stored in air and in the dark, had a

reduction of the photocurrent by up to 60 % after 120 h of storage. Reference

cells under continuous irradiation, however, showed a decrease of almost all

cell parameters. This reduction could be attributed to the increase of oxygen


48

induced trap states with activation energies of about 100 meV [18]. An

interesting phenomenon was observed by Aguirre et al. who demonstrated the

formation of meta stable charge transfer states as a first step of photo-induced

degradation [166]. On two different conjugated polymer systems parasitic

radical cations on the polymer side chain were found, if the samples were

irradiated in air. In a further study Seemann et al. showed with CELIV studies

that parasitic charges were formed when P3HT:PCBM cells were illuminated

and exposed to air. This parasitic behavior could be reversed when the cells

were stored in vacuum. Seemann et al. concluded that oxygen forms together

with P3HT a weakly bound P3HT+O2 complex which acts as a reversible trap

state [118]. Finally Manceau and co-workers developed a rule of thumb for the

chemical stability of conjugated polymers under continuous irradiation. In a

detailed study it was shown that a polymer with a small number of side chains

had a much more stable chemical behavior as polymers having a large number

of side chains [162].

4.2 Degradation of electrodes

In order to achieve high efficient organic solar cells the right choice of

electrodes is essential. State of the art OPV devices mainly consist of two types

of electrodes. As top electrode and high conductive metal is used while as

bottom electrode a TCO is used which enables the transmission of light into the

AL. Especially the work function (WF) of the electrodes is primarily responsible

for the built in potential and optimal charge carrier extraction. Any variation

would automatically result in a reduction of PCE and FF.

4.2.1 Metal electrode

Commonly used electrode metals are aluminum (Al), calcium (Ca), and

silver (Ag). In particular, due to the low work function of Ca, electrodes

consisting of a composition of Ca and Al show a high chemical reaction with

water and oxygen [67,167]. Norrman and co-workers showed with spatially

resolved TOF-SIM (time of flight secondary ion mess spectroscopy)


49

measurements that water and oxygen can diffuse through the metal grains or

pinholes resulting in a modification of the AL/electrode interface due to

corrosion [24,168]. The result is the formations of voids or insulating patches

leading to a rapid reduction of charge carrier extraction and an increase of Rs. It

was assumed that this might lead to delamination of electrode/active layer

interface. With similar investigations, Hermenau and co-workers demonstrated

that the diffusion of water has the a stronger influence on the modification of the

electrode/active layer interface than oxygen [109]. One possibility to avoid the

degradation of the metal electrode is the use of more inert metal electrodes

such as silver [169]. In contrast to Al and Ca shows Ag a less distinct reaction

with water and oxygen resulting in lift times of several hundred hours [19,170].

However, Llyod and co-worker demonstrated that the WF of a 100 nm thick Ag

electrode deposited on a P3HT/ITO substrate decreased from around -4.3 eV

up to -4.5 eV in the presence of oxygen. In average an increase of about

0.23 eV was observed which was attributed to the formation of Ag2O at the Ag

film [171].

4.2.2 Degradation of hole transport layer

The most investigated HTL material is PEDOT:PSS. The polymer is ionic

and dissolved in a solution of water. Afterwards, this solution can be deposited

to thin films on several substrates by using different coating techniques. To

evaporate the water content of the PEDOT:PSS films additional annealing steps

over 100°C are required. After annealing the PEDOT:PSS films show an

increased conductivity, an improved morphology, and strong reduction of weight

[172–174]. With the help of additives the WF of PEDOT:PSS can be easily

changed [175]. Some formulations of PEDOT:PSS are highly acidic which might

induce in conjunction with moisture from the air or residual water in the cell

stack, corrosion of metal electrodes [176]. De Jong and co-worker showed that

PEDOT:PSS in a normal device geometry might also influence the ITO

electrode. The group stored glass slides with an ITO/PEDOT:PSS/PPV stack in

an nitrogen atmosphere at 100°C. With Rutherford backscattering (RBS) it was

possible to show that small amounts of In from the ITO defused into the


50

PEDOT:PSS layer after 2500 h of storage. Interestingly the In uptake was

accelerated by samples which were stored in air at 8°C. The group concluded

that due to the hydroscopic nature of PSS water is absorbed from the air, which

then causes an etching of the ITO surface [177]. Kawano et al. investigated as

one of the first groups the influence of PEDOT:PSS on the long-term stability of

MDMO-PPV:PCBM solar cells with a normal device structure. Therefore, non-

encapsulated test samples were stored in humid and dry air. It was found the

samples under illumination degrade faster in humid air (> 40 RH) than in dry air

or in a nitrogen atmosphere [23]. Similar to de Jong, the authors attributed the

shorter life time of the cells stored in humid air to the hydroscopic nature of the

PEDOT:PSS. It was concluded that due to the absorption of water, the series

resistance of the PEDOT:PSS layer increased. The thermal stability of

PEDOT:PSS was investigated by Vitoratos and co-workers. In particular, the

group investigated the electrical conductivity of PEDOT:PSS films in

dependency of annealing time at 120°C. An exponential decrease of

conductance with a half-life of 50 h was observed [173]. Lloyd and co-workers

investigated the degradation behavior of inverted P3HT:PCBM solar cell with

and without PEDOT:PSS as hole transport layer [178]. After 600 h of

continuous irradiation at 70 mW/cm² the sample comprising an PEDOT:PSS

layer showed a fatal break down of the Jsc. This decrease could be attributed to

the formation of dark spots (see figure 4.2) in the device which grew in size with

increasing measurement time.

Figure 4.2: LBIC images of an inverted P3HT:PCBM solar cell with PEDOT:PSS as HTL at different aging stages. On the left side the fresh state can be seen while on the right side the degraded stage is shown. The cell was stored in air and continuously irradiation with light provided from a sulfur plasma lamp. The light intensity was ~70 mW/cm². Image is reproduced with permission from [178].


51

The authors suggested, similarly to Kim and co-workers [179], that H2O

diffuses through pinholes of the electrodes and dissociates to H+ and OH- ions.

These ions locally dedope PEDOT:PSS to an non-conducting state. The result

is an subsequent increase of Rs of the PEDOT:PSS layer or, in extreme cases,

a local delamination of the PEDOT:PSS/Ag interface. Voroshazi et al.

demonstrated that PEDOT:PSS favors an accelerated reduction of the device

performances of OPV devices with a normal device geometry [180]. In particular,

it was found that PEDOT:PSS accelerates oxidation at cathode by absorbing

moisture from the air. Due to the substitution of PEDOT:PSS with MoO3 an

enhanced stability could be observed. Norrman and co-workers investigated the

degradation behavior of inverted P3HT:PCBM solar cells stored in different

atmospheres [181]. With the help of chemical surface analysis techniques such

as X-ray photoelectron spectroscopy and TOF-SIM measurements, the group

showed a dynamic chemical behavior at the HTL/active layer interface. It was

found that PEDOT:PSS tends to phase separation and that the PEDOT and the

PSS phase showed different reactivity towards oxygen with selective oxidation

of the PEDOT phase and subsequent oxidation throughout the device layers

beneath the PEDOT phase [181]. Feron et al. examined with help of LBIC

measurements the diffusive ingress of water into inverted OPV devices which

were stored in air at different relative humidities. It was found that the water

ingress clearly occurred at pinholes of the metal electrode and at the device

edges [182]. It was assumed that moisture mainly diffuses though the

PEDOT:PSS layer. Using the second Fick law of diffusion a diffusion coefficient

could be quantified with D = 5∙10-6 cm²/s.

The cases described above, show clearly that the use of PEDOT:PSS as

HTL entails various problems, which directly affects the long-term stability of

OPV devices. To overcome these problems current research mainly focused on

the replacement of PEDOT:PSS with high WF metal oxides such as MoOx or

WOx. The advantage of these metal oxides is a higher chemical stability and life

time of OPV devices [19,69,70,169,180].


52

4.2.3 Degradation of electron transport layer

Material compositions which are commonly used for electron transport

layers are zinc oxide (ZnO) or aluminum doped zinc oxide (AZO) [73,183]. Both

materials are mainly used for solar cells with an inverted structure and provide a

high electron mobility [184,185]. Especially the conductance of ZnO can be

varied by oxygen doping or by doping with Zn atoms [186]. Interestingly cells

which incorporate a ZnO layer show a distinct s-shape formation after

processing. This phenomenon is widely discussed in the literature

[183,184,187–189]. It is assumed that due to a reversible process bound

oxygen radicals react with ZnO at the ZnO/active layer interface and form a

counter diode. The result in a s-shape deformation of the JV characteristics

which can be cured by UV radiation.

Figure 4.3: Absorption and desorption of oxygen at the ZnO interface, Image is reproduced with permission from [68]

One of the first explanations of this memory effect was made by Verbakel

et al. [190]. The group demonstrated on ZnO nanoparticle films that the

conductivity could be strongly improved after a short treatment of the irradiation

of UV light or the appliance of a high forward biases (see Figure 4.3). It was

concluded that due to both the UV and forward bias treatments bound oxygen

from the ZnO surface is desorbed leading to an improvement of the n-type

conductivity. Interestingly, using these treatments the initial state of the devices


53

was restored. Lilliedal and co-workers showed a similar phenomenon on OPV

modules based on an inverted device geometry with P3HT:PCBM as active

layer [191]. The investigated modules showed a distinct s-shape deformation

JV-characteristics guided by complete reduction of photocurrent extraction. Due

to continuous UV irradiation the s-shape deformation disappeared resulting in

normal JV characteristic and improved charge carrier mobility. Further

investigations showed a reappearance of the s-shape deformations after a

certain time of dark storage. A more detailed study about the impact of oxygen

and moisture on electron transport layers was presented by Litzov and co-

workers [192]. Therefore the group investigated the optical and electrical

properties of two different types of AZO layers, processed with high and low

temperature treatments, in dependency of damp heat storage condition in dark.

In both cases the work function and conductivity of the films showed a clear

degradation upon damp heat exposure which indicates an interaction between

the AZO surface and water or oxygen. The result was a formation of a

secondary depletion region limiting the charge carrier extraction and device

stability.


54

5. Methods for device characterization

55

5 Methods for device characterization

5.1 Lifetime evaluation of OPV devices

Since OPV solar cells show efficiencies above 11%, their long term stability

gains more interest in both the scientific and industrial community. In particular,

a life time of several years is required to join the market and being competitive

to other standard cell technologies. However, due to the use of different

semiconductor polymers, device geometries, and encapsulation materials OPV

devices show a large deviation in terms of their long term behavior. Several

degradation studies show the relation between environmental conditions (e. g.

light, temperature, moisture) and degradation phenomena which directly

influence the life time [68,193,194]. Furthermore, due to the lack of standard

procedures describing accurate life time experiments, the test conditions often

significantly vary during the different degradation studies. To overcome these

limitations the first two international summits on OPV stability (ISOS1&2)

established a set of guidelines [195] which allow comparable and accurate life

time studies of different OPV technologies. As consensus four test conditions

are defined and refer to:

Shelf life studies under ambient conditions (ISOS-D1) or

controlled environmental conditions (ISOS-D2 & D3)

Photo-aging studies under indoor conditions (ISOS-L)

Outdoor studies in the field under both real and artificial sun light

(ISOS-O)


56

Temperature cycling studies between different temperature and

humidity settings (ISOS-T).

The purpose of the guidelines is not to build up a rigorous standard procedure

but rather a standardized protocol which enables a simplified comparison of

stability measurements of different research groups. Therefore, each test

protocol is separated into three levels of procedures; basic, intermediate, and

advanced. However, the core of the procedures is the characterization of test

samples stored at different environmental conditions by measuring the JV

characteristic either in-situ or after well-defined time intervals. From the time

depending characteristics of the cell parameters, extracted from the JV

characteristics, the life time of an OPV device is determined as shown in Figure

5.1. Here, the life time is defined as the difference of the time parameters T0

and T80 which represents the time interval between an initial value E0 and its

decrease of 20% (E80). Since OPV devices show a rapid and exponential decay

(burn-in) of cell parameters under certain degradation conditions (e.g. light) a

second definition of life time is introduced.

Figure 5.1: Typical degradation characteristic of an OPV. The eight parameters are used to define the device lifetime. Parameters labeled with an s refer to a stabilized state after the burn in period. Image was reproduced with permission from [194]

As can be seen in Figure 5.1, after a certain time TS the burn in period is

stabilized and the PCE characteristic turns into a linear behavior. Starting from


57

this point the life time is defined as the time interval between TS and TS80.

Depending on the material system some OPV devices show a relatively strong

stability with only small changes within the first hundred or thousand hours of

investigation. In this case the life time of a device can be extrapolated by

applying a linear regression to the stabilized and linear decrease of the

degradation characteristics. Using this method Peters and co-worker

demonstrated lifetimes of about 7 years for OPV devices with an

PCDTBT:PCBM active layer [196]. For this study the devices were degraded

under continuous irradiation at 1000 W/m² and the life time was extrapolated by

dividing the measured life time (TS80-TS) by the annual amount of sun hours on

earth.

5.1.1 Sample holder for degradation cycles

Since degradation refers to a continuous decrease of quality describing

parameters, a reliable measurement of these parameters is essential to

understand different degradation phenomena in solar cells. Especially, a

reproducible connecting of the electrodes is one of the major concerns, since

the electrodes might be corroded or damaged during degradation and

characterization. Furthermore, mounting and demounting of the samples

causes variations of contact resistances and limits an accurate measurement of

Jsc, FF, Voc, and PCE (see table 5.1). To overcome these limitations the test

samples need to be fixed into sample holders during the entire degradation

study. Therefore, a new sample holder was developed during this study in order

to enable similar connecting conditions of a test sample for JV measurements

after different degradation stages. A photograph of representative sample

holder can be seen in Figure 5.2.


58

Figure 5.2: Photograph of a representative sample holder used for degradation studies, developed at ZAE-Bayern.

The holder consists of a printed circuit board (PCB) with ten embedded gold

pins. Each gold pin is spring loaded which allows an optimal electric contact.

Due to the device geometry of the test samples, the two pins on the top and

bottom of the PCB contact the ITO electrode while the other pins on the right

and left side contact the Ag electrodes. After placing the test sample on top of

the pins, the sample is fixed with a front plate, screws and nuts to enable a

constant contact pressure of the pins on the electrodes during an experiment.

JV characteristics are measured by connecting the holder to a relay board

which selectively controls the single OPV cells of a substrate. The reliability of

the measured cell parameters for a fixed test sample and a test sample which is

mounted and demounted to the holder can be seen in Table 5.1.

Table 5.1: Cell parameters of a test cell place at two different holder systems. The cell parameters are extracted averaged from ten illuminated JV-characteristics of a test cell measured with an AM1.5G spectrum and 1000W/m².

test cell Jsc

(mA/cm²)

Voc

(V)

FF

(%)

PCE

(%)

fixed 8.4 ± 0.23% 0.57 ± 0.67% 60 ± 0.44% 2.9 ± 0.19%

(de)mounted 8.5 ± 2.42% 0.56 ± 0.43% 60 ± 3.43% 2.8 ± 3.62%


59

From the relative errors it can be seen that mounting and demounting of a test

sample induces variations of extracted cell parameters for Jsc, FF, and PCE and

limit an accurate data extraction. If the test sample is fixed to a holder during the

degradation study, a high repetition rate could be achieved resulting in only

minor deviations of the measured cell parameters.

5.1.2 JV characterization

The JV characterization belongs to one of the most used characterization

method in order to quantify cell specific cell parameters such as Voc, Jsc, FF,

and PCE (see Figure 2.10). For a JV measurement sequences of biases are

applied to a sample while the current is simultaneously measured either under

dark or irradiated (light) conditions. Especially, under light conditions the used

light source needs to fulfill a set of requirements which are related to both the

irradiation power and the spectral distribution of light. First, the emitted light

spectrum should be similar to the AM1.5G spectra of the sun and second the

irradiation power needs to be 1000 W/m². Different light sources (e.g. xenon

lamps and metal halide lamps) in combination with daylight filter can be used to

fulfill the mentioned requirement [195]. However, most of these sources show

small spectral variations resulting in a spectral mismatch to the AM1.5G

spectrum. An elegant way to overcome these limitations is the use of an LED

based solar simulators as done for all illuminated JV measurement for this study.

Therefore, the light source is made of a 30x12 LED array consisting of 23

different LED spectra (see Figure 5.3). The spectral emission of the LED array

ranges from 400nm till 1100nm and is shown in Figure 5.4. The intensity of the

different LEDs was previously adjusted to match the AM1.5G spectrum at 1000

W/m² as described in [197]. To avoid overheating, the LED array is cooled by a

water controlled heat sink which is continuously set to 25 °C. Homogenization of

the light is achieved using a 30 cm long reflector tunnel. The irradiated area is

about 15 cm x 15 cm with a homogeneity mismatch of less than 2%. The

electrical characterization of the test samples either under dark or light

conditions is realized with a two wire connection and a source measure unit


60

B2900 from Agilent in bias range between -0.5 V – 1.5 V for single layer devices

and -0.5 V - +2 V for multilayer devices.

Figure 5.3: LED solar simulator used for illuminated JV characteristics

Figure 5.4: Irradiation spectra from the sun (AM1.5) and the LED solar simulator.


61

5.1.3 Setup of photo-degradation

The investigation of the photo-stability is described in the ISOS-L protocols.

Here, in the basic version (ISOS-L1) a test sample is continuously irradiated

with light at 1000 W/m² while an emission spectrum of the light source close to

the AM 1.5G spectrum is preferred but not required. However, most of the light

sources used for photo-degradation operate at high temperatures (>2000 K).

The life times of theses light sources are limited (e.g. xenon lamp ~2000 h) and

implications due to temperature degradation of the test samples need to be

considered as well.

One way to overcome these limitations is the use of high power light

emitting diodes (HP-LEDs) as light sources for photo-degradation experiments

as done in this study (see chapter 6). For all photo-degradation experiments in

HP-LEDs from Bridgeluxe (BXRA-30E0800-B-00) were used. The spectral

emission of the HP-LEDs ranges from 400 nm and 750 nm with a local

maximum around 445 nm (see Figure 5.5). The emitted light is homogenized

using a parabolic and highly reflective reflector. For the injection of current

resulting in the emission of light two different standard power supplies made by

Agilent (type: E3640A) and Hewlet Packert (type: E2632A) are used. To avoid

overheating, the LEDs are mounted on a passive heat sink which is cooled by

air.

Before each photo-degradation experiment, the light intensity of a HP-

LED is adjusted to 1000 W/m² by using a test cell with an active layer material

similar to the test samples. To neglect the spectral mismatch between the HP-

LEDs and an AM1.5 spectrum, the light intensity of the each LED was adjusted

until Jsc,HP-LED of the test cell matched the Jsc value measured under the solar

simulator at 1000W/m².


62

Figure 5.5: Emission spectrum of the LEDs used for photo aging. Inset: Photograph of the setup used for photo-aging of organic solar cells. One LED is placed above the investigated sample which is mounted in a sample holder. Image is reproduced with permission from [198]

5.2 Voc-Measurement of thin film solar cells within a

module

As described in section 3, a defect creates an alternative current path resulting

in a decrease of the electrical power output and open circuit voltage of a cell

(Voc,cell). The influence of the defect on the short circuit current (Isc) is negligible

[38]. In order to quantify the defect induced drop of Voc a new method to

measures the Voc,cell in a module was developed. During taking measurements

of Voc,cell, the whole solar module except for one cell is covered by a house

made mask. The mask is made of two polyester blades; each attached to a

piece of black velvet and fitted to the required size to cover the module. The two

parts of the mask are positioned in parallel with the cell of interest. The

uncovered cell is irradiated with light while the module terminals are connected

to a Keithley 2000 multimeter for measuring the open circuit voltage of single

and uncovered cell. Figure 5.6 shows the equivalent circuit diagram of the setup

used for the Voc,cell determination.

*This section is adopted with permission from ref [129] (Copyright Elsevier)


63

Figure 5.6: Circuit diagram showing the open circuit voltage evaluation of a single thin film solar cell within a module with solar cells each connected in series. Extracted with permission from [129] (with the permission of Elsevier

The irradiated cell (1) is shown by the one-diode-module as found in the middle

of Figure 5.6. The covered cells are represented by the same model, but

without a current source (2). The multimeter acts as a high ohmic load (1∙107 kΩ

in a voltage range of 0-1 V [199]) in the circuit and due to its high ohmic input

resistance, a low photocurrent flows through the covered cells and the

multimeter. In the covered cells however, no photocurrent is generated. Hence,

the covered cells operate in a reverse direction resulting in a blocking of the

diodes. Since the current only flows through the parallel and series resistances,

the covered cells may be replaced by a high ohmic series resistance Rsp,i. Thus,

the resistance Rsp,i of a covered cell is calculated by:

𝑅𝑠𝑝,𝑖 = 𝑅𝑝,𝑖 + 𝑅𝑠,𝑖 Eq. 10

Replacing the cell under illumination by a voltage source having the voltage

Voc,cell, the voltage measured by the multimeter Vm may be calculated by:

𝑉𝑚 = 𝑉𝑜𝑐,𝑐𝑒𝑙𝑙 ∙𝑅𝑚

𝑅𝑚+∑ 𝑅𝑠𝑝,𝑖𝑁𝑖=1

Eq. 11


64

with Rm as input resistance of the multimeter and N as the number of covered

cells. As long as the sum of Rsp << Rm, the influence of Rsp may be neglected

and the Vm is equivalent to Voc,cell. For estimating the inherent uncertainty in the

measurement, the absolute error of Vm may be calculated by using the total

differential (with Rsp,i as variables) resulting in:

∆𝑉𝑚 = ∑ |𝜕𝑉𝑚

𝜕𝑅𝑠𝑝,𝑖| ∙ ∆𝑅𝑠𝑝,𝑖 = ∑ |

−𝑉𝑜𝑐,𝑐𝑒𝑙𝑙∙𝑅𝑚

𝑅𝑚+∑ 𝑅𝑠𝑝,𝑖𝑁𝑗=1

| ∙ ∆𝑅𝑠𝑝,𝑖𝑁𝑖=1

𝑁𝑖=1 Eq. 12

Based on the measured data presented in [10] a single CIGS cell revealed an

Rp of 1732 Ω∙cm² and an Rs of 0.31 Ω∙cm² for ideal cells. This gives a Rsp of

about 1732.31 Ω∙cm². From equation 11 it may be seen, that the Vm of a cell

strongly depends on the relation between Rsp and the input resistance (Rm) of

the multimeter. If an ideal CIGS solar module (one without defects), 1 irradiated

cell, and 66 covered cells, no ΔVm can be calculated and Vm is then only limited

by the relation between the sum of Rsp and the input resistance Rm (Eq. 11). In

this case the difference of the measured and real Voc,cell for a module would be

less than 0.00048 %. Assuming a module with a Rsp of 1 Ω∙cm² for defective

(totally shunted) cell, the absolute error of Rsp (ΔRsp) would result in

1731.31 Ω∙cm². Accordingly, the ΔVm for this system can be calculated as

410 µV. The maximal deviation of the measured Vm will not be more than

0.06 % and can be neglected for further Voc investigations.

5.2.1 Cell voltage of modules with randomly distributed defects

In order to measure the defect induced drop of Voc of a single cell in a

thin film solar module the method described in the previous section was used.

The chosen test module was special because it contained a large variety of

different kinds of defects. The reliability of the above mentioned method was

tested by replacing the mask on the investigated cells. Therefore the Voc of

each cell within a module was measured five times in a row. The highest

relative error of Voc,cell estimation was evaluated with 3% using the standard

deviation of the five measurement. The local cell voltage distribution of the

*This section is adopted with permission from ref [129] (Copyright Elsevier)


65

investigated test module can be seen in Figure 5.7. Each data point represents

the averaged Voc,cell of five independent Voc,cell measurements.

Figure 5.7: Local cell voltages of a CIGS-module with 36 connected cells in series as measured by the above mentioned mask method (at 30 W/m² illumination power). The sum of the individual Voc-cell (12.40 V) is equivalent to the measured Voc of the whole module (12.41 V). Image was reproduced with permission from [129] (copyright Elsevier)

In general one can see that each defect induces a drop of maximal Voc,cell

indicated by the dotted line at 0.52 V. Especially severe defects (e.g. in cell 4

and 10) provoke a strong drop of Voc,cell resulting in very low cell voltages (<

50 mV). Due to the series connection, the sum of the Voc,cells must be equal to

the Voc of the non-covered, fully irradiated module. Regarding the data in Figure

8.2 the sum of the single cell voltages was 12.40 V, which is almost equal to the

measured voltage of the entire module (12.41 V). Assuming a module with non-

defective cells comparable to the non-defective cells in Figure 5.7 a module

voltage of 18.7 V was calculated. The voltage loss induced by the defects inside

the module could be therefore related to 6.3 V. In conclusion, the applied

method to measure the open cell voltage of single cells works well for ideal cells

and also for cells with severe defects.


66

5.3 Lock-in Imaging

Since high advanced IR cameras provide a noise equivalent temperature

difference (NETD) of about 10-20 mK [97] the detection of temperatures

differences lower than 10 mK is limited. Moreover, due to the continuous

excitation at a conservative IR experiment, a defect produces continuously heat

which diffuses into the surrounding cell. Hence, the size of a hot spot seemed to

be blurred and a precise localization inside the cell is not possible [200]. To

overcome these limitations the concept of lock-in thermography (LIT) was

developed where temperature differences of less than 100 µK can be detected

[151]. In recent years, several lock-in based imaging methods such as dark

lock-in thermography (DLIT), illuminated lock-in thermography (ILIT) or

electroluminescence lock-in (ELLI) have been published which differ in the way

of excitation and spectral sensitivity.

5.3.1 Theory of lock in based imaging

The principle of lock-in based imaging is described in [128,201] and will

only be summarized and described in the case of lock-in thermography. Here,

the dissipated power of a sample is periodically amplitude modulated by an

external excitation source (e.g. power supply). Each pulse induces a thermal

wave [202] inside the solar cell which interacts with discontinuities leading to an

detectable heat dissipation. Since high advanced IR camera provide certain

frame rate the generated IR images can be digitally processed according to the

lock-in principle. Hence, two primary correlation signals (images) can be

calculated and interpreted as the in-phase signal S0°(x,y) and the out-of-phase

signal S-90(x,y) with x,y as pixel position of the image. A schematic illustration of

this process is presented in Figure 5.8.


67

Figure 5.8: Schematic image of lock-in calculation. The IR images are reproduces with permission

from [129] (copyright Elsevier)

As can be seen in Figure 5.8, the two primary lock-in images are taken in a

correlation phase angle of 0° and 90°. The signal S for each pixel can be

described as

𝑆𝑖,𝑗 = (𝑓𝑖,𝑗 ∙ 𝑔𝑖,𝑗)(𝜏) = ∫ 𝑓𝑖,𝑗(𝑡) ∙ 𝑔𝑖,𝑗(𝑡)𝑑𝑡∞

0 Eq. 13

with f(t) as detected signal, g(t) as correlation function, t as time, and i,j as

position parameters of a pixel. Commonly used correlation functions for lock-in

based imaging are sine and cosine or two rectangle functions which have a

phase shift of 90° to each other. The two primary lock signals can be seen as

the real and complex imaginary part in complex analysis and can be, therefore,

used to calculate two secondary lock in signals:

𝐴(𝑥, 𝑦) = √𝑆0°(𝑥, 𝑦)² + 𝑆−90°(𝑥, 𝑦)² Eq. 14

𝜑(𝑥, 𝑦) = arctan(−𝑆−90°(𝑥,𝑦)

𝑆0°(𝑥,𝑦)) Eq.15

with A(x,y) as phase independent amplitude signal and φ(x,y) as phase signal

of the surface temperature.


68

Since the S0° and S-90° signal show a strong proportionality to the dissipated

power of a heat source, the amplitude image shows also a similar

proportionality of its dissipated power. At this point it is worth to note that during

the lock-in calculation the temperature information gets lost, as S0° and S-90° are

correlation signals and only show similarities between the captured signal and

the reference signal. As can be seen from Eq.15, the phase image, on the other

hand, is related to the quotient of the S0° and S-90° image and is independent

from the power dissipation and IR emissivity ε of the heat source.

Heat diffusion

Due to the periodically excitation a sample will slowly heat up until an

equilibrium state is reached. In this state the periodic heating will superimposed

with cooling to the surrounding material and the pulsed heat source act as a

harmonic oscillator [202]. The temperature distribution can be described by

using the non-steady state heat diffusion equation in 3 dimensional isotopic and

homogeneous solids:

𝑐𝑇𝜌𝜕𝑇

𝜕𝑡= 𝜆𝑇Δ𝑇 + 𝑝 Eq. 16

with T as temperature, t as time, cT and specific heat, ρ material density, λT as

heat conductivity. Its solution for an point heat source with the power

P(t)=P0sin(ω·t) is:

𝑇(𝑟, 𝑡) =𝐴

𝑟exp (

−𝑟

Λ) 𝑒𝑥𝑝 (𝑖 (𝜔𝑡 −

𝑟

Λ)) Eq. 17

Λ = √2𝜆

𝜌∙𝑐𝑇∙𝜔 Eq. 18

𝐴 = (𝑃0

2𝜋𝜆) Eq. 19

Here, r is the radial distance from the heat source center, ω is 2πflock-in, A the

amplitude factor, and Λ the heat diffusion length. From equations 16 it can be

seen that the temperature modulated amplitude diverges from source position


69

and decreased with 1/r, while Λ strongly depend on the lock-in frequency and

reduces with 1/√flock-in. Depending on cT, λ, and ω the blurring effect of a hot

spot can be influenced. For continuous excitation Λ is infinite while for high lock-

in frequencies Λ is strongly reduced and the detection of the defect’s shape is

only limited by the optical resolution of the imaging system.

5.3.2 Imaging setup

In Figure 5.9 a schematic diagram of the imaging setup used for all

imaging investigation in this work is presented. For an imaging experiment a

high sensitive IR or EL camera is placed above a test sample. Depending on

the used imaging technique, different excitation modes can be chosen. In case

of ILIT sixteen white LEDs (Citizen Electronics) are used for periodic excitation

of the test samples.

Figure 5.9: Schematically sketch of Lock-in setup for DLIT, ELLI and ILIT applications

Each LED has an electric power of 7 W and a spectral range from 410 nm to

780 nm with a spectral peak at 440 nm. The irradiated area comprises

40 cm x 40 cm with an illumination power of 30 W/m². The maximal deviation of

illumination power over this area is ± 6 %. The temperature of the samples and


70

the LEDs is controlled with a water cooled heat sink regulated by a Neslab

TermoFlex2500 (Thermo Scientific Inc., USA). For DLIT and ELLI

measurements the sample is excited with a power source (E3640A) made by

Agilent. To enable a pulsed electrical excitation a house made switch is set in

series between the sample and power supply and is connected with a trigger

signal provided by the IR camera.

5.3.3 IR camera

For all IR investigations a highly advanced EQUUS 327 SM IR camera

(IRCAM GmbH, Erlangen, Germany) controlled by a computer for real time

analysis is used. The camera is equipped with a focal plane array InSb detector

which is cooled (~77K) by a sterling engine. Due to detector cooling the thermal

energy of the charge carriers inside a pixel is strongly reduced resulting in

increased thermal response of the detectors. The detector has spatial resolution

of 640x512 pixels and is highly responsive in a spectral range between 1µm

and 5µm. The noise equivalent temperature difference (NETD) is 20 mK. The

typical frame rate of the detector is 100Hz. As lens system an IR objective with

a focal distance of 25 mm and a transparency >90% in a spectral range

between 2 µm and 5 µm is used. A trigger signal provided by the camera is

used to synchronize the lock-in algorithm with the pulsed excitation.

5.3.4 ELLI camera

For EL investigation a highly advanced EQUUS 327NM EL camera

(IRCAM GmbH, Erlangen, Germany) equipped with an FPA InGaAs with a

spatial resolution of 640 x 512 pixels is used. The detector is cooled by a

Pelletier element and has a spectral response between 0.9 µm and 1.7 µm. The

frame rate is 100 Hz. As lens system a NIR objective with a focal distance of

30 mm and a transparency >90% in a spectral range between 0.6µm and 1.7µm

is used. Similar to the IR camera a trigger signal is provided by the EL camera

which can be used for lock-in measurements.

6. Photo-degradation of organic tandem solar cells

71

6 Photo-degradation of organic tandem

solar cells

Photovoltaic devices based on organic semiconductors (OPVs) hold great

promise as a cost-effective renewable energy platform because they can be

processed from solution and deposited on flexible plastics using roll-to-roll

processing. Despite important progress and reported power conversion

efficiencies of more than 11% the rather limited stability of this type of devices

raises concerns towards future commercialization. The tandem concept allows for

both absorbing a broader range of the solar spectrum and reducing thermalization

losses.

Recently, Anderson et al. presented first stability data of OPV tandem

devices stored under irradiated and dark conditions (see Figure 6.1). Therefore,

OPV tandem modules with efficiencies of about 1.7% with eight cells connected in

series were investigated. The stability data are presented in Figure 6.1.

*This chapter is adopted from ref [198] with the permission of the Royal Society of Chemistry


72

Figure 6.1: Operational stability of organic tandem solar modules in the dark (top) and light (bottom). The modules had a Ag/HC PEDOT:PSS/ZnO/MH301:PCBM/PFN/HTL PEDOT:PSS/ZnO /MH306:PCBM/PFN/HTL PEDOT:PSS/HC PEDOT:PSS/Ag structure and were processed in an roll-to-roll production line. Image was reproduced from [203]

Under Dark storage conditions, it can be seen that the shelf life time of the test

devices demonstrates maintenance of Voc over 800 h while the loss in PCE is

mainly governed by the decrease of Jsc. in the study it was assumed that the

losses in FF and Jsc might be due the formation of gradual traps at the ZnO site.

Under light, the degradation of all module parameters was observed to be

accelerated and life times of 12 h for T80 or 140 h for Ts80 were calculated.

However, current tandem OPV devices suffer from a rather poor long-term

stability and are far beyond to be competitive with standard solar cell technologies.

To understand the factors which limit the life time of OPV tandem devices an

enhanced study with respect to the photo-stability and the influence of UV light on

the device performance is presented.

6.1 Device fabrication and used materials

All solar cells used for this study are based on an inverted tandem structure

comprising sub-cells of the blends poly(3-hexylthiophene):phenyl-C60-butyric acid

methyl ester (P3HT:PC60BM) and diketopyrrolopyrrolequinquethio-phene:phenyl-

C70-butyric acid methyl ester (pDPP5T-2:PC70BM) as active layers. Due to its

known stability and its absorption spectrum being complementary to that of the


73

low band gap polymer pDPP5T-2 [196] (see Figure 6.2), the standard

P3HT:PC60BM as active layer of the bottom sub cell was used (see Figure 6.3).

Figure 6.2: Absorption spectra of the active materials used for light absorption in the tandem cell. The inset shows the chemical structure of the polymer pDPP5T-2. Reproduced with permission from [198]

As recombination layer zinc oxide (ZnO) in combination with poly(3,4-

ethylenedioxythiophene (PEDOT) was used. For charge carrier extraction

aluminum-doped ZnO (AZO) layer was used as interface layer between ITO

and active layer [72]. The interface between ZnO and pDPP5T-2:PCBM was

modified by coating a thin barium hydroxide (Ba(OH)2) layer on top of ZnO to

enhance the photovoltaic performance [204,205]. For a better understanding of

the degradation behavior, inverted single OPV cells based on the polymer

blends used in the tandem structure were fabricated and compared with the

tandem cell data. To obtain representative variations of device performance vs.

time five devices of each solar cell type were fabricated and encapsulated. To

ensure a reliable and reproducible protection along with a minimization of

extrinsic effects due to water and oxygen, a glass-on-glass encapsulation

geometry was chosen. The quality of the encapsulation was probed using ELLI

imaging before and after the photo-aging process (Figure A1). In the following


74

the fabrication parameters of the used tandem cells as well as the respective

sub-cells will be explained.

6.1.1 Tandem cell layer deposition

All photovoltaic devices were fabricated by doctor blading under ambient

conditions using an inverted device structure (Figure 6.3).

Figure 6.3: Schematic device representation of the tandem and single cells investigated in the present photo-degradation study. Reproduced with permission from [198]

Laser-patterned ITO coated glass substrates (area of 2.5×2.5 cm2) were

successively cleaned in an ultrasonic cleaner using acetone and isopropanol.

After drying, a 40 nm thick AZO layer was coated on top of the substrates,

which were then annealed for 10 minutes on a hot plate at 140 °C. For the

tandem solar cells, a chlorobenzene based solution of P3HT:PC60BM (1:1 wt%,

32 mg/ml in total) was coated on top of the AZO layer to form a 130 nm thick

bottom active layer. Subsequently, the intermediate layer was deposited by

successively blading a 40 nm thick PEDOT HIL3.3 and 30 nm thick ZnO layer.

These layers were dried at 70 °C for 5 min in air. The ZnO layer was modified

by coating a very thin (20nm) Ba(OH)2 film on top (7 mg/ml in 2-

methoxyethanol). Afterwards, an 80 nm thick layer of pDPP5T-2:PC70BM (1:2

wt.%, dissolved in a solvent consisting of 90% chloroform and 10%

dichlorobenzene with a total concentration of 24 mg/ml) was deposited as the

top active layer. At the end, a 10 nm MoOx layer and 100 nm Ag layer were

evaporated in order to form the top electrode. All fabricated solar cells had an

active area of 10.4 mm². The active layer thicknesses in the tandem structure


75

were tuned to guarantee photocurrent balancing between the sub-cells (see

Figure A2). The single solar cell devices were prepared in a way equivalent to the

tandem solar cell device fabrication protocol described above. For intimate

comparison, the thicknesses of the active layers were chosen to be the same as

in the tandem structure, i.e., 130 nm for P3HT:PC60BM and 80 nm for pDPP5T-

2:PC70BM single cell devices. Optical simulation results for the tandem solar cell

geometry based on the transfer matrix approach are depicted in the ESI (Figure.

A2).

6.1.2 Device Encapsulation

In order to neglect degradation due to extrinsic effects the devices were

encapsulated in a glass-on-glass encapsulation geometry. Therefore a

dispenser robot from I&J Fisnar Inc. (I&J 4100-LF) was used to distribute the

adhesive Katiobond LP655 from DELO GmbH & Co KGaA on top of the

completed OPV devices. After the adhesive deposition a second barrier glass

with a thickness of about ~100 µm was used to complete the glass on glass

encapsulation. Finally, the epoxy was cured for one minute inside a UVACUBE

100 from Hönle AG equipped with an iron doped lamp.

6.2 Photo-aging without UV light

6.2.1 Photo-aging of tandem cells

The tandem cells as well as the corresponding single sub-cells were

aged under continuous white light irradiation without UV component at

≈1000 W/m² (see section 5.1). The absence of UV light during photo-

degradation was essential for studying the effect of UV light treatment on the

lifetime of the devices. During photo-degradation the open circuit condition was

provided to the test cells. The initial device performance of the tandem cells and

the respective sub-cells under AM1.5G conditions was 4.4% ±0.2% (tandem),


76

2.8%±0.1% (P3HT), and 4.1%±0.2% (pDPP5T-2), respectively (Figure A3).

Representative JV characteristics of the investigated test devices can be seen

in Figure 6.4.

Figure 6.4: JV-characteristics of a representative Tandem and respective sub-cells measured at t=0 h. Reproduced with permission from [198]

To eliminate the double diode effect which is intrinsic to devices

incorporating ZnO and/or AZO [191,206], a single light soaking (LS) treatment

by irradiating the test cells with UV light (365 nm) for the time of 10 s (prior to

starting J-V characterization) was applied to the test cells. Afterwards the

subsequent decay of the UV light soaking state under continuous photo-aging

using white light was investigated.


77

Figure 6.5: Lifetime of the UV light soaking state under continuous photo-aging. The plots show the device parameters Voc, Isc, FF and PCE over time for the different types of single and tandem solar cells upon initial UV light soaking (365 nm, 10 s) and under continuous photo-aging using white light (400 – 750 nm) without UV component at 1000 W/m². UV light treatment was repeated after 60 hours. Each data point represents the average values of 5 solar cells and is normalized to the initial value at t = 0 hours. Reproduced with permission from [198]

Figure 6.5 shows two consecutive 60 h photo-degradation cycles for

elucidating the effect of the initial light-soaking step. These cycles are within the

burn-in period of photo-degradation, which typically follows an exponential

decay of the initial device efficiency (see also Figure 6.8) [196]. Within the first

cycle of continuous irradiation, the PCE of the tandem and the P3HT based

sub-cells decreased to 60% and 50% of the initial value, respectively. In both

cases, the loss in PCE is mainly dominated by losses in Jsc (≈15 – 20%) and FF

(25 – 40%), which can be most likely related to a reduction in charge carrier

extraction and the accumulation of carriers in the device, respectively [190,207].

The latter leads to the formation of an S-shape in the J-V-characteristics (see

Figure 6.6) [191]. The Voc, on the other hand, remains fairly stable. Interestingly,

under the same conditions the solar cell parameters and the overall PCE of the

pDPP5T-2 based single cells remained almost intact throughout the same

period of time. After 60 h the UV treatment was repeated. Notably, all solar cell


78

parameters were almost fully restored and followed the same degradation

pattern as in the first cycle. This behavior suggests that the burn in period that is

typically observed in the first hours of degradation is triggered by a reversible

reaction in these types of devices. A plausible explanation would be that the

conductivity of ZnO degrades with time due to the presence of traces of oxygen,

while UV treatment can release oxygen and restore its electronic properties

[22,190].

Figure 6.6: S-shape formation in the JV characteristic of a representative OPV tandem cell. The device performance was probed immediately upon UV light soaking (0 h, black curve), after 60 h under continuous photo-aging (red curve) and after repeated UV treatment (60 h, black dashed curve; LS means after light soaking). Reproduced with permission from [198]

6.2.2 Darkaging of OPV tandem cells

In order to gain an additional insight into the effect of the UV light-soaking

process, the transient behavior of the UV treatment in the dark was studied.

Figure 6.7 shows the periodically measured change of the photovoltaic

parameters for tandem cells upon a single UV light-soaking step. Here, the cells

were stored in the dark between JV characterization. Upon UV light soaking, the

FF increases dramatically (≈45%) while Jsc increases by about 5% and Voc

barely changes. It is well documented in the literature that UV radiation can


79

improve the electronic properties (conductivity) of the ZnO layer as well as the

contact at the ZnO interface [22,190,191]. This is most likely the reason for the

JV characteristics translating from a double-diode type behavior (S-shape) to a

diode behavior with high FF. The long-term behavior of UV treatment is much

less documented.

Figure 6.7: Long-term decay of the UV light soaking (LS) state in the dark. Each data point represents the average value of 5 tandem cells. The filled symbols represent the condition after immediate light soaking, whereas the hollow symbols represent the temporal decay of the LS state. The data were extracted from J-V-measurements using an AM1.5 solar simulator and an illumination intensity of 1000 W/m². The tandem cells were stored in the dark in between J-V characterization. Reproduced with permission from [198]

In Figure 6.7 it can be seen that the light soaking state remains constant

for about 10 h after which the photovoltaic performance decays sharply.

Moreover, Figure 6.6 reveals that the light soaking state features a half-time of

about 200 h and is, therefore, expected to contribute decisively to the burn-in

period of OPVs containing ZnO.

6.3 Photo-aging with UV lights soaking

From the previous result, the importance of UV light exposure during

continuous 1-sun irradiation can be inferred. Therefore, a long-term and photo-


80

aging test, in which the solar cells were exposed to UV light for 10 s prior to

each JV measurement, was designed. Figure 6.8 shows the average long-term

evolution of Voc, Jsc, FF, and PCE for the tandem cells and respective sub-cells

under continuous white light illumination with intermittent UV treatment.

Figure 6.8: Photo-aging of single and tandem OPV cells. The graphs show the average long-term temporal evolution of PCE, Voc, Jsc, and FF for the different single and tandem cells under continuous white light illumination. The photovoltaic parameters were extracted from J-V-measurements using an AM1.5 solar simulator at 1000 W/m². Before each J-V measurement the samples were UV treated (365 nm, 10 s). Each data point represents the average value of 5 tandem devices, 5 DPP devices, and 5 P3HT devices. Reproduced with permission from [198]

It can be seen that the burn-in period extends to ≈800 h, after which the

decay of the PCE follows a close to linear trend (Figure 6.8). Remarkably, in the

long-term measurements the tandem cells showed the most stable behavior by

losing only 11% of the initial value after 2000 h. The PCE of P3HT and

pDPP5T-2 based single sub-cells followed a similar decay with losses of 16%

and 15%, respectively (see Figure 6.9)


81

Figure 6.9: Relative change of device performance after 2000 h of continuous white light illumination. The relative change is calculated based on the performance of the single and tandem devices shown in Figure 6.8. The tandem cells showed better overall device stability. Reproduced with permission from [198]

Overall, the loss in PCE is mainly determined by a loss in FF (5 – 11%)

and a modest decay in Jsc (0 – 8%). Impressively, in the case of the tandem and

pDPP5T-2-based single cells almost no current losses throughout 2000 h of

light exposure were observed. To put the results into perspective, PCE drops in

the range of 10 – 20% for P3HT single cells and 25% for PCDTBT-based

single cell devices under 1-sun exposure and within similar periods of time

which have been reported in the past [19,67,196].

It is assumed that the reasons for the enhanced long-term stability of the

presented solar cells are manifold. Specifically, an inverted device structure was

used which enable elimination of reactive metal interfaces [208]. Additionally,

the widely used but moisture sensitive and reactive PEDOT:PSS was replaced

with MoOx as the top buffer layer. The benefit of using the chemically more inert

MoOx for better device stability has been shown before [209,210]. Moreover,

the ETL/pDPP5T-2 (ETL: electron transporting layer) interface was modified

with a hole blocking Ba(OH)2 layer resulting in more stable tandem and

pDPP5T-2 single sub-cells. Barium and Ba(OH)2 interlayers have been proven

to reduce exciton quenching and trap induced recombination at cathode

interfaces as well as improve the overall device efficiency in the case of OLEDs


82

and OPVs [204,205,211]. Based on the findings during the experiment, it is

assumed that Ba(OH)2 could also contribute to stabilizing the ETL/polymer

interface by reducing electronic trap formation and oxygen adsorption. Indeed,

comparison of the temporal evolution of photovoltaic device performance in the

case of P3HT:PCBM with and without Ba(OH)2 suggests that Ba(OH)2 may

increase the operating lifetime (see Figure 6.10).

Figure 6.10: Comparison of operating lifetime of P3HT:PC60BM cells with and without and additional Ba(OH)x interlayer. The Ba(OH)x was deposited on top of the ZnO ETL to reduce interfacial recombination center. See also Figure A4 and Figure A5. Reproduced with permission from [198]

Furthermore, periodic UV light soaking during prolonged operation is

expected to desorb oxygen trapped at the surface of ZnO and AZO. This step is

likely to prevent significant conductivity losses of the ETLs ZnO and AZO,

contributing to a larger FF throughout the lifetime measurements [190,212].

For an estimation of the lifetime of the tandem solar cells, a linear

regression to the slowly, linearly decreasing PCE data points and extended this

line up to 80% of the initial value was applied (see Figure 6.11).


83

Figure 6.11: Life time extrapolation from tandem PCE data presented in the experiment shown in Figure 6.7. The red line dependents and linear regression through slowly and linear decreasing PCE data points starting at t=1000 h. Reproduced with permission from [198]

In doing so, an extrapolated operating lifetime of 27000 h can be

extracted. Considering an average 1500 hours of sunshine per year in central

Europe, this represents, under the current conditions, a best case lifetime of

18 years. A more conservative lifetime for the test cells could be derived by

accounting for the error bars of the experiment, which still resulted in a lifetime

of 8 years. It is important to note that the presented operating lifetime was

extrapolated from cells, which were aged under open circuit and under indoor

conditions using a LED based solar simulator that does not emit radiation in the

180 – 400 nm wavelength range. The absence of UV light, which has been

shown to accelerate degradation through bond scission and free radical

formation in OPV semiconducting polymers, [27] may artificially increase the

lifetime of our cells. Furthermore, for outdoor conditions in the field, there are

influences from other sources such as natural thermal cycling, shading, and

humidity cycling, which need to be taken into account.


84

6.4 Conclusion

In summary an organic tandem cell with a PCE loss of only 11% within the first

2000 h of operation was demonstrated. The stable tandem operation was

possible by choosing active layer materials that can be processed in air and by

adopting an inverted device geometry, in which MoOx as a replacement for

PEDOT and Ba(OH)2 as hole blocking layer at the ETL/blend interface was

used. Moreover, it was confirmed that the well-characterized importance of UV

light treatment for devices based on ZnO is also an essential requirement for

attaining long-term device stability, which requires periodic light soaking by UV-

photons. This procedure mainly prevents the formation of S-shaped J–V

performance. While the current generation of organic tandem devices does still

require some UV light soaking, it is anticipated that the final OPV product will

not rely on UV light treatment. As such, future studies should primarily foster

materials development of new absorbers with enhanced photo and structural

stability as well as alternative electron transport layers that are not subject to

the requirement of photo-doping. Moreover, fully solution processed tandem

OPVs on flexible plastics with state-of-the-art encapsulation need to be

demonstrated for certifying market readiness. Assuming that the slow

degradation rate observed in this work can be further improved, organic tandem

solar cells with operating lifetimes comparable to traditional PV are thought

possible in the forthcoming future.

7. Temperature and moisture induced degradation of inverted and organic P3HT:PCBM solar cells

85

7 Temperature and moisture induced

degradation of inverted and organic

P3HT:PCBM solar cells

Solar cells and modules based on organic absorber materials have experienced

a steady increase in power conversion efficiency [7,9]. Record PCEs for single

OPV devices above 11% are reported and 21% are predicted [13,76,213,214].

The prospect of implementing fully solution based device fabrication steps

makes this solar technology uniquely suited for roll-to-roll printing applications

with potentially very low costs. In addition to the need for higher efficiencies, the

long-term stability of OPVs is essential for practical and economically viable

outdoor applications [198]. Investigating the underlying degradation

mechanisms and macroscopic defects in organic solar cells is thus an

indispensable condition for achieving competitive lifetimes [27,68,193].

*This chapter including all images is adopted from the paper „Water ingress in encapsulated inverted

organic solar cells: correlating infrared imaging and photovoltaic performance“. The Manuscript is

submitted (30th of May, 2015) to the scientific journal Advanced Energy Materials with the number of

submission aenm.201501065.


86

There are many aspects that can limit the lifetime of OPVs [27,166]. Most

organic semiconducting materials are highly reactive in the presence of oxygen,

water, high temperatures and light, which can induce thermal and photo-

oxidation reactions as well as local trap formation and structure variations in the

active layer [215]. In a device, adverse environmental conditions can lead to

morphological changes, compositional gradients, and electronic defects at

interfaces and electrodes [118,193].

While a clear correlation between observed performance loss and cause

is often difficult, the underlying degradation mechanisms depend critically on the

active materials and the device geometry.[68,193] It has been shown that

conventional device architectures employing low work function metals typically

break down of oxygen and water [171]. The diffusion of water preferentially

occurs through grain boundaries and pinholes, leading to corrosion phenomena

at the electrode and electrode/active layer interface, but also via the edges of

the device. As a result, a drop in short circuit current and fill factor due to loss in

contact area and the formation of resistive metal oxides is typically observed

[24,68,168,181]. To overcome this problem, high barrier encapsulation

materials with water vapor transmission rates (WVTR) much better than

~ 10 -3 g/(m²day) are commonly used in order to achieve module lifetimes of

several thousands of hours [216–218]. Importantly, the presence of water may

accelerate the aging process and could become a crucial degradation factor in

climates with high levels of humidity. Specifically, it is known that the presence

of water the fill factor of the JV characteristics due to S-shape formation, which

originates from deleterious chemical reactions at the active layer/interface.

In this study both the kinetics of water ingress and the primary reaction

site of water in inverted P3HT:PCBM solar cells with thick glass barriers are

investigated. In detail, the temporal evolution of the solar cell parameters Voc,

Jsc, FF, and PCE at the temperatures (relative humidity, RH) 7 °C (51%), 20 °C

(63%), 50 °C (20%), and 65 °C (20%) with a moisture variability of ± 5% was

investigated. Samples maintained at 7 °C were kept in a refrigerator while

samples maintained at 20 °C, 50 °C and 65 °C were stored in a heating oven

with controlled heating performance.


87

In a second set of experiments, the solar cell performance under

controlled temperature (65°C) and moisture (85% RH) settings by storing the

devices in a climatic chamber (CTS) was investigated. Temperature and

moisture fluctuations were kept below 1%. For each degradation setting, a set

of 10 cells was fabricated and averaged of the performance. All cells were kept

in the dark to minimize photo-oxidation reactions.

7.1 Test conditions and device fabrication

7.1.1 Organic solar cell preparation and characterization

All interface and active layers were processed in ambient atmosphere

using doctor blading. The test cells were prepared on ITO-coated glass

substrates (25 mm x 25 mm x 1.3 mm from Osram) with a sheet resistance of

5 Ω/sq [192]. The substrates were successively sonicated in acetone and

isopropanol for 10 min each. After blow-drying with nitrogen, the aluminum-zinc-

oxide (AZO) precursor was coated on top of ITO. Hydrolysis of the AZO

precursor was achieved by annealing the samples at 140 °C for 10 min [72].

Solutions of the photovoltaic blend layer components P3HT (Rieke Metals) and

PC61BM (Solenne BV) were prepared separately in chlorobenzene at a

concentration of 2 wt.% and stirred for 120 min at 60°C. Afterwards, the

P3HT:PCBM (1:1) blend solution was deposited on top of the AZO layer. As

electron blocking/hole extraction layer, a 100 nm thick PEDOT:PSS film was

doctor-bladed on top of the P3HT:PCBM film. The whole stack was annealed on

a hot plate at 140 °C for 10 min. A silver (Ag) film with a thickness of 100 nm

was evaporated as top electrode. The active area of one cell was 10.4 cm².

Figure 7.1a shows the device geometry and the final solar cell stack layout. For

encapsulation, a thick (0.7 mm) glass barrier and an ultra violet curable epoxy

adhesive from DELO (Katiobond LP 655) was used. For current-voltage

characterization, the test cells were removed from the climate chamber and the

measurements were performed in ambient atmosphere using an Agilent 2900A

source measure unit in a two-wire configuration. The cells were irradiated with a


88

light emitting diode (LED) solar simulator featuring an AM 1.5G spectrum at

1000 W/m² (FUTURELED GmbH, Berlin, Germany) [197]. The intensity of the

LEDs was calibrated with an ISE Fraunhofer certificated silicon calibration

photodiode.

Figure 7.1: a) Cross section of the inverted P3HT:PCBM devices used in this study. Water ingress is symbolized with red arrows. b) Schematic top view of a test device as shown in a).

In order to satisfy comparability between cells exposed to different

conditions a rigorous selection protocol was developed, in which solar cells with

a short circuit current density (Jsc) of less than 7.7 mA/cm², open circuit voltage

(Voc) of less than 0.58 V, a fill factor (FF) of less than 60, and PCE of less than

2.5% were discarded from the degradation studies. A representative JV curve

for one test solar cell that complies with the selection criterion can be seen in

Figure B1.

7.1.2 Imaging parameters and settings

During the degradation study electroluminescence (EL) and infrared (IR)

imaging based on ELLI, DLIT, ILIT as well as PL imaging was carried out. For

ELLI and DLIT measurements, two Equus 327k NM IR cameras (IRCAM GmbH,

Erlangen, Germany) equipped with a cooled indium-gallium-arsenide (InGaAs)

FPA detector (640x512) and a cooled indium-antimonite (InSb) based focal

plane array (FPA) detector (640 x 512), respectively, were used.

To avoid perturbations owing to temperature drift, the cells were allowed

to reach thermal equilibrium prior to applying a voltage. In order to minimize

implications due to the heat diffusion length the lock-in frequency was set to


89

7 Hz for all imaging techniques [219]. The image acquisition time for DLIT and

ILIT was 45 min, while the acquisition time for ELLI was 60 seconds. Solar cells

investigated with ILIT were measured under Voc condition and excited with

pulsed white light coming from an LED-array (FUTURELED GmbH, Berlin,

Germany). The spectral emission of the LEDs extends from 400 nm to 750 nm

with a spectral peak at around 440 nm. The light intensity was set to 1000 W/m²,

adjusted with an ISE Fraunhofer certified silicon calibration photodiode. Due to

glass thickness constraints the cells were irradiated from the front (ITO) side

and investigated from the backside for all ILIT measurements. Solar cells

investigated with DLIT and ELLI measurements were excited with a pulsed

voltage of 1 V (forward bias). In this case, the IR and EL radiation of the solar

cells was detected from the front side (see Figure B2). For measuring the PL

signal, the cells were scanned with an Argon ion laser using 488 nm wavelength.

The PL emission was detected by a germanium detector (ADC 403L) and

integrated over the wavelength range 610 – 840 nm. Details regarding the PL

setup have been described before [220].

7.2 Moisture induced degradation of inverted OPC

devices

7.2.1 Heat- and damp heat-induced degradation

In order to study the impact of moisture on the stability of the

encapsulated and inverted OPV devices, an experiment in which 10 cells stored

at 65 °C and 85% relative humidity (65 °C/85% RH) and 65 °C and 20% RH,

respectively, was carried out. For this purpose, the cells were stored in a dark

climate chamber with controlled relative humidity and temperature. The device

parameters at 65 °C/85% RH decay quickly within the first 300 h compared to

the reference measurement at 65 °C/20%RH (Figure 7.2). This fast decrease is

mainly due to the emergence of an S-shape deformation in the JV

characteristics of the test cells (see Figure 7.3). Analysis of the acceleration

factors by evaluating the slopes of the traces of FF (for details see further


90

below) reveals a faster degradation rate at Jsc and FF, for cells stored under

65 °C / 85% RH than compared to those stored at 65 °C /20% RH. Considering

that the temperature is the same in both experiments, the change in oxygen

concentration can be neglected. Taking into account the absolute values of

humidity dissolved in the air at 65 °C (136.2 g/m³ and 32.1 g/m³ at 85% RH and

20% RH, respectively) the result from Figure 7.2 strongly suggests that higher

content of moisture is mainly responsible for the fast aging process and S-

shape deformation in this type of devices.

Figure 7.2: Impact of relative humidity at 65 °C on the temporal evolution of Voc, Jsc, FF, and PCE for encapsulated P3HT:PCBM solar cells of inverted architecture.


91

Figure 7.3: Periodically measured JV characteristics of an inverted P3HT:PCBM solar cell stored in a climate chamber at 65°C/85%RH.

To study the thermal stability of inverted P3HT:PCBM solar cells under

varying moisture settings, a new experiment in which a set of test cells were

stored at different temperature settings was carried out. Figure 7.4 shows the

long-term behavior of glass-encapsulated P3HT:PCBM solar cells for different

storage temperatures in the dark. Each data set is normalized to the initial value

at 0 h and represents the average value from 10 test cells. All devices were

stored in the dark at room temperature for 45 min prior to JV characterization to

minimize temperature-induced variations in the JV performance.


92

Figure 7.4: Long-term behavior of P3HT:PCBM solar cells stored at different temperatures (7 °C/51%RH, 20 °C/63%RH, 50 °C/20%RH, 65 °C/20%RH) in the dark. Each data point represents the average value for 10 solar cells and is normalized to the initial value at t = 0 h. The solar cell parameters were extracted from J-V measurements under 1000 W/m² using an AM1.5G solar simulator spectrum.

The lifetime performance shown in Figure 7.4 can be categorized into

two regimes: a regime for T ≤ 20 °C where degradation is slowed down and a

regime of accelerated degradation for T ≥ 50 °C. Below 20 °C, a decelerated

decrease in device performance of the encapsulated solar cells is observed.

This is valid even after more than 28 months (20,000 h) of storage. For instance,

after 20,000 h and at T = 20 °C (7 °C), Jsc decreases by 15% (2%), the FF

drops by 17% (12%), and the overall PCE decreases by about 28% (15%).

Remarkably, the Voc remains almost unchanged for all temperatures,

suggesting that the morphology and recombination characteristics of the cells

are not significantly altered throughout the long investigation time. By applying a

linear regression to the data points of PCE for 20 °C (7 °C), a shelf lifetime of

about 2.2 (3.5) years is extrapolated (80% of the initial PCE, Figure B3). At

temperatures ≥ 50 °C, the photovoltaic performance is dominated by early

losses, mainly in terms of Jsc and FF, which decrease to 79% (78%) and 57%

(67%) of the initial value, respectively, for 50 °C (65 °C) after 6200 h (2100 h).

Overall, the PCE drops by about 52% after 6200 h for solar cells stored at 50 °C

and 46% after 2100 h for cells stored at 65 °C. At the end of the investigations,

nearly all JV curves reveal a distinct S-shape (Figure 7.3). This observation is

not fully understood and may result from several causes, such as the loss of


93

conductance in the ZnO layer [22,68] or moisture-induced aging [181,221]. It is

assumed that the fast decrease of device performance for T > 50 °C is mainly

related to the diffusion of water through the edges of the glass–glass

encapsulation. Ongoing from 50 °C to 65 °C, aging accelerates by a factor of

five, which is associated with a combined effect of temperature and moisture.

Note that the difference in oxygen concentration at this small temperature

difference can be considered negligible [222].

7.2.2 Determination of the activation energy

Based on the experimental data shown in Figure 7.4, the quantification of

the activation energy of the performance loss from the temperature depending

acceleration factors of the investigated test cells is attempt by applying the

Arrhenius model:

kdeg = A ∙ exp (−Ea

kBT), Eq. 20

where Ea is the activation energy (in eV) of the degradation mechanism,

kB the Boltzmann constant, T the temperature, and A a pre-exponential factor.

The reaction rate coefficient kdeg is obtained from applying a linear degradation

kinetics to the experimental Jsc data of Figure 7.4a [223]:

Jsc = Jsc(0) ∙ (1 − kdeg ∙ t). Eq. 21

Following a procedure by Brabec et al., the acceleration factor K for

degradation is extracted from the ratio of the degradation rate constants, i.e., by

combining Eq. 20 and Eq. 21, [223]:

K =k`

kdeg= exp [

Ea

kB(1

T−

1

𝑇`)]. Eq 22


94

Figure 7.5: Temperature dependence of the acceleration factor K for Jsc, FF, and PCE as extracted from the experimental data shown in Figure 7.4. The reference temperature T’ is 293 K.

Figure 7.5 shows the acceleration factor K = kdeg`/kdeg vs. 1/T, referenced

to T = 293 K, as determined from the temperature dependent losses of Jsc

(Figure 7.4 a). The same approach was applied for analyzing the temporal

decay of FF and PCE, giving rise to similar activation energies of 440 meV (Jsc),

470 meV (FF), and 460 meV (PCE). According to Figure 7.5, the degradation

occurs more than 10 times faster at T = 65 °C than at T = 20 °C. Generally, the

degradation rate is determined by a superposition of effects originating from the

presence of intrinsic (e.g., radicals, metal traces) and extrinsic (e.g., O2, H2O)

contaminants. Since water has been identified as the dominant degradation

reactant, Ea can be associated with the activation energies for diffusion and

reaction of water with solar cell components as well as with temperature

dependence of the amount of water dissolved in the encapsulated test cells,

which in turn comprise the factor of humidity and solubility. Previously, Brabec

et al. reported activation energies of 300 – 350 meV for encapsulated, MDMO-

PPV:PCBM based solar cells of regular architecture [223]. Kinloch et al.

reported activation energies for the displacement of different adhesives by

water in the order of 170 meV – 400 meV [178,179]. However, a direct

comparison is only meaningful for identical device geometries when measured

under the same storage conditions. Here, it is emphasized that the acceleration


95

factor can be primarily considered as a phenomenological measure for

establishing a stability metrics, comparing the optoelectronic robustness of the

material system and/or the quality of the sealing.

7.2.3 IR imaging of moisture-induced degradation at

65°C/85%RH

The data presented in the previous sections show the global change of

the solar cell specific parameters. However, it would be highly desirable to

correlate the loss in device performance with local changes in device behavior.

To further elucidate the effects of moisture diffusion on the local change of OPV

device performance different imaging techniques including ELLI, DLIT, and

photoluminescence imaging were applied on the same device area. All lock-in

images are represented as the 0° (in-phase) signal. Technical details regarding

the setup and the interpretation of the signals are presented in section 5 and the

Supporting Information (Appendix B).

At first the ELLI measurements of the test cells (Figure 7.6) will be

discussed. EL radiation was previously identified as originating from interfacial

charge-transfer (CT) state luminescence [226]. Here electroluminescence

imaging was possible using a sensitive indium-gallium-arsenide (InGaAs) focal-

plane array with a spectral response matching the CT emission of the

P3HT:PCBM blend (Figure B4)

At time zero, the emitted EL-radiation is nearly homogeneous over the

active areas of the cells. After ≈160 h of storage in the dark at 65°C/85%RH, the

emitted EL-radiation starts to bleach out from the outside towards the center of

the substrate while maintaining homogenous intensity in the remaining active

areas of the devices (see also Figure 7.7). Particularly, the three cells at the top

of the 160 h image show a strong and localized reduction of the emitted ELLI-

signal. The concentric damage is clearly discernible after a storage time of

> 500 h. At this point, only the center shows EL radiation while close to the

corners and edges the EL is almost fully suppressed.


96

Figure 7.6: ELLI images measured at different storage times using a pulsed voltage of 1 V (forward bias). The cells were kept in the dark at 65°C/85% RH.

In order to reveal if IR thermography might be a more sensitive tool than

ELLI, DLIT, ILIT, and ELLI images were measured (Figure 7.7) [227] of the

same solar cell as a function of storage time under 65 °C/85%RH.

In general DLIT and ILIT measurements reveal local thermal loss

processes which can be attributed to non-radiation recombination mechanism

induced by injected current or light. In both cases the DLIT and ILIT

measurements show a distinct intensity change in the degraded areas similar to

the EL disruption observed at the ELLI measurements. In addition to the loss of

EL and IR emission in degraded areas, a relative increase of the EL and IR

radiation in the unaffected parts of the cell is observed (Figure 7.7, top and

center). This is clearly apparent from line scans taken along the long side of the

active area of the solar cell (Figure B5). The increased signal might be

attributed to decreasing active area leading in an increasing potential difference

at the electrode of the unaffected cell parts. Hence, the bias point at the dark

characteristic of the non-defective cell parts is shifted to a slightly higher applied

voltage resulting in an increased injected current under the dark condition. The

result would be an increased ELLI signal of the cells. Note that the scratch

discernible in all images of Figure 7.7 does not affect the electrical properties of

the solar cell and that the disruption and increase of the ELLI signal is also

apparent in other devices (see Figure 7.6). Notably, ILIT shows an opposite IR


97

dissipation behavior compared with the DLIT and ELLI measurements, i.e., the

ILIT signal increases in the defective cell parts, which could be related to a

disruption at the electrode/active layer interface. It is generally possible that

charge carriers under open circuit conditions flow through the electrode to

recombine via the shunt resistance of the solar cell [14,228]. In case of a

disturbed interface, however, the electrodes are not accessible and the charge

carriers are likely to recombine in the active layer, leading to a local temperature

increase. Considering the notably higher signal to noise ratio of ELLI images at

relatively short acquisition times (60 s compared to 45 min for DLIT and ILIT),

ELLI emerges as the most promising technique for fast, in-situ imaging of OPV

cells and modules.

Figure 7.7: Top to bottom: ELLI, DLIT, and ILIT images of the same cell measured at different storage times. The sample was stored in the dark at 65 °C/85%RH. For ELLI and DLIT measurements the cell was biased in forward direction with a pulsed voltage of 1 V.

While the diffusion of moisture is apparent in EL and IR images (Figure

7.7), the signals can be affected by both, electro-optical changes of the

semiconductor material and the quality of the contact, making it difficult to

identify the layer where the damage actually takes place. To discriminate


98

between degradation effects occurring in the bulk active layer and at the active

layer interface ELLI images and PL scans of fresh and degraded solar cells

were recorded (Figure 7.8 and Figure B6). PL, when performed under open

circuit conditions is exclusively related to the light induced radiative

recombination behavior of the light induced exactions inside the semiconductor

[129]. The potentially large volume of active material sampled by diffusing

excitons, makes PL spectroscopy a very sensitive tool for the detection of

morphological changes or trap formation in bulk heterojunction layers. A change

in the PL signal can, therefore, be employed as probe for changes in exciton

and charge carrier dynamics in the bulk active layer.

Figure 7.8: Comparison of the electroluminescence (EL) and photoluminescence (PL) signal of a fresh and a degraded cell. The cells were kept in the dark at 65 °C/85% RH (see also Figure B7). PL was excited at 488 nm and the detection range was 610 – 840 nm (see Figure B4).

As opposed to ELLI, the PL map of the degraded solar cell remains

relatively homogenous over the whole active area (Figure 7.8 and Figure B7).

Importantly, the PL signal does not follow the same behavior as the ELLI signal.

Since the PL radiation originates directly from the active layer and is mostly

decoupled from influences related to the electrodes [229], it is concluded that

diffusion of moisture mainly affects the electrode/active layer interface, while the

P3HT:PCBM layer remains mostly intact.


99

For the solar cells shown in Figure 7.8, the active EL emitting area is

correlated with the photovoltaic performance as a function of the degradation

time (Figure B8). For the first 150 h the active EL area and the photovoltaic

parameters remain unchanged, corresponding to the diffusion time of water

through the rim of adhesive. After 150 h, moisture starts to affect device

performance and the ELLI area starts to drop rapidly, even faster than the

photovoltaic parameters. Interestingly, there is little change in Jsc with

decreasing active ELLI area, while Voc and FF decrease noticeably with

degradation time. This suggests that the traces of water diffusing into the solar

cell and turning off EL emission do barely affect net charge generation and

charge carrier extraction. Furthermore, a close to constant Jsc with simultaneous

decrease in Voc indicates that the work function of the affected contact is likely

to undergo change under damp heat conditions.[192,230] Recently, Reinhardt

et al. showed that a change in work function, more specifically, a reduced

charge selectivity at the contacts due to surface recombination, can lead to a

concomitant decrease of Voc and EL intensity [99]. This is consistent with our

observations and further suggests that the degradation is occurring at the

contact rather than in the bulk active layer.

Overall, comparison of electroluminescence imaging and photovoltaic

performance during accelerated temperature and humidity testing of OPVs

reveals that ELLI features much better sensitivity for predicting moisture-

induced device failure than I-V measurements.

7.2.4 Moisture diffusion in encapsulated devices

As shown in the previous section, the presence of moisture at the

semiconductor/electrode interfaces is responsible for the observed S-shape

formation in our organic solar cells. To find countermeasures for preventing

moisture-induced failure modes in encapsulated organic photovoltaics, it is

important to understand quantitatively the transport of moisture through the

packaged device. In the following, it will be shown that classical diffusion theory

is applicable for describing isothermal moisture transport upon side ingress of

water into the packaging of our samples. To that end, encapsulated large area


100

(1.44 cm²) OPV devices and Ca corrosion tests of the same geometry were

prepared and degraded at 65 °C/85% RH (Figure B9). The diffusion of water in

the calcium samples is measured optically by monitoring the transition of the

films from opaque to transparent, while in the case of OPV devices detecting

the decrease of the electroluminescent area using ELLI allows to analyze the

moving waterfront.

In order to understand the kinetics of moisture diffusion inside

encapsulated OPV devices the diffusion coefficient D for water in the adhesive

was calculated by applying Fick’s second law of diffusion to a calcium test:

c(x, t) = cs ∙ (1 − erf (x

2√Dt)), Eq. 23

where cs = c(x=0,t) is the saturation concentration and erf(x) is the Gauss

error function. The diffusion coefficient reflects the speed at which the moisture

diffuses through the adhesive. Equation 23 can be rearranged to solve for D:

D = (dx

d√t⁄

2∙erf−1(1−ckcs))

2

, Eq. 24

The concentration ck represents the critical concentration of water

necessary for consuming the calcium film and was estimated to be 4∙10-4

mol/cm3 for a 100 nm thick film. The saturation concentration cs was measured

gravimetrically to be 3.2×10-3 mol/cm³ (Table B1). The thickness of the adhesive

film was 15 µm. The calcium test (Figure 7.9 and figure B9), which featured the

same encapsulation architecture as the test devices, shows that the position x

of the waterfront as a function of √t follows a linear behavior, as expected from

Fick’s law, with dx/d√t = 3.3×10-6 m/√s. From the optical calcium test

D = 2.1∙10-12 m²/s using Eq. 24 is extracted. Alternatively, D is calculated using

the water vapor transmission rate (WVTR) of the adhesive as provided by the

manufacturer (Appendix B), giving rise to D = 1.4×10-12 m2/s, which matches the

value from the calcium test closely. Similar diffusion coefficients for water in

lamination adhesives have been recently presented by Michels et al. [231] ELLI

images of large area OPV devices show a similarly linear x vs. √t, also obeying


101

Fick‘s law of diffusion. Determination of the diffusion coefficient of water in the

device requires specific knowledge of ck or WVTR in the case of a full device

and was outside the scope of the present study.

The value of the diffusion coefficient is important when considering the

water permeation path inside encapsulated devices. To test whether the

degradation kinetics is mainly determined by the diffusion of water through the

adhesive or through one of the active layers of the device solar cells with grid

finger electrodes as well as with different thicknesses of silver were prepared

(Figure B11). Metal electrodes may offer barrier protection against vertical

diffusion when water is primarily transported via the adhesive. ELLI images of

devices with grid finger electrodes show that the EL signal emerging from areas

in direct contact with adhesive and from areas protected by finger electrodes

deteriorates with very similar kinetics, suggesting that the waterfront progresses

through one of the active layers of the device stack under the silver electrode

(Figure B12 and B13). Additionally, the observation that EL degradation

patterns for solar cells with thick silver metal electrodes resemble those with

thin electrodes further eliminates the possibility of water diffusing through the

electrodes (Figure S14). Considering the device geometry and previous reports

using similar device layouts, PEDOT:PSS emerges as the most likely candidate

for efficient transport of moisture in inverted devices [178,180]. In fact, Feron et

al. recently measured the diffusion coefficient of water in PEDOT:PSS to be D =

5.0×10-10 m2∙s-1, i.e., approximately two orders of magnitude larger than D for

our adhesive [182]. To confirm this hypothesis and identify the site of

predominant degradation due to damp heat full devices and samples missing

PEDOT:PSS layer and top electrode (i.e., ITO/ZnO/P3HT:PCBM) were exposed

to damp heat conditions (Figure S15). The results show that the degradation

effects are strongest in complete devices while only minor impact is observed in

the case of ITO/ZnO/P3HT:PCBM samples with fresh PEDOT:PSS/silver top

electrode. Since there is no indication of corroded silver electrodes or silver

interfaces it is conclude that degradation of the PEDOT:PSS contact is primarily

responsible for early device performance losses under damp heat conditions.

Finally, assuming that the linear behavior shown in Figure 7.9 is valid at

longer times, the shelf life of an OPV device for a given adhesive geometry and

environment by extrapolating the kinetics of water transport in the protective


102

adhesive can be predicated (Figure 9). By extending dx/d√t, a rim of adhesive

of 9.3 cm would be sufficient to achieve a lifetime of t≈100,000 h under

65 °C/85%RH.

Figure 7.9: a) Water permeation model for inverted and encapsulated organic solar cells based on ITO/Al-ZnO/P3HT:PCBM/PEDOT:PSS/Ag. b) Experimentally determined kinetics of water diffusion in epoxy adhesive using calcium test: penetration distance vs. t1/2. The error bars are

approximated reading errors. The slope dx/d√t is dx/d√t = 3.3∙10-6 m/√s. Similar results were

obtained with large area calcium tests (Figure B10)

7.3 Conclusion

Considering that the diffusion of moisture during manufacturing, storage,

or operation is thought to be one of the major reliability concerns in

encapsulated organic electronic devices, identifying the underlying degradation

mechanisms represents a prerequisite for developing strategies towards long-

lived organic photovoltaics. Here, it was shown that by combining spatial


103

information on charge transfer luminescence and charge carrier recombination

with JV measurements as a function of controlled aging conditions important

insight into possible degradation paths in OPVs can be gained. By investigating

the temporal JV performance under controlled temperature and moisture

settings water was identified as the main promoter for performance loss in

inverted P3HT:PCBM based solar cells. The degradation related acceleration

factors and the activation energy for the diffusion of water were quantified by

applying the Arrhenius model (≈450 meV). Furthermore, it was shown that

selective electrical and optical excitation using lock-in imaging in the IR allows

to visualize the loss in active area due to the permeation of water and,

importantly, to distinguish between degradation occurring in the bulk

heterojunction and at the electrode/active layer interface. By comparing ELLI,

DLIT, ILIT and PL measurements it was proved that the presence of moisture

mainly affects the local electroluminescence and the active layer electrode

interface of inverted organic solar cells. This suggests that in the presence of

moisture charge carrier extraction is most likely inhibited, which could be one

possible cause for the temporal decay in photovoltaic performance.

Under damp heat PEDOT:PSS was shown to degrade more faster than

ZnO. The hypothesis is that water upon diffusing through the rim of adhesive is

quickly transferred to the PEDOT:PSS layer at the contact area between

adhesive and PEDOT:PSS. The water diffusion then progresses via the PEDOT

layer. As a result, the charge carrier extraction is most likely inhibited in area

affected by water ingress, which could be on reason of S-shape formation and

temporal decay of photovoltaic performance. This mechanism is expected to be

valid as long as the diffusion coefficient of water in the adhesive is much more

smaller than the diffusion coefficient of water in the PEDOT:PSS layer.

Furthermore, the diffusion of water can be approximated using Fick`s law of

diffusion

Finally, among the IR imaging techniques studied in this work, ELLI

represents the fastest and most responsive optical imaging modality. It thus

bears the potential to be employed as a standard characterization technique for

both studying the local radiative recombination behavior and the quality of

barrier materials in organic solar cells. This result further strengthens the

importance of IR imaging techniques not only as a quality control tool but also


104

as a means for localizing, studying and preventing the origin of local

degradation effects in organic solar cell performance, particularly when

combined with spectroscopic information.

8. The electrical and thermal characterization of macroscopic defects in thin film solar modules

105

8 The electrical and thermal

characterization of macroscopic

defects in thin film solar modules

Due to the opportunity to print organic solar cells on light weight and flexible

substrates, OPV devices bear the potential for a cost-efficient mass production.

In order to guarantee a successful commercialization of OPV modules, these

devices need to fulfill different requirements which focus on the efficiency, long-

term stability, and fabrication. With efficiencies above 10% OPV cells already

proved their potential to be competitive with standard thin film technologies

[232]. Furthermore, several groups as well as the findings presented in

section 6 show that life times of several years are possible [196,233]. In terms

of manufacturing, the main challenges are mainly related to upscaling (transition

from small area single solar cells to large area solar modules) [234]. Here, the

introduction of performance reducing defects due to improper device fabrication

is one of the major concerns of manufacturers. The amount of the performance

reduction depends predominantly on the defect type, defect size, and defect

position in the cell as well as the illumination conditions [14]. Especially, under

low light irradiation condition (< 100 W/m²) even “weak” defects may lead to a

pronounced loss of electrical cell output.

* This chapter including all images beginning from Figure 8.2 is adopted with permission from ref

[129] (Copyright Elsevier)


106

During the last decade different thermographic methods have been developed

to characterize several thermal losses in thin film solar cell technologies. ILIT

measurements have been proven to be sensitive enough for a contactless, fast,

and non-destructive characterization of defects in organic solar modules [32].

The possibility of a contactless module characterization during different

production steps and allows an early sorting of defective devices and an

improvement of the production line’s throughput rate. To do so, it is crucial to

have a firm physical and electrical understanding of defects and their impact on

the surrounding cell. Especially a highly advanced analysis of both the

quantitative influence on device performance and quantitative defect

parameters such as resistance are essential. Recently, Besold et al. presented

a study in which randomly distributed and fabrication related defects in OPV

modules were assigned with a distinct ohm value [122]. From JV characteristics

at different light intensities, the group demonstrated that bulk heterojunction

solar cells exhibit a photo shunt, which means that the overall value of the

parallel resistance depends reciprocally on light intensity and decreases with

increasing irradiation. Cells containing one or more defects did not followed this

reciprocal trend. The defects could be, therefore, attributed to purely ohmic.

Based on ILIT measurement at different irradiation conditions and 2

dimensional LT spice simulations each defect was associated with a distinct

ohm value in which the IR emission of a defect was set in relation with its

simulated electrical power dissipation.

The authors mentioned that the findings of the study can be easily transferred to

other thin film solar technologies and clarify the importance of quantitative

analyses of defect related losses in solar modules. However, organic solar

modules are still in the pre-commercialization stage and large improvements

need to be done for market readiness. Therefore, for an efficient and

quantitative analysis of defects, one needs to focus on already highly advanced

thin film solar technologies. Especially solar cells and modules based on

CuInGaSe2 (CIGS) have been successfully developed into high performance

power converters of solar energy with a global market share of about 2% [4].

These devices belong to the highest advanced thin film solar technologies and

show among all other technologies the currently highest values of PCE for

single cells (21.7%) and modules (15%-17%) [10–12]. Similar to OPV devices,


107

CIGS solar modules suffer from a large variety of fabrication related defect

which can be possibly introduced to the module during the different layer

deposition steps. The end result is that defects cause a decrease in the module

electrical output. Therefore a firm physical and electrical understanding of these

defects is essential to improve both module efficiency and fabrication.

8.1 Description of test modules

The investigated test samples consisted of 15 non encapsulated CIGS modules

with 67 solar cells connected in series. The modules were fabricated at a large

industrial production line for CIGS modules and had a size of 28 cm x 28 cm.

The electrical power output (Pmpp) of each module was 9-10 W at 1000 W/m²

with an irradiation spectrum of AM 1.5G. A photograph and an ILIT

measurement of one test modules can be seen in Figure 8.1

Figure 8.1: a) Photograph of a 28 cm x 28 cm CIGS test module with 67 cells connected in series. b) ILIT measurement (amplitude signal) of the test module presented in in the photograph. The brightest spot (marked with arrows) indicates defects which reducing the maximal power output of the module c) Schematic cross section of a CIGS solar module. The interlayer between the ZnO window layer and CIGS absorber layer represents the CdS buffer layer. Illustration is simplified and not in scale. Image a) and b) are reproduced and modified from [120]


108

At this point only a short and summarized description of the device fabrication

will be given. A detailed description of the different production steps can be

found in [38,235–238]. As back contact of the individual cells, a molybdenum

(Mo) layer was used, which was sputtered on a 3 mm thick glass substrate

(soda-lime glass). To separate the back contacts of the cells, a first patterning

step (P1) was realized with a laser scribe. Next, the CIGS absorber layer was

deposited by co-evaporation on top of the Mo back electrodes. This was

followed by a buffer layer of CdS deposited using a chemical bath deposition

process. A second patterning step (P2) ensures the series connection between

the cells by mechanical scribing down to the Mo layer. The front contact was

then deposited by sputtering subsequently an intrinsic zinc oxide (ZnO) and an

aluminum doped zinc oxide (ZnO:Al) window layer. A final patterning step (P3)

separates the cells by mechanical scribing again down to the Mo layer (see

figure 8.1 c).

8.2 Influence of a defect on its surrounding cell

measured by ILIT

Figure 8.2 shows a magnification of a representative defect from one of the test

modules, visualized by an ILIT-Voc measurement. According to the lock-in

thermography principles it is possible to display four images: the 0° image (in-

phase), the -90°image (out of phase), the amplitude image A(x,y), and the

phase image Φ(x,y). Details about the different signals obtained by DLIT and

ILIT measurements can be found in Breitenstein et al [128].


109

Figure 8.2: A defect (bright spot) identified by an ILIT-Voc measurement (illumination power of 30 W/m², 1Hz lock-in frequency, measuring time of 10 min): a) 0°-image, b) -90°-image, c) amplitude image and d) phase image. Reproduced with permission from [129]

The small vertical and dark lines which can be seen in Figure 8.3 a), b),

and c) are the patterning lines while the bright spot represents the defect. The

vertical cell region around the defect shows a lower IR-emission compared to

non or less-defective regions. This region appears darker because of the

attraction of light induced charge carriers as described in section 3.3. It can be

seen from the images a), b) and c) in Figure 8.3 that the IR-emission of the

adjacent neighbor cells is slightly increased (“cross talk”) when compared with

to the outer neighbor cells. This behavior can be attributed to a charge carrier

injection from the defect into the adjacent neighbor cells [14] resulting in an

increased non-radiative recombination of charge carriers in the neighbor cells

around the defect.

In order to qualitatively confirm the influence of the defect on its

surrounding cell and the recombination behavior of light induced charge carriers

Figure 8.3 a) shows a referenced line scan over the defective cell presented in

Figure 8.2. The ILIT signal was referenced to an ILIT line scan of an ideal

neighbor cell, resulting in a referenced IR-emission Sdiff(x) = S(x) – Sreference(x),

with x as the position in the cell. The reference cell was chosen as close as

possible to the defective cell in order to minimize both the spatial inhomogeneity

of illumination and the local variations of material properties. However, to avoid


110

referencing with an affected neighbor cell (“cross talk”), the second ideal

neighbor cell was chosen as the reference cell.

Figure 8.3: a) Line scan (Sdiff) of a defective cell (amplitude image) in a CIGS solar module (see Figure 8.2). b) Simulation of the local electrical power density of a CIGS cell with a defect.

The cell length of 28 cm is equivalent to 475 pixels of the camera

detector. This results in a spatial resolution of about 0.6 mm x 0.6 mm per pixel.

Figure 8.3 a) shows a defect as a distinct maximum caused by highly localized

IR-emissions at 18 cm. As can be seen in the graph within the first 6 cm (Figure

8.3 a) left part), the IR-emission is approximately zero meaning that there is no

difference between the IR-emissions of the defective cell and the reference cell.

In these regions the heat generation is mainly induced by thermalization and

trap state assisted recombination and not influenced by the defect. Continuing

after the first 6 cm the influence of the defect becomes noticeable which is

reflected in the negative values in Figure 8.3 a). The IR-emission decreases

continuously towards the defect, where a large peak exists (joule heating). The

data from the experiment shows a similar behavior (a distinct peak and slightly

negative values for the signal around the defect) when compared with a line

scan of the simulated electrical power density of a cell with an embedded defect

(Figure 8.4.b). A closer description of the cell parameters and simulation model

used for the simulations can be seen in the appendix and in reference [14],


111

respectively. The simulated ohmic defect had an area of 0.001 cm² with a

resistance of 0.1 Ω∙cm² and was placed at the same position as in the

experiment. It can be seen that electrical power density shows a similar non-

symmetric behavior to the data found from experiment, indicating that the

applied model incorporates the main processes for heat dissipation. Here, it

should be mentioned that the aim of the simulation was not to calculate the IR-

emission precisely but rather to gain a better understanding of the dissipation

processes in dependency of the charge carrier generation.

In a further study the influence of light on the impact of a defect on its

surrounding cell was investigated. Therefore Figure 8.5 shows two line scans

over a defective cell with an embedded defect close to the cell center. The two

line scans show that with increasing light intensity from 50 W/m² to 380 W/m²

the influence length decreased from around 10 cm to 5 cm. Especially, under

low light irradiation conditions the influence of the defect goes over the entire

cell length.

Figure 8.4: Line scans of a defective cell (amplitude image) in CIGS solar module irradiated with different light intensity. The module was investigated for both experiments with a lock-in frequency of 1Hz and a measuring time of 10 min.

This decreasing influence of a defect in dependency of increasing light

intensity can be attributed to a shunt screening behavior which is explained in


112

detail in the section 3.2 the references [14,120]. DLIT and ELLI measurements

in which the injected current was adjusted to the respective photocurrent of the

ILIT measurement showed the same behavior (see Appendix C). One

conclusion from these results is that the defects in thin film modules have an

influence region from where they attract electrons. This influence region can be

quantified by an influence distance. However, defects are generally randomly

distributed on the solar cell. As the influence distance is rather large if the

sample is irradiated with lower light intensity (<100 W/m²), the evaluation of the

influence distance is ambiguous. Determining the influence distance also

depends on various factors such as defect position, number of defects in a cell,

and defect type (ohmic and non-ohmic). Small differences in those factors make

the finding of the influence distance difficult and in some cases impossible.

8.3 Relation between Voc,cell and IR- emission of a defect

In order to understand the dependency between thermal power

dissipation and defect resistance a thermal-electric simulation was performed.

The simulation based on the same model used for the simulations of Figure 8.4.

In a first step the temperature distribution T(x,y) over the entire defective cell for

each defect parameter was simulated. In the second step, this temperature

distribution was converted into a power distribution PSB(x,y) (Stefan-Boltzmann).

For all simulations the ambient temperature was set to 300 K. The power

dissipation was then integrated over its area of the defect and set in relation to

the simulated cellular open circuit voltage. Therefore the Voc,cell was defined as

the average voltage between the front and back contacts divided by the whole

area of the simulated cell.


113

Figure 8.5: Simulated Voc,cell, current through the defect (defect current), and power dissipation of the defect depending on the defect resistance provided a defect size of 0.001 cm². To compare the different simulations the values of Voc,cell, defect current, and power dissipation are normalized to their maximum value. Reproduced with permission from [129]

Figure 6.5 shows the normalized electrical current flow through the

defect, the voltage of Voc,cell and the power dissipation of a defect in

dependency of the defect resistance. For high resistances, the current flow is

relatively low. This results in a high cellular voltage and relatively low power

dissipation. With decreasing resistances, the current and the power dissipation

increase until a maximum is reached. For relatively low resistances the current

is constant (defect current saturation) and the power dissipation is mainly

dominated by the decreasing voltage. The result is a low cellular voltage and a

low heat dissipation of the defect.

In the experiment the IR-emission of a defect was found by integrating

the IR-signal over its area. For the IR signal evaluation the -90° signal of the

ILIT investigations was used. The area of a defect was defined as the dark

region inside the white ring, as shown in the phase image (Figure 8.2 d) [128].

The phase signal shows the area where the IR-signal of the defect is well in

phase with the light excitation (black dot), slightly in phase with the light

excitation (white ring) and not in phase at all (noisy background).Then the IR-

signal of the defect was integrated over this area and compared with the cellular


114

voltage with the method explained in section 5.1 and demonstrated in section

8.2. In Figure 8.6 the data from the experiment is summarized. For the

experiment, the defect’s IR-emission and Voc,cell of 220 cells from 15 CIGS thin

film modules was measured. Each cell contained one or more defects. In order

to estimate the drop of Voc,cell caused by a defect, the Voc,cell of multiple ideal

cells in each CIGS test module were measured.

a)

b)

Figure 8.6: a) Normalized open circuit voltage of different cells vs. the defect IR-emission. b) Simulated open circuit cell voltage depending on the defect’s power dissipation calculated using the Stefan Boltzmann law for different defect resistances and two defect sizes. The black arrow indicates the defect resistance varied from high to low. Reproduced with permission from [129]


115

As shown in Figure 8.6 a) the maximum Voc,cell of the modules may vary

as the material compositions of different modules are generally not completely

identical. For better comparability, the Voc,cell for every module was normalized

using the largest measured Voc,cell of a non-defective cell. The relative error of

measuring the Voc,cell was previously determined of 3%. The IR-emissions of

defects in one cell had been summarized for all cells containing more than one

defect. Figure 8.6 a) shows that cells with a relatively high Voc,cell contain defects

with a relatively low IR-emission. With decreasing cellular voltage the IR-

emission increases. This may be explained by a decreasing resistance from the

defects (see Figure 8.6). A IR-emission maximum is reached at around 15 -

20% loss in Voc,cell. Beyond this maximum, the IR-emission decreases along

with a further loss in Voc,cell. This reduction, despite the severe loss in Voc,cell, is

due to a very low resistance for such defects. As a result, the IR-emission is

also low. However, the data points are widely scattered. Defect parameters

such as size, type, and position may influence the decrease in Voc,cell and IR-

emission. In Figure 8.6a), these possible influences are not considered. To

show the influence of the defect size on the cell, the power dissipation of defect

two different sizes was simulated using the same model as shown in Figure 8.6.

The graph presented in Figure 8.6 b) resembles the behavior of the data of the

experiment. For high resistances, the IR-emission of a defect and the loss of

Voc,cell increases with decreasing resistance until a maximum is reached. From

both graphs presented in Figure 8.6, a clear relation between the IR-emission

and the loss of Voc,cell for defects with high resistances can be seen.

Interestingly, from the experimental data over 95% of the investigated defects

induces a drop of Voc,cell less than 20%. This behavior is changing for

intermediate and stronger (lower resistance) defects. As the defect’s size was

not taken into account the influence of size may be responsible for the

scattering of the Voc, cell. When more defective cells are considered, the relation

of the defect IR-emission and Voc,cell will be improved.

However, the data from the experiment as well as the simulation show an

ambiguous relation between IR-emission and defect induced drop of Voc,cell.

One way to avoid the ambiguous behavior is to irradiate the modules with a

higher illumination power. Further simulations of Frank Fecher and co-workers


116

[14] revealed a spatial dependency of defects in the cell on the Voc,cell and their

IR-emission. Therefore the power dissipation of a defect and cellular voltage at

different defect positions in the cell (data not shown) were compared. One

conclusion from the simulation was that the defect position plays an important

role for the quantitative interpretation of the influence of a defect on its

surrounding cell. Nevertheless, defects in industrial produced CIGS thin film

solar modules are in generally randomly distributed over the whole module

without control of size, resistance, and type. In order to investigate a detailed

dependency of the defect position on its IR-emission and Voc,cell, a reliable

method to artificially create a defect is needed for an experimental proof.

8.4 Conclusion

In this study it was shown that ILIT measurement can be used to characterize

different light induced loss processes of charge carriers in thin film solar cells

and modules. The investigated modules were fabricated in a large industrial

production line for CIGS modules. Besides the detection of defect induced hot

spots, a precise localization of these and the characterization their influence on

the surrounding cell under the Voc condition was possible. This influence

strongly depends on the irradiation power and the sheet resistances of the

electrodes. It was shown that the influence on the surrounding cell shrank with

increasing irradiation power resulting in a distinct shunt screening behavior

[120]. The investigation of 220 defective cells under low light irradiation

revealed that the IR emission of a defect was found to be increasing for a

relatively small decrease in cellular voltage until a maximum was reached

(around 15% – 20% of the Voc of a healthy cell). Beyond this maximum the IR-

emission decreased with decreasing cellular voltage. Computer simulations

confirmed this behavior and showed that this was due to a saturation of current

through the defect. Hence, the influence of a defect can be classified into weak,

intermediate, and strong. It was shown that the majority of the investigated

defects (95%) only induced a drop of Voc-cell less than 20%. The characterization

of weak and intermediate defects showed a clear relation between IR emission

and Voc,cell while for strong defects the data of the experiments are scattered.


117

Therefore the characterization of strong defects using one ILIT measurement is

only limited and can lead to misinterpretations. This limitation might be

overcome using high irradiation powers (~1000 W/m²) during the ILIT

experiment, since the thermo-electric influence of a defects strongly depends on

the irradiation power. However, ILIT imaging enables a fast, contactless, non-

destructive characterization of solar cells and modules. Especially under the Voc

conditions ILIT measurements can be easily integrated as quality control tool in

large industrial production lines for PV modules. Since the presented

experiments show basic loss processes of light induce charge carriers the

presented finding can be transferred to other solar cell technologies.


118

9. Summary and Outlook

119

9 Summary and Outlook

9.1 Summary

During the last decade, different thin film solar cell technologies based on

organic or inorganic semiconductors have been developed to high efficient

energy converters. These cells are made of different thin films with thicknesses

ranging from several hundred nanometers up to several micro meters. However,

an improper layer deposition as well as different degradation phenomena might

result in the formation of defects, leading to a decrease of electric output and

stability. Therefore, it is crucial to have a firm physical and electrical

understanding of defects and different degradation phenomena in order to

improve their reliability.

The experiments in this study were designed to improve the current

understanding of defect and degradation related loss processes found in thin

film solar cells. To achieve this, local cell information about different loss

processes provided from imaging methods based on lock-in such as ILIT or

ELLI were set in relation with electrical characterization experiments of the

investigated test samples. The combination of imaging experiments and

electrical characterization experiments as a function of different degradation

stages especially helped to enable the characterization of several degradation

paths within organic solar cells.

The first part of this thesis presents organic tandem solar cells with power

losses of around 11% within the first 2000 h of operation (chapter 6). The long

operation mode was achieved by choosing an inverted device geometry, in

which MoOx was used as a replacement for PEDOT:PSS. Furthermore, the

ETL/active layer interface was modified by adding a Ba(OH)2 layer to improve

PCE and charge carrier extraction. During the investigation, the importance of


120

UV light treatments for OPV devices comprising a ZnO interlayer was confirmed.

Without the UV light treatment, the test samples showed a rapid S-shape

deformation which resulted in a decreased efficiency under continuous

irradiation. With similar degradation studies on the respective sub-cells (single

junction cell), the origin of the S-shape deformation could be investigated and

attributed to one of the sub-cells of the tandem cell. The investigations

presented in this study revealed that periodic UV light treatments represent an

essential requirement for achieving long term device stability for OPV samples

compromising a ZnO interlayer.

The second part of this study focuses on the investigation of moisture

induced degradation phenomena in encapsulated and inverted organic solar

cells based on a P3HT:PCBM active layer (chapter 7). The test cells were

stored in different environmental conditions with controlled temperature and

moisture setting. With more than 20,000 h of storage time one of the longest

lifetime studies reported in literature was carried out. During different

degradation experiments it was identified that water is the main promoter for

performance loss in the investigated test cells. The quantification of the

degradation related acceleration factors reveal an active energy for the diffusion

of water of about 450 meV. The use of IR and EL imaging techniques visualized

the loss in active area due to the diffusion of water through the device. With the

comparison of electroluminescence lock-in measurements and

photoluminescence measurements, it was shown that the presence of moisture

mainly affects the local electroluminescence and the active layer/electrode

interface. Furthermore, the use of different Ag electrode geometry identified that

the lateral diffusion of moisture is mainly through one of the cell interlayers.

In the third part, this study focuses on the local characterization of the

thermo-electric influence of macroscopic defects in CIGS solar modules

(chapter 8). Therefore, 15 test modules with 67 cells connected in series,

provided from a large industrial production line for CIGS modules, were

investigated. Each module contained several fabrication related defects. The

modules were IR characterized using illuminated lock-in thermography under

both low light and the Voc conditions. It was shown that each defect induced an


121

influence region in which power is extracted from the cell. To combine the IR

emission of a defect with the local module voltage (cell voltage) a new method

of voltage determination for cells in a module was developed. Investigating 220

defective cells revealed that the IR emission of a defect increases for a

relatively small decrease in cellular voltage until a maximum is reached (around

15% – 20% of the Voc of a non-defective cell). Beyond this maximum the IR-

emission decreases with decreasing cellular voltage. Computer simulations

confirmed this behavior and attributed that this was due to a saturation of

current through the defect.

9.2 Outlook

The presented investigations in this study impressively show the

importance of high advanced device characterization in order to understand

different degradation phenomena and defect induced loss processes in thin film

solar cells and modules. It was shown that different IR or NIR imaging

techniques can be used to provide a complementary insight into several local

aging processes. A firm and physical understanding of the origin of different

loss processes is essential to improve the reliability of PV devices and opens

the way for a future commercialization.

The investigations in chapter 6 show that the current generation of organic

tandem devices, comprising a ZnO interlayer, still requires a UV light soak

treatment. Since the irradiation of UV light is also known to be responsible for

photo-oxidation and the destruction of the active layer material, it is anticipated

that the final OPV product will not rely on UV light treatment. Therefore future

investigations on the long term stability of tandem OPV devices should primarily

focus on the development of new absorbers materials with both an enhanced

photo and structural stability. Furthermore, the replacement of the ZnO electron

transport layers with alternative materials which do not need to be subjected to

UV photodoping is required. Finally, fully solution processed tandem OPVs on

flexible plastics with state-of-the-art encapsulation need to be demonstrated to

certify market readiness.


122

The investigations in chapter 7 show the importance of high advanced

encapsulation materials in order to obtain long lifetimes of OPV devices under

real weather conditions. The study only focuses on single cells which were

sealed between two rigid glass sheets and stored under controlled temperature

and moisture settings. Detailed investigations under real weather conditions are

needed in order to confirm the observations made in the experiments. During

the investigation it was shown that the device stability can be seen as a function

of the encapsulation quality. To improve the device stability, new and cost-

efficient encapsulation materials need to be developed. Furthermore, the

development of new and moisture inert cell layers is needed. Finally, among the

IR imaging techniques studied in this study, ELLI represents the fastest and

most responsive optical imaging modality. It thus bears the potential to be

employed as a standard characterization technique for both studying the local

radiative recombination behavior and the quality of barrier materials in organic

solar cells.

In general during the investigations in chapter 7 and chapter 8 it was

shown that IR and NIR imaging techniques can be used for a fast, reliable, and

non-destructive characterization of different loss processes in thin film solar

cells. Especially, ILIT imaging under the Voc conditions enables a contactless

characterization of PV devices at different fabrication steps. Hence, this imaging

technique can be used as a quality control tool in large industrial production

lines for PV modules at different production steps. Since the presented

experiments in chapter 8 show basic loss processes of light induced charge

carriers, the finding can be easily transferred to other solar cell technologies.

Appendix A

123

Appendix A

Figure A1. Electroluminescence lock-in (ELLI) imaging of photo-aged OPV tandem cells. a) ELLI image of photoaged tandem cells after 2000 h of continuous white light illumination. No significant change in EL radiation was observed during the 2000 h of continuous irradiation, suggesting stable packaging. (b) and (c) show ELLI images of fresh and aged P3HT:PCBM single cells, respectively (see Figure 1 for device geometry). The aging process was accelerated by storing the sample in a climate chamber (dark) with 65°C/85% relative humidity for 168 h. After a storage time of 168 h, a clear disruption of the ELLI signal becomes apparent due to encapsulation failure (c). For ELLI imaging an Equus 327k NM infrared (IR) camera (IRCAM GmbH, Erlangen, Germany) was used, equipped with a cooled indium-gallium-arsenide (InGaAs) FPA detector (640x512). The spectral response of the detector is between 0.8 µm and 1.8 µm. The camera was run at a frame rate of 60 Hz and standard lock-in detection was applied. The cells were biased with 1.5 V in forward direction.

*This part is adopted from ref [198] with the permission of the Royal Society of Chemistry

Appendix A

124

Figure A2. Optically simulated short-circuit current based on the transfer matrix approach for tandem solar cells consisting of P3HT:PC[60]BM as the front cell and pDPP5T-2:PC[70]BM as the back cell. The short-circuit current (Jsc) was calculated assuming IQEs of 80% and 65% for the front and back cell, respectively. The thicknesses of the remaining layers where chosen to be as described in the section 6. The asterisk represents the predicted Jsc of the tandem cell studied in this work.

Appendix A

125

Figure A3. EQE spectra of organic single and tandem solar cells based on the polymer blends P3HT:PC[60]BM (bottom cell of tandem structure) and pDPP5T-2:[70]PCBM (top cell of tandem structure). The EQE spectra of the tandem solar cell were recorded under blue and infrared light bias.

Figure A4. Photovoltaic device parameters of inverted P3HT:PCBM single solar cells with and without a ≈20 nm thick Ba(OH)2 interfacial layer under continuous photo-aging. The device geometry was the same as in Figure 6.10 of the main text: ITO/AZO/Ba(OH)2/P3HT:PCBM/PEDOT:PSS/Ag. At time t=0 the devices were UV light soaked for 10 s. Continuous photo-aging occurred under white light illumination using high power LEDs without UV component (400 – 750 nm) at ≈1000 W/m². UV light treatment was repeated after 78 h for cells without Ba(OH)2 and after 340 h for cells containing a thin Ba(OH)2 layer. Importantly, devices with Ba(OH) feature significantly improved overall device stability. Moreover, no light soaking effect is apparent upon repeating the UV treatment as opposed to devices without Ba(OH)2. Each data point represents an average value of 5 solar cells and is normalized to t = 0. The photovoltaic parameters were extracted from J-V curves measured under an AM1.5 solar simulator at 1000 W/m².

Appendix A

126

Figure A5. Temporal evolution of the EQE under continuous white light irradiation (left) and the extracted short circuit currents (right) for a tandem solar cell consisting of P3HT:PC[60]BM as the front cell and pDPP5T-2:PC[70]BM as the rear cell with the stack geometry: ITO/AZO/Ba(OH)2/P3HT:PC[60]BM/ Pedot/ZnO/Ba(OH)2/pDPP5T2:PC[70]BM/MoOx/Ag. The normalized short circuit currents were extracted from the J-V performance of the tandem cell (black squares) and from the EQE spectra of the front (red circles) and rear (blue triangles) sub-cells. The tandem cell was UV light soaked for 10 s at t=0 and maintained under open circuit conditions. No periodic UV light treatment was applied. The EQE spectra of the tandem solar cell were recorded under blue and infrared light bias. The temporal evolution of the short circuit current extracted from the J-V curves tracks the short circuit current originating from the P3HT:PC[60]BM sub-cell.

Appendix B

127

Appendix B

IR imaging:

Electroluminescence (EL) and infrared (IR) imaging based on ELLI, DLIT,

ILIT as well as PL imaging were carried out during the experiments. For ELLI

and DLIT measurements, two Equus 327k NM IR cameras (IRCAM GmbH,

Erlangen, Germany) equipped with different types of detectors were used

(Figure S1). The IR-camera used for DLIT investigations was equipped with a

cooled indium-antimonite (InSb) based focal plane array (FPA) detector

(640 x 512). The InSb detector is highly responsive in the spectral range 1.5 µm

– 5 µm and has a noise equivalent temperature difference of less than 20 mK.

For ELLI, a cooled indium-gallium-arsenide (InGaAs) FPA detector (640x512)

with an optimum spectral response between 0.8 µm and 1.8 µm was used. Both

cameras were run at a frame rate of 100 Hz and controlled with a computer for

real-time lock-in calculations. The cameras were equipped with a 25 mm focal

lens imaging system featuring a spectral transparency >90% (IRCAM GmbH,

Erlangen, Germany).

Calcium deposition:

The Ca layers were thermal evaporated glass substrates

(25 mm x 25 mm x 1.3 mm) similar to the glass substrates used for the cell

fabrication. The substrates were successively sonicated in acetone and

isopropanol for 10 min each. Thermal evaporation of the Ca layers with a

thickness of 100 nm was done inside a glove box with a nitrogen protection

atmosphere. After evaporation the samples were encapsulated using a thick

(0.7 mm) glass barrier and an ultra violet curable epoxy adhesive from DELO

(Katiobond LP 655).

*This part including all images is adopted from the paper „Water ingress in encapsulated

inverted organic solar cells: correlating infrared imaging and photovoltaic performance“. The

Manuscript is submitted (30th of May, 2015) to the scientific journal Advanced Energy

Materials with the number of submission aenm.201501065.

Appendix B

128

Table B1: Data to calculate D from Eq 6.

Data Value

Saturation concentration of adhesive cs

(65°C/ 85%RH)

3.7∙10-3 mol/cm³

Saturation concentration of adhesive cs

(60°C/ 90%RH)

3.2∙10-3 mol/cm³

dx/d√t 4∙10-6 m/√s

Determination of D using first law of Fick

In order to understand the diffusion kinetic of moisture inside the encapsulated

OPV devices the diffusion constant D was calculated by applying Fick’s 1st law

which can be written as

J = −D ∙∆c

l, Eq. S1

where J (mol cm-2s-1) is the diffusion flux of moisture, D (m2/s) is the diffusion

coefficient, and Δc as the concentration difference between of saturation

concentration cs and material concentration c0 across the membrane thickness l.

The Diffusion coefficient reflects the speed in with the moisture diffuses though

the adhesive of our sample and can be written by rearranging of Eq 1 as:

D0 =J∙l

(cs−c0), Eq. S2

For the diffusion flux of moisture through the adhesive (DELO Katiobond

LP655) we used the water vapor transmission rate (WVTR = 6.1 g m-2 d-1 at

60°C/90%RH) as given from the technical data sheet [239]. The saturation

concentration cs was measured gravimetrically to be 3.2∙10-3 mol/cm³ (SI). The

thickness of our adhesive film was measured with 15 µm (cell stack ~150 nm).

Using these input parameters Eq. S2 gives rise to D0 = 1.05∙10-12 m2∙s-1. The

calculated diffusion constant refers to water diffusion through the adhesive at

60° C and a relative humidity of 90%. For our test conditions (65°C/85%RH) D

Appendix B

129

needs to be corrected. We use the determined Ea for the diffusion process and

a modification of Eq. S1. In this case D is defined as

𝐷 = 𝐷0 ∙ 𝑒𝑥𝑝 (𝐸𝑎

𝑅∙ (

1

𝑇−

1

𝑇0)). Eq. S3

and can be calculated to D = 1.37∙10-12 m2∙s-1, which is in close agreement to

the value reported in the main text.

Figure B1: JV characteristic of a test device which fulfills the section criteria mentioned in section 7.1. The table show averaged values of Jsc, Voc, FF, and PCE of the test cells at t = 0 h.

Appendix B

130

Figure B2. Schematic illustration of the lock-in setup used for ELLI, DLIT and ILIT investigations

Figure B3. Extrapolated lifetime of inverted OPV cells. Long-term PCE decay of inverted P3HT:PC60BM based solar cells. Each data point represents an average value of 10 devices. For estimating the accelerated lifetime, we applied a linear fit to the data points and extended the fit to where the efficiency drops to 80% of the initial value (red and black line).

Appendix B

131

Figure B4. ELLI line scans (top to bottom) of the solar cell shown in Figure 7.6 as a function of the storage time at 65°C/85 %.

Figure B5 Normalized PL and EL spectra of the encapsulated and inverted P3HT:PCBM solar cell. For the excitation of PL a light beam of an argon ion laser with an wavelength of 488nm was used. EL radiation was measured with an injection current of 48 mA/cm².

Appendix B

132

Figure B6. ELLI images of fresh and aged a pulsed voltage of 1 V (forward bias). The cells were kept in the dark at 65°C/85%RH. The excitation conditions were similar to those described in section 2.3. The difference in contrast to the measurements presented in Fig 7.6 arises from larger distance between sample and camera during the ELLI measurements.

Figure B7. Spatial distribution of PL signal of a fresh (left) and degraded (right) cell presented in Figure 7.8.

Appendix B

133

Figure B8: Active cell area vs. different cell parameters. For the analysis the left bottom cell presented in Figure B6 was taken into account.

Figure B9: Comparison of the diffusion kinetics of moisture though the adhesive of Ca test and OPV device. The P3HT:PCBM cell was fabricated under the same condition as described in section 7.1.1. For both experiments the samples were stored in dark at 65°C/85%. In terms of a better comparison of the moisture diffusion through the adhesive the active area of a cell was set to 1.44 cm². The red square angle represents the position of the active area. During the experiment a change of contrast was observed at the Ca test. This change might be due to a changing layer thickness and can be related to a primary diffusion of moisture through the adhesive material.

Appendix B

134

Figure B10: Experimentally determined kinetics of water diffusion in epoxy adhesive using calcium test: penetration distance vs. t1/2. The error bars are approximated reading errors. The slope dx/dh is dx/d√t = 3.9∙10-6 m/√s

Figure B11: Degradation characteristics of cell parameters of an encapsulated and inverted P3HT:PCBM solar cell with a grid finger electrode made of silver. The cell was stored under controlled moisture and temperature setting at 65°C/85% RH. The cell was ELLI investigated with a pulsed injection current of 80 mA, a lock in frequency of 1 Hz, and a measure time of 60 s.

Appendix B

135

Figure B12: (left) ELLI image of a fresh OPV device with a grid finger electrode. (right) Line scans of local ELLI emission of a fresh and aged device. Both ELLI images were measured with an injection current of 80 mA. The increased ELLI emission of the aged OPV device can be attributed to an increased current density of the active area which is due to the decreasing functionality of the electrode/active layer interface.

Figure B13: (left) ELLI images of an aged OPV device with a grid finger electrode. The continuous and dashed lines refer to line scans presented in the right graph. (right) Line scans of local ELLI emission. The ELLI image was measured with an injection current of 80 mA.

Appendix B

136

Figure B14: ELLI images of encapsulated and inverted P3HT:PCBM solar cells measured after different degradation stages. The samples were stored under controlled temperature and moisture settings at 85°C/85%RH. Each sample was investigated with a lock in frequency of 1 Hz, measure time of 15 s, and an injection current of 10 mA

Appendix B

137

Figure B15: Degradation of full (ITO/ZnO/P3HT:PCBM/PEDOT:PSS/Ag) and partial (AL: ITO/ZnO/P3HT:PCBM) devices under damp heat (DH: 85°C/85%RH). Full devices were measured before and after DH. Partial devices were finalized and measured after exposure to DH.

before DH after DH AL DH

2

4

6

8

Jsc

Conditions (DH= 85°C, 85% RH Damp Heat)

JSC

0.0

0.2

0.4

0.6

0.8

VOC

Voc

before DH after DH AL DH0

1

2

3

PC

E

Conditions (DH= 85°C, 85% RH Damp Heat)

PCE

0.2

0.4

0.6

0.8

FF

FF

-0.4 0.0 0.4 0.8

-10

0

10

Curr

ent D

ensity [m

A/c

m2]

Bias [V]

Complete device

before 85/85

after 85/85 (2hrs)

Incomplete device

85/85 after AL (2hrs)

Appendix C

138

Appendix C

Figure C1: Shunt screening behavior of a defect in a CIGS cell investigated with ELLI, DLIT, and ILIT. For all measurement the same imaging settings were used. The injection current for DLIT and ILIT measurements was adjusted to the Isc of the respective light intensity used for the ILIT investigations

Appendix C

139

Figure C2: Comparison of shunt screening measurements at different excitation modes. The injection currents are related to the Isc of the respective light intensity

Simulation model:

We developed 2D-simulations (finite element method) by using the software

COMSOL Multiphysics and a multi-diode network model. Specifically, we

simulated a submodule with 5 cells in order to take into account the influence of

the adjacent neighbor cells on the middle cell's electrical properties. A 2D-

network was generated using a geometric dependent mesh, with mesh

Appendix C

140

refinement towards small objects such as defects. The electrical parameters

(such as ideality factor and parallel resistance) were found using the IV-curves

found through our measurements. The sheet resistances were 18 Ω/sq. for

ZnO:Al and 1.25 Ω/sq. for Mo. The simulated cells were 28 cm long and 0.4 cm

wide which was essentially equal to the original cell area of our samples. The

ohmic defect was set in the center of the cell in the center of the submodule

with an area of 0.001 cm² and a resistance of 0.1 Ω∙cm²

Bibliography

141

Bibliography

[1] S.B. Darling, F. You, The case for organic photovoltaics, RSC Advances. 3 (2013) 17633–17648.

[2] E. Becquerel, La Lumière, ses causes et ses effects, 1st ed., University of Michigan Library, 1867.

[3] M.A. Green, The Path to 25 % Silicon Solar Cell Efficiency : History of Silicon Cell Evolution, Progress in Photovoltaics: Research and Applications. 17 (2009) 183–189.

[4] K.P. Bhandari, J.M. Collier, R.J. Ellingson, D.S. Apul, Energy payback time (EPBT) and energy return on energy invested (EROI) of solar photovoltaic systems: A systematic review and meta-analysis, Renewable and Sustainable Energy Reviews. 47 (2015) 133–141.

[5] K. Yamamoto, A. Nakajima, M. Yoshimi, T. Sawada, S. Fukuda, T. Suezaki, et al., A high efficiency thin film silicon solar cell and module, Solar Energy. 77 (2004) 939–949.

[6] C. Gretener, J. Perrenoud, L. Kranz, L. Kneer, R. Schmitt, S. Buecheler, et al., CdTe / CdS thin fi lm solar cells grown in substrate con fi guration, Progress in Photovoltaics: Research and Applications. (2013) 1580–1586.

[7] C.J. Brabec, S. Gowrisanker, J.J.M. Halls, D. Laird, S. Jia, S.P. Williams, Polymer-fullerene bulk-heterojunction solar cells., Advanced Materials (Deerfield Beach, Fla.). 22 (2010) 3839–56.

[8] M. Powalla, P. Jackson, W. Witte, D. Hariskos, S. Paetel, C. Tschamber, et al., High-efficiency Cu(In,Ga)Se2 cells and modules, Solar Energy Materials and Solar Cells. (2013) 1–8.

[9] NREL, Research Cell Efficiency Records, Press Release. (2014).

[10] P. Jackson, D. Hariskos, E. Lotter, S. Paetel, R. Wuerz, M. Richard, et al., New world record efficiency for Cu (In, Ga) Se2 thin film solar cells beyond 20%, Progress in Photovoltaics: Research and Applications. 19 (2011) 894–897.

[11] P. Jackson, D. Hariskos, R. Wuerz, O. Kiowski, A. Bauer, T.M. Friedlmeier, et al., Properties of Cu(In,Ga)Se 2 solar cells with new record efficiencies up to 21.7%, Physica Status Solidi (RRL) - Rapid Research Letters. 9999 (2014) n/a–n/a.

Bibliography

142

[12] M.A. Green, K. Emery, Y. Hishikawa, W. Warta, E.D. Dunlop, Solar cell efficiency tables (version 43), Progress in Photovoltaics: Research and Applications. 22 (2014) 1–9.

[13] Heliatek, Heliatek consolidates its technology leadership by establishing a new world record for organic solar technology with a cell efficiency of 12%, Press Release. (2013).

[14] F.W. Fecher, A. Pérez Romero, C.J. Brabec, C. Buerhop-Lutz, Influence of a shunt on the electrical behavior in thin film photovoltaic modules – A 2D finite element simulation study, Solar Energy. 105 (2014) 494–504.

[15] K. Kawano, C. Adachi, Evaluating Carrier Accumulation in Degraded Bulk Heterojunction Organic Solar Cells by a Thermally Stimulated Current Technique, Advanced Functional Materials. 19 (2009) 3934–3940.

[16] R. De Bettignies, J. Leroy, M. Firon, C. Sentein, Accelerated lifetime measurements of P3HT:PCBM solar cells, Synthetic Metals. 156 (2006) 510–513.

[17] M. Manceau, A. Rivaton, J.-L. Gardette, S. Guillerez, N. Lemaître, Light-induced degradation of the P3HT-based solar cells active layer, Solar Energy Materials and Solar Cells. 95 (2011) 1315–1325.

[18] J. Schafferhans, A. Baumann, A. Wagenpfahl, C. Deibel, V. Dyakonov, Oxygen doping of P3HT:PCBM blends: Influence on trap states, charge carrier mobility and solar cell performance, Organic Electronics: Physics, Materials, Applications. 11 (2010) 1693–1700.

[19] E. Voroshazi, B. Verreet, T. Aernouts, P. Heremans, Long-term operational lifetime and degradation analysis of P3HT:PCBM photovoltaic cells, Solar Energy Materials and Solar Cells. 95 (2011) 1303–1307.

[20] M. Manceau, S. Chambon, A. Rivaton, J.-L. Gardette, S. Guillerez, N. Lemaître, Effects of long-term UV–visible light irradiation in the absence of oxygen on P3HT and P3HT:PCBM blend, Solar Energy Materials and Solar Cells. 94 (2010) 1572–1577.

[21] M.O. Reese, A.J. Morfa, M.S. White, N. Kopidakis, S.E. Shaheen, G. Rumbles, et al., Pathways for the degradation of organic photovoltaic P3HT:PCBM based devices, Solar Energy Materials and Solar Cells. 92 (2008) 746–752.

[22] A. Manor, E. a. Katz, T. Tromholt, F.C. Krebs, Electrical and Photo-Induced Degradation of ZnO Layers in Organic Photovoltaics, Advanced Energy Materials. 1 (2011) 836–843.

[23] K. Kawano, R. Pacios, D. Poplavskyy, J. Nelson, D.D.C. Bradley, J.R. Durrant, Degradation of organic solar cells due to air exposure, Solar Energy Materials and Solar Cells. 90 (2006) 3520–3530.

Bibliography

143

[24] K. Norrman, N.B. Larsen, F.C. Krebs, Lifetimes of organic photovoltaics: Combining chemical and physical characterisation techniques to study degradation mechanisms, Solar Energy Materials and Solar Cells. 90 (2006) 2793–2814.

[25] H. Hintz, H.-J. Egelhaaf, L. Luer, J. Hauch, H. Peisert, T. Chasse, Photodegradation of P3HT−A Systematic Study of Environmental Factors, Chemistry of Materials. 23 (2011) 145–154.

[26] A. Seemann, H.-J. Egelhaaf, C.J. Brabec, J. a. Hauch, Influence of oxygen on semi-transparent organic solar cells with gas permeable electrodes, Organic Electronics. 10 (2009) 1424–1428.

[27] J.U. Lee, J.W. Jung, J.W. Jo, W.H. Jo, Degradation and stability of polymer-based solar cells, Journal of Materials Chemistry. 22 (2012) 24265–24283.

[28] J. Isenberg, A.S.H. van der Heide, W. Warta, Investigation of series resistance losses by illuminated lock-in thermography, Progress in Photovoltaics: Research and Applications. 13 (2005) 697–703.

[29] C. Buerhop, J. Bachmann, Infrared analysis of thin-film photovoltaic modules, Journal of Physics: Conference Series. 214 (2010) 012089.

[30] O. Breitenstein, J.P. Rakotoniaina, M.H. Al Rifai, Quantitative evaluation of shunts in solar cells by lock-in thermography, Progress in Photovoltaics: Research and Applications. 11 (2003) 515–526.

[31] O. Breitenstein, J. Bauer, T. Trupke, R.A. Bardos, On The Detection of Shunts in Silicon Solar Cells by Photo-and Electroluminescence Imaging, Progress in Photovoltaics: Research and Applications. (2008) 325–330.

[32] J. Bachmann, C. Buerhop-Lutz, R. Steim, P. Schilinsky, J. a. Hauch, E. Zeira, et al., Highly sensitive non-contact shunt detection of organic photovoltaic modules, Solar Energy Materials and Solar Cells. 101 (2012) 176–179.

[33] J. Bauer, O. Breitenstein, J.-M. Wagner, Lock-in Thermography: A Versatile Tool for Failure Analysis of Solar Cells, Device Failure Analysis. 11 (2009) 6–12.

[34] C. Buerhop, D. Schlegel, M. Niess, C. Vodermayer, R. Weißmann, C.J. Brabec, Reliability of IR-imaging of PV-plants under operating conditions, Solar Energy Materials and Solar Cells. 107 (2012) 154–164.

[35] J. Isenberg, W. Warta, Spatially resolved evaluation of power losses in industrial solar cells by illuminated lock-in thermography, Progress in Photovoltaics: Research and Applications. 12 (2004) 339–353.

Bibliography

144

[36] T. Kirchartz, U. Rau, M. Kurth, J. Mattheis, J.H. Werner, Comparative study of electroluminescence from Cu(In,Ga)Se2 and Si solar cells, Thin Solid Films. 515 (2007) 6238–6242.

[37] P. Chawla, S. Singh, S.N. Sharma, An insight into the mechanism of charge-transfer of hybrid polymer:ternary/quaternary chalcopyrite colloidal nanocrystals., Beilstein Journal of Nanotechnology. 5 (2014) 1235–44.

[38] A. Luque, S. Hegedus, Handbook of Photovoltaic Science and Engineering, 1st ed., John Wiley & Sons Ltd., 2002.

[39] A. Chirilă, S. Buecheler, F. Pianezzi, P. Bloesch, C. Gretener, A.R. Uhl, et al., Highly efficient Cu(In,Ga)Se2 solar cells grown on flexible polymer films., Nature Materials. 10 (2011) 857–61.

[40] M. Bar, S. Nishiwaki, L. Weinhardt, S. Pookpanratana, W.N. Shafarman, C. Heske, Electronic level alignment at the deeply buried absorber/Mo interface in chalcopyrite-based thin film solar cells, Applied Physics Letters. 93 (2008) 042110.

[41] B.J. Stanbery, Copper Indium Selenides and Related Materials for Photovoltaic Devices, Critical Reviews in Solid State and Materials Sciences. 27 (2002) 73–117.

[42] F. Kessler, D. Rudmann, Technological aspects of flexible CIGS solar cells and modules, Solar Energy. 77 (2004) 685–695.

[43] M.P. Suryawanshi, G.L. Agawane, S.M. Bhosale, S.W. Shin, P.S. Patil, J.H. Kim, et al., CZTS based thin film solar cells: a status review, Materials Technology: Advanced Performance Materials. 28 (2013) 98–109.

[44] W. Wang, M.T. Winkler, O. Gunawan, T. Gokmen, T.K. Todorov, Y. Zhu, et al., Device characteristics of CZTSSe thin-film solar cells with 12.6% efficiency, Advanced Energy Materials. 4 (2014).

[45] M. Morkel, L. Weinhardt, B. Lohmüller, C. Heske, E. Umbach, W. Riedl, et al., Flat conduction-band alignment at the CdS/CuInSe2 thin-film solar-cell heterojunction, Applied Physics Letters. 79 (2001) 4482–4484.

[46] D. Schmid, M. Ruckh, F. Grunwald, H.W. Schock, Chalcopyrite/defect chalcopyrite heterojunctions on the basis of CuInSe2, Journal of Applied Physics. 73 (1993) 2902–2909.

[47] T. Dullweber, G. Hanna, U. Rau, H.W. Schock, New approach to high-efficiency solar cells by band gap grading in Cu(In,Ga)Se2 chalcopyrite semiconductors, Solar Energy Materials and Solar Cells. 67 (2001) 145–150.

Bibliography

145

[48] M. Turcu, O. Pakma, U. Rau, Interdependence of absorber composition and recombination mechanism in Cu(In,Ga)(Se,S)2 heterojunction solar cells, Applied Physics Letters. 80 (2002) 2598–2600.

[49] I.M. Kötschau, H.W. Schock, Depth profile of the lattice constant of the Cu-poor surface layer in (Cu2Se)1−x(In2Se3)x evidenced by grazing incidence X-ray diffraction, Journal of Physics and Chemistry of Solids. 64 (2003) 1559–1563.

[50] K. Ramanathan, H. Wiesner, S. Asher, D. Niles, R.N. Bhattacharya, High-Efficiency Cu(In,Ga)Se2 Thin Film Solar Cells Without Intermediate Buffer Layers, in: 2nd World Conference and Exhibition on Photovoltaic Solar Energy Conversion, NREL, Vienna, Austria, 1998: p. 477.

[51] T. Nakada, M. Mizutani, 18% Efficiency Cd-Free Cu(In, Ga)Se2 Thin-Film Solar Cells Fabricated Using Chemical Bath Deposition (CBD)-ZnS Buffer Layers, Japanese Journal of Applied Physics. 41 (2002) L165–L167.

[52] D. Hariskos, S. Spiering, M. Powalla, Buffer layers in Cu(In,Ga)Se2 solar cells and modules, Thin Solid Films. 480-481 (2005) 99–109.

[53] G. Hadziioannou, G.G. Malliaras, Semiconducting Polymers, 2nd ed., WILEY-VCH Verlag GmbH & Co. KGaA, 2008.

[54] N. Greeves, http://www.chemtube3d.com/orbitalsethene.htm, (2014).

[55] C. Deibel, V. Dyakonov, Polymer–fullerene bulk heterojunction solar cells, Reports on Progress in Physics. 73 (2010) 096401.

[56] F. Liu, Y. Gu, J.W. Jung, W.H. Jo, T.P. Russell, On the morphology of polymer-based photovoltaics, Journal of Polymer Science Part B: Polymer Physics. 50 (2012) 1018–1044.

[57] P.W.M. Blom, V.D. Mihailetchi, L.J.A. Koster, D.E. Markov, Device Physics of Polymer:Fullerene Bulk Heterojunction Solar Cells, Advanced Materials. 19 (2007) 1551–1566.

[58] H. Kim, J.S. Horwitz, G. Kushto, A. Pique, Z.H. Kafafi, C.M. Gilmore, et al., Effect of film thickness on the properties of indium tin oxide thin films, Journal of Applied Physics. 88 (2000) 6021.

[59] K.-H. Choi, J.-A. Jeong, J.-W. Kang, D.-G. Kim, J.K. Kim, S.-I. Na, et al., Characteristics of flexible indium tin oxide electrode grown by continuous roll-to-roll sputtering process for flexible organic solar cells, Solar Energy Materials and Solar Cells. 93 (2009) 1248–1255.

[60] J. Wu, H.A. Becerril, Z. Bao, Z. Liu, Y. Chen, P. Peumans, Organic solar cells with solution-processed graphene transparent electrodes, Applied Physics Letters. 92 (2008) 263302.

Bibliography

146

[61] Y. Zhou, F. Zhang, K. Tvingstedt, S. Barrau, F. Li, W. Tian, et al., Investigation on polymer anode design for flexible polymer solar cells, Applied Physics Letters. 92 (2008) 233308.

[62] J.L. Blackburn, T.M. Barnes, M.C. Beard, Y.-H. Kim, R.C. Tenent, T.J. McDonald, et al., Transparent conductive single-walled carbon nanotube networks with precisely tunable ratios of semiconducting and metallic nanotubes., ACS Nano. 2 (2008) 1266–74.

[63] J. Krantz, M. Richter, S. Spallek, E. Spiecker, C.J. Brabec, Solution-Processed Metallic Nanowire Electrodes as Indium Tin Oxide Replacement for Thin-Film Solar Cells, Advanced Functional Materials. 21 (2011) 4784–4787.

[64] R. Po, C. Carbonera, A. Bernardi, N. Camaioni, The role of buffer layers in polymer solar cells, Energy & Environmental Science. 4 (2011) 285.

[65] S. Karg, J.C. Scott, J.R. Salem, M. Angdopoulos, Increased brightness and lifetime of polymer light-emitting diodes with polyaniline anodes, Synthetic Metals. 80 (1996) 111–117.

[66] Y. Cao, G. Yu, C. Zhang, R. Menon, A.. Heeger, Polymer light-emitting diodes with polyethylene dioxythiophene–polystyrene sulfonate as the transparent anode, Synthetic Metals. 87 (1997) 171–174.

[67] M. Jørgensen, K. Norrman, F.C. Krebs, Stability/degradation of polymer solar cells, Solar Energy Materials and Solar Cells. 92 (2008) 686–714.

[68] M. Jørgensen, K. Norrman, S. a Gevorgyan, T. Tromholt, B. Andreasen, F.C. Krebs, Stability of polymer solar cells., Advanced Materials (Deerfield Beach, Fla.). 24 (2012) 580–612.

[69] T. Stubhan, N. Li, N. a. Luechinger, S.C. Halim, G.J. Matt, C.J. Brabec, High Fill Factor Polymer Solar Cells Incorporating a Low Temperature Solution Processed WO 3 Hole Extraction Layer, Advanced Energy Materials. 2 (2012) 1433–1438.

[70] T. Stubhan, T. Ameri, M. Salinas, J. Krantz, F. Machui, M. Halik, et al., High shunt resistance in polymer solar cells comprising a MoO 3 hole extraction layer processed from nanoparticle suspension, Applied Physics Letters. 98 (2011) 253308.

[71] S. Schubert, H. Klumbies, L. Müller-Meskamp, K. Leo, Electrical calcium test for moisture barrier evaluation for organic devices, Review of Scientific Instruments. 82 (2011) 094101.

[72] T. Stubhan, H. Oh, L. Pinna, J. Krantz, I. Litzov, C.J. Brabec, Inverted organic solar cells using a solution processed aluminum-doped zinc oxide buffer layer, Organic Electronics. 12 (2011) 1539–1543.

Bibliography

147

[73] T. Stubhan, I. Litzov, N. Li, M. Salinas, M. Steidl, G. Sauer, et al., Overcoming interface losses in organic solar cells by applying low temperature, solution processed aluminum-doped zinc oxide electron extraction layers, Journal of Materials Chemistry A. 1 (2013) 6004.

[74] N. Li, D. Baran, K. Forberich, F. Machui, T. Ameri, M. Turbiez, et al., Towards 15% energy conversion efficiency: a systematic study of the solution-processed organic tandem solar cells based on commercially available materials, Energy & Environmental Science. 6 (2013) 3407.

[75] J. You, L. Dou, K. Yoshimura, T. Kato, K. Ohya, T. Moriarty, et al., A polymer tandem solar cell with 10.6% power conversion efficiency., Nature Communications. 4 (2013) 1446.

[76] T. Ameri, N. Li, C.J. Brabec, Highly efficient organic tandem solar cells: a follow up review, Energy & Environmental Science. 6 (2013) 2390.

[77] T. Ameri, G. Dennler, C. Lungenschmied, C.J. Brabec, Organic tandem solar cells: A review, Energy & Environmental Science. 2 (2009) 347.

[78] G. Dennler, M.C. Scharber, T. Ameri, P. Denk, K. Forberich, C. Waldauf, et al., Design rules for donors in bulk-heterojunction tandem solar cells-towards 15 % energy-conversion efficiency, Advanced Materials. 20 (2008) 579–583.

[79] O. Adebanjo, P. Maharjan, P. Adhikary, M. Wang, S. Yang, Q. Qiao, Triple Junction Polymer Solar Cells, Energy & Environmental Science. 6 (2013) 3150.

[80] L. Dou, W.H. Chang, J. Gao, C.C. Chen, J. You, Y. Yang, A selenium-substituted low-bandgap polymer with versatile photovoltaic applications, Advanced Materials. 25 (2013) 825–831.

[81] L. Dou, J. You, J. Yang, C.-C. Chen, Y. He, S. Murase, et al., Tandem polymer solar cells featuring a spectrally matched low-bandgap polymer, Nature Photonics. 6 (2012) 180–185.

[82] K. Li, Z. Li, K. Feng, X. Xu, L. Wang, Q. Peng, Development of large band-gap conjugated copolymers for efficient regular single and tandem organic solar cells, Journal of the American Chemical Society. 135 (2013) 13549–13557.

[83] Y. Zhou, C. Fuentes-Hernandez, J.W. Shim, T.M. Khan, B. Kippelen, High performance polymeric charge recombination layer for organic tandem solar cells, Energy & Environmental Science. (2012).

[84] M. Powalla, M. Cemernjak, J. Eberhardt, F. Kessler, R. Kniese, H.D. Mohring, et al., Large-area CIGS modules: Pilot line production and new developments, Solar Energy Materials and Solar Cells. 90 (2006) 3158–3164.

Bibliography

148

[85] K. Matsunaga, T. Komaru, Y. Nakayama, T. Kume, Y. Suzuki, Mass-production technology for CIGS modules, Solar Energy Materials and Solar Cells. 93 (2009) 1134–1138.

[86] P. Reinhard, A. Chirilǎ, P. Blösch, F. Pianezzi, S. Nishiwaki, S. Buecheler, et al., Review of progress toward 20% efficiency flexible CIGS solar cells and manufacturing issues of solar modules, IEEE Journal of Photovoltaics. 3 (2013) 572–580.

[87] F.C. Krebs, T. Tromholt, M. Jørgensen, Upscaling of polymer solar cell fabrication using full roll-to-roll processing., Nanoscale. 2 (2010) 873–86.

[88] D. Angmo, T.T. Larsen-Olsen, M. Jørgensen, R.R. Søndergaard, F.C. Krebs, Roll-to-Roll Inkjet Printing and Photonic Sintering of Electrodes for ITO Free Polymer Solar Cell Modules and Facile Product Integration, Advanced Energy Materials. 3 (2013) 172–175.

[89] T. Mäkelä, S. Pienimaa, S. Jussila, H. Isotalo, Lithographic patterning of conductive polyaniline, Synthetic Metals. 101 (1999) 705–706.

[90] T. Granlund, T. Nyberg, L.S. Roman, M. Svensson, O. Inganäs, Patterning of polymer light-emitting diodes with soft lithography, Advanced Materials. 12 (2000) 269–273.

[91] P. Kubis, N. Li, T. Stubhan, F. Machui, G.J. Matt, M.M. Voigt, et al., Patterning of organic photovoltaic modules by ultrafast laser, Progress in Photovoltaics: Research and Applications. (2013).

[92] R. Tipnis, J. Bernkopf, S. Jia, J. Krieg, S. Li, M. Storch, et al., Large-area organic photovoltaic module—Fabrication and performance, Solar Energy Materials and Solar Cells. 93 (2009) 442–446.

[93] P. Kubis, L. Lucera, F. Machui, G. Spyropoulos, J. Cordero, A. Frey, et al., High precision processing of flexible P3HT/PCBM modules with geometric fill factor over 95%, Organic Electronics: Physics, Materials, Applications. 15 (2014) 2256–2263.

[94] M.L. Crozier, P. Adamson, A. Brunton, S. Henley, J.D. Shephard, G. Kartopu, et al., Recent developments toward a one step thin-film PV interconnection process using laser scribing and inkjet printing ., in: IEEE 40th Photovoltaic Specialists Conference (PVSC), 2014.

[95] S.M. Sze, K.K. Ng, Physics of Semiconductor Devices, 3rd ed., John Wiley & Sons Ltd., 2006.

[96] W. Shockley, H.J. Queisser, Detailed Balance Limit of Efficiency of p-n Junction Solar Cells, Journal of Applied Physics. 32 (1961) 510.

[97] C. Corsi, History highlights and future trends of infrared sensors, Journal of Modern Optics. 57 (2010) 1663–1686.

Bibliography

149

[98] K.L. Luke, L.-J. Cheng, Analysis of the interaction of a laser pulse with a silicon wafer: Determination of bulk lifetime and surface recombination velocity, Journal of Applied Physics. 61 (1987) 2282.

[99] J. Reinhardt, M. Grein, C. Bühler, M. Schubert, U. Würfel, Identifying the Impact of Surface Recombination at Electrodes in Organic Solar Cells by Means of Electroluminescence and Modeling, Advanced Energy Materials. 4 (2014) n/a–n/a.

[100] T. Kirchartz, B.E. Pieters, J. Kirkpatrick, U. Rau, J. Nelson, Recombination via tail states in polythiophene:fullerene solar cells, Physical Review B. 83 (2011) 115209.

[101] M. Saad, a. Kassis, Effect of interface recombination on solar cell parameters, Solar Energy Materials and Solar Cells. 79 (2003) 507–517.

[102] S. Kouijzer, S. Esiner, C.H. Frijters, M. Turbiez, M.M. Wienk, R. a. J. Janssen, Efficient Inverted Tandem Polymer Solar Cells with a Solution-Processed Recombination Layer, Advanced Energy Materials. 2 (2012) 945–949.

[103] C. Groves, J.C. Blakesley, N.C. Greenham, Effect of charge trapping on geminate recombination and polymer solar cell performance., Nano Letters. 10 (2010) 1063–9.

[104] U. Lemmer, R.F. Mahrt, Y. Wada, a. Greiner, H. Bassler, E.O. Göbel, Time resolved luminescence study of recombination processes in electroluminescent polymers, Applied Physics Letters. 62 (1993) 2827.

[105] M. Gloeckler, J.R. Sites, W.K. Metzger, Grain-boundary recombination in Cu(In,Ga)Se[sub 2] solar cells, Journal of Applied Physics. 98 (2005) 113704.

[106] S. Martinuzzi, M. Stemmer, L. De Photo, E.A. Dofauts, Mapping of defects and their recombination strength by a light-beam- induced current in silicon wafers, Material Science and Engineering. 24 (1994) 152–158.

[107] R. Scheer, H.-W. Schock, Chalcogenide Photovoltaics: Physics, Technologies, and Thin Film Devices, WILEY-VCH Verlag GmbH & Co. KGaA, 2011.

[108] D. Abou-Ras, T. Kirchartz, U. Rau, Advanced Characterization Techniques for Thin Film Solar Cells, 1st ed., WILEY-VCH Verlag GmbH & Co. KGaA, 2011.

[109] M. Hermenau, M. Riede, K. Leo, S. a. Gevorgyan, F.C. Krebs, K. Norrman, Water and oxygen induced degradation of small molecule organic solar cells, Solar Energy Materials and Solar Cells. 95 (2011) 1268–1277.

Bibliography

150

[110] J. Isenberg, W. Warta, Realistic evaluation of power losses in solar cells by using thermographic methods, Journal of Applied Physics. 95 (2004) 5200–5209.

[111] O. Breitenstein, J.P. Rakotoniaina, M.H. Al Rifai, M. Werner, Shunt types in crystalline silicon solar cells, Progress in Photovoltaics: Research and Applications. 12 (2004) 529–538.

[112] T.M.H. Tran, B.E. Pieters, A. Gerber, U. Rau, On the impact of defects in solar modules and the interaction between monolithically interconnected cells, in: 2014 IEEE 40th Photovoltaic Specialist Conference (PVSC), IEEE, 2014: pp. 0435–0438.

[113] R. Rösch, D.M. Tanenbaum, M. Jørgensen, M. Seeland, M. Bärenklau, M. Hermenau, et al., Investigation of the degradation mechanisms of a variety of organic photovoltaic devices by combination of imaging techniques—the ISOS-3 inter-laboratory collaboration, Energy & Environmental Science. 5 (2012) 6521.

[114] H. Hoppe, J. Bachmann, B. Muhsin, K.-H. Drue, I. Riedel, G. Gobsch, et al., Quality control of polymer solar modules by lock-in thermography, Journal of Applied Physics. 107 (2010) 014505.

[115] R. Rösch, F.C. Krebs, D.M. Tanenbaum, H. Hoppe, Quality control of roll-to-roll processed polymer solar modules by complementary imaging methods, Solar Energy Materials and Solar Cells. 97 (2012) 176–180.

[116] a. Gerber, V. Huhn, T.M.H. Tran, M. Siegloch, Y. Augarten, B.E. Pieters, et al., Advanced large area characterization of thin-film solar modules by electroluminescence and thermography imaging techniques, Solar Energy Materials and Solar Cells. (2014) 1–8.

[117] M. Seeland, R. Rösch, H. Hoppe, Luminescence imaging of polymer solar cells: Visualization of progressing degradation, Journal of Applied Physics. 109 (2011) 064513.

[118] A. Seemann, T. Sauermann, C. Lungenschmied, O. Armbruster, S. Bauer, H.-J. Egelhaaf, et al., Reversible and irreversible degradation of organic solar cell performance by oxygen, Solar Energy. 85 (2011) 1238–1249.

[119] T. Minh, H. Tran, B.E. Pieters, A. Gerber, U. Rau, I.- Photovoltaics, et al., On the Impact of Defects in Solar Modules and the Interaction between Monolithically Interconnected Cells, IEEE Journal of Photovoltaics. 1 (2014) 435–438.

[120] V.G. Karpov, G. Rich, A. V. Subashiev, G. Dorer, Shunt screening, size effects and I /V analysis in thin-film photovoltaics, Journal of Applied Physics. 89 (2001) 4975.

Bibliography

151

[121] B.E. Pieters, Spatial Modeling of Thin-Film Solar Modules Using the Network Simulation Method and SPICE, IEEE Journal of Photovoltaics. 1 (2011) 93–98.

[122] S. Besold, U. Hoyer, J. Bachmann, T. Swonke, P. Schilinsky, R. Steim, et al., Quantitative imaging of shunts in organic photovoltaic modules using lock-in thermography, Solar Energy Materials and Solar Cells. 124 (2014) 133–137.

[123] O. Breitenstein, J.P. Rakotoniaina, Electrothermal simulation of a defect in a solar cell, Journal of Applied Physics. 97 (2005) 074905.

[124] T. Ott, F.R. Runai, F. Schwäble, T. Walter, 2D network simulation and luminescence characterization of Cu (In, Ga) Se2 thin film modules, Progress in Photovoltaics: Research and Applications. 20 (2012) 600–605.

[125] F.W. Fecher, J. Adams, A. Vetter, C. Buerhop-Lutz, C.J. Brabec, Loss analysis on CIGS-modules by using contactless, imaging illuminated lock-in thermography and 2D electrical simulations, in: 2014 IEEE 40th Photovoltaic Specialist Conference (PVSC), IEEE, 2014: pp. 3331–3334.

[126] G. Acciani, O. Falcone, S. Vergura, Analysis of the thermal heating of poly-Si and a-Si photovoltaic cell by means of Fem, in: International Conference on Renewable Energies and Power Quality, 2010: pp. 1–5.

[127] M. Simon, E.L. Meyer, Detection and analysis of hot-spot formation in solar cells, Solar Energy Materials and Solar Cells. 94 (2010) 106–113.

[128] O. Breitenstein, M. Langenkamp, W. Warte, Lock-in Thermography, Basics and Use for Evaluating Electronic Devices and Materials, 2nd ed., Springer: Berlin, Heidelberg, New York, 2010.

[129] J. Adams, A. Vetter, F. Hoga, F. Fecher, J.P. Theisen, C.J. Brabec, et al., The influence of defects on the cellular open circuit voltage in CuInGaSe2 thin film solar modules—An illuminated lock-in thermography study, Solar Energy Materials and Solar Cells. 123 (2014) 159–165.

[130] H. Straube, M. Siegloch, A. Gerber, J. Bauer, O. Breitenstein, Illuminated lock-in thermography at different wavelengths for distinguishing shunts in top and bottom layers of tandem solar cells, Physica Status Solidi (C). 8 (2011) 1339–1341.

[131] O. Breitenstein, J.P. Rakotonianina, M. Kaes, T. Pernau, S. Seren, G. Hahn, et al., Lock-in thermography—A universal tool for local analysis of solar cells, 20th Photovoltaic Solar Energy Conference. 20 (2005) 590–593.

[132] A. Helbig, T. Kirchartz, R. Schaeffler, J.H. Werner, U. Rau, Quantitative electroluminescence analysis of resistive losses in Cu(In, Ga)Se2 thin-film modules, Solar Energy Materials and Solar Cells. 94 (2010) 979–984.

Bibliography

152

[133] J. Bachmann, C. Buerhop-Lutz, C. Deibel, I. Riedel, H. Hoppe, C.J. Brabec, et al., Organic solar cells characterized by dark lock-in thermography, Solar Energy Materials and Solar Cells. 94 (2010) 642–647.

[134] J.M. Bennett, E.J. Ashley, Infrared Reflectance and Emittance of Silver and Gold Evaporated in Ultrahigh Vacuum, Applied Optics. 4 (1965) 221.

[135] U. Hoyer, L. Pinna, T. Swonke, R. Auer, C.J. Brabec, T. Stubhan, et al., Comparison of Electroluminescence Intensity and Photocurrent of Polymer Based Solar Cells, Advanced Energy Materials. 1 (2011) 1097–1100.

[136] P.K. Kuo, Z.J. Feng, T. Ahmed, L.D. Favro, R.L. Thomas, J. Hartikainen, Parallel Thermal Wave Imaging Using a Vector Lock-In Video Technique, Photoacoustic and Photothermal Phenomena. 58 (1988) 415–418.

[137] G. Busse, Optoacoustic phase angle measurement for probing a metal, Applied Physics Letters. 35 (1979) 759.

[138] G. Busse, D. Wu, W. Karpen, Thermal wave imaging with phase sensitive modulated thermography, Journal of Applied Physics. 71 (1992) 3962.

[139] J. Rantala, D. Wu, G. Busse, NDT of polymer materials using lock-in thermography with water-coupled ultrasonic excitation, NDT&E International. 31 (1998) 43–49.

[140] F. Padinger, T. Fromherz, P. Denk, C.J. Brabec, J. Zettner, T. Hierl, et al., Degradation of bulk heterojunction solar cells operated in an inert gas atmosphere: a systematic study, Synthetic Metals. 121 (2001) 1605–1606.

[141] O. Breitenstein, M. Langenkamp, P. Rakotoniaina, J. Zettner, The imaging of shunts in solar cells by infrared lock-in thermography, in: Proc. 17th Eur. Photovoltaic Solar Energy Conference, 2001: p. 1499.

[142] T.J. McMahon, T.J. Berniard, D.S. Albin, Nonlinear shunt paths in thin-film CdTe solar cells, Journal of Applied Physics. 97 (2005) 054503.

[143] O. Breitenstein, J. Bauer, A. Lotnyk, J.M. Wagner, Defect induced non-ideal dark I-V characteristics of solar cells, Superlattices and Microstructures. 45 (2009) 182–189.

[144] O. Breitenstein, Nondestructive local analysis of current–voltage characteristics of solar cells by lock-in thermography, Solar Energy Materials and Solar Cells. 95 (2011) 2933–2936.

[145] O. Breitenstein, J. Bauer, P.P. Altermatt, K. Ramspeck, Influence of Defects on Solar Cell Characteristics, Solid State Phenomena. 156-158 (2009) 1–10.

Bibliography

153

[146] S. Rißland, O. Breitenstein, High resolution saturation current density imaging at grain boundaries by lock-in thermography, Solar Energy Materials and Solar Cells. 104 (2012) 121–124.

[147] O. Breitenstein, Local efficiency analysis of solar cells based on lock-in thermography, Solar Energy Materials and Solar Cells. 107 (2012) 381–389.

[148] K. Ramspeck, K. Bothe, D. Hinken, B. Fischer, J. Schmidt, R. Brendel, Recombination current and series resistance imaging of solar cells by combined luminescence and lock-in thermography, Applied Physics Letters. 90 (2007) 153502.

[149] T.Y. Chung, C.H. Wang, K.J. Chang, S.Y. Chen, H.H. Hsieh, C.P. Huang, et al., Evaluation of the spatial distribution of series and shunt resistance of a solar cell using dark lock-in thermography, Journal of Applied Physics. 115 (2014).

[150] T. Chung, C.-H.A. Cheng, C.-H. Wang, Response to “Comment on ‘Evaluation of the spatial distribution of series and shunt resistance of a solarcell using dark lock-in thermography’” [J. Appl. Phys. 116, 046101 (2014)], Journal of Applied Physics. 116 (2014) 046102.

[151] O. Breitenstein, M. Langenkamp, Lock-in contact thermography investigation of lateral electronic inhomogeneities in semiconductor devices, Sensors and Actuators A: Physical. 4247 (1998) 0–4.

[152] M. Kaes, S. Seren, T. Pernau, G. Hahn, Light-modulated lock-in thermography for photosensitive pn-structures and solar cells, Progress in Photovoltaics: Research and Applications. 12 (2004) 355–363.

[153] K. Ramspeck, K. Bothe, J. Schmidt, R. Brendel, Correlation between spatially resolved solar cell efficiency and carrier lifetime of multicrystalline silicon, Journal of Materials Science: Materials in Electronics. 19 (2008) 4–8.

[154] A. Vetter, F. Fecher, J. Adams, R. Schaeffler, J.-P. Theisen, C.J. Brabec, et al., Lock-in thermography as a tool for quality control of photovoltaic modules, Energy Science & Engineering. 1 (2013) 12–17.

[155] J. Adams, F. Fecher, R. Schäffler, R. Auer, C.J. Brabec, C. Buerhop-Lutz, Influence of Macroscopic Defects at Different Position in CIGS Modules on Optical Imaging Techniques, in: EU PVSEC, 2011: pp. 3001 – 3003.

[156] U. Rau, Reciprocity relation between photovoltaic quantum efficiency and electroluminescent emission of solar cells, Physical Review B. 76 (2007) 085303.

[157] T. Fuyuki, H. Kondo, T. Yamazaki, Y. Takahashi, Y. Uraoka, Photographic surveying of minority carrier diffusion length in

Bibliography

154

polycrystalline silicon solar cells by electroluminescence, Applied Physics Letters. 86 (2005) 1–3.

[158] D. Hinken, K. Ramspeck, K. Bothe, B. Fischer, R. Brendel, Series resistance imaging of solar cells by voltage dependent electroluminescence, Applied Physics Letters. 91 (2007) 182104.

[159] T. Kirchartz, J. Mattheis, U. Rau, Detailed balance theory of excitonic and bulk heterojunction solar cells, Physical Review B - Condensed Matter and Materials Physics. 78 (2008).

[160] U. Hoyer, M. Wagner, T. Swonke, J. Bachmann, R. Auer, A. Osvet, et al., Electroluminescence imaging of organic photovoltaic modules, Applied Physics Letters. 97 (2010).

[161] M. Manceau, A. Rivaton, J.-L. Gardette, S. Guillerez, N. Lemaître, The mechanism of photo- and thermooxidation of poly(3-hexylthiophene) (P3HT) reconsidered, Polymer Degradation and Stability. 94 (2009) 898–907.

[162] M. Manceau, E. Bundgaard, J.E. Carlé, O. Hagemann, M. Helgesen, R. Søndergaard, et al., Photochemical stability of π-conjugated polymers for polymer solar cells: a rule of thumb, Journal of Materials Chemistry. 21 (2011) 4132.

[163] A. Abdou, S. Holdcroft, Mechanisms of Photodegradation of Poly( 3-alkylthiophenes) in Solution, Macromolecules. 26 (1993) 2954–2962.

[164] M.O. Reese, A.M. Nardes, B.L. Rupert, R.E. Larsen, D.C. Olson, M.T. Lloyd, et al., Photoinduced Degradation of Polymer and Polymer-Fullerene Active Layers: Experiment and Theory, Advanced Functional Materials. 20 (2010) 3476–3483.

[165] M.S.A. Abdou, S. Holdcroft, Solid-state photochemistry of TT- conjugated poly ( 3-alkylthiophenes ), Can. J. Chem. 1901 (1995) 1893–1901.

[166] A. Aguirre, S.C.J. Meskers, R.A.J. Janssen, H.J. Egelhaaf, Formation of metastable charges as a first step in photoinduced degradation in pi-conjugated polymer:fullerene blends for photovoltaic applications, Organic Electronics: Physics, Materials, Applications. 12 (2011) 1657–1662.

[167] J.-S. Park, H. Chae, H.K. Chung, S.I. Lee, Thin film encapsulation for flexible AM-OLED: a review, Semiconductor Science and Technology. 26 (2011) 034001.

[168] K. Norrman, S. a Gevorgyan, F.C. Krebs, Water-induced degradation of polymer solar cells studied by H2(18)O labeling., ACS Applied Materials & Interfaces. 1 (2009) 102–12.

Bibliography

155

[169] H.-L. Yip, A.K.-Y. Jen, Recent advances in solution-processed interfacial materials for efficient and stable polymer solar cells, Energy & Environmental Science. 5 (2012) 5994.

[170] B. Zimmermann, U. Würfel, M. Niggemann, Longterm stability of efficient inverted P3HT:PCBM solar cells, Solar Energy Materials and Solar Cells. 93 (2009) 491–496.

[171] M.T. Lloyd, D.C. Olson, P. Lu, E. Fang, D.L. Moore, M.S. White, et al., Impact of contact evolution on the shelf life of organic solar cells, Journal of Materials Chemistry. 19 (2009) 7638.

[172] Y. Kim, a Ballantyne, J. Nelson, D. Bradley, Effects of thickness and thermal annealing of the PEDOT:PSS layer on the performance of polymer solar cells, Organic Electronics. 10 (2009) 205–209.

[173] E. Vitoratos, S. Sakkopoulos, E. Dalas, N. Paliatsas, D. Karageorgopoulos, F. Petraki, et al., Thermal degradation mechanisms of PEDOT:PSS, Organic Electronics. 10 (2009) 61–66.

[174] J. Huang, P.F. Miller, J.C. de Mello, a. J. de Mello, D.D.C. Bradley, Influence of thermal treatment on the conductivity and morphology of PEDOT/PSS films, Synthetic Metals. 139 (2003) 569–572.

[175] a. M. Nardes, M. Kemerink, M.M. de Kok, E. Vinken, K. Maturova, R. a. J. Janssen, Conductivity, work function, and environmental stability of PEDOT:PSS thin films treated with sorbitol, Organic Electronics. 9 (2008) 727–734.

[176] F.C. Krebs, Stability and Degradation of Organic and Polymer Solar Cells, 1st ed., John Wiley & Sons Ltd., 2012.

[177] M.P. De Jong, L.J. Van Ijzendoorn, M.J.A. De Voigt, Stability of the interface between indium-tin-oxide in polymer light-emitting diodes, App. Phys. Lett. 77 (2000) 2255–2257.

[178] M.T. Lloyd, C.H. Peters, A. Garcia, I. V. Kauvar, J.J. Berry, M.O. Reese, et al., Influence of the hole-transport layer on the initial behavior and lifetime of inverted organic photovoltaics, Solar Energy Materials and Solar Cells. 95 (2011) 1382–1388.

[179] J.S. Kim, P.K.H. Ho, C.E. Murphy, N. Baynes, R.H. Friend, Nature of non-emissive black spots in polymer light-emitting diodes by in-situ micro-raman spectroscopy, Advanced Materials. 14 (2002) 206–209.

[180] E. Voroshazi, B. Verreet, A. Buri, R. Müller, D. Di Nuzzo, P. Heremans, Influence of cathode oxidation via the hole extraction layer in polymer:fullerene solar cells, Organic Electronics. 12 (2011) 736–744.

Bibliography

156

[181] K. Norrman, M. V Madsen, S.A. Gevorgyan, F.C. Krebs, Degradation patterns in water and oxygen of an inverted polymer solar cell, Journal of the American Chemical Society. 132 (2010) 16883–16892.

[182] K. Feron, T.J. Nagle, L.J. Rozanski, B.B. Gong, C.J. Fell, Spatially resolved photocurrent measurements of organic solar cells: Tracking water ingress at edges and pinholes, Solar Energy Materials and Solar Cells. 109 (2013) 169–177.

[183] S.R. Ferreira, P. Lu, Y. Lee, R.J. Davis, J.W.P. Hsu, Effect of Zinc Oxide Electron Transport Layers on Performance and Shelf Life of Organic Bulk Heterojunction Devices, The Journal of Physical Chemistry C. 115 (2011) 13471–13475.

[184] F.C. Krebs, M. Jørgensen, K. Norrman, O. Hagemann, J. Alstrup, T.D. Nielsen, et al., A complete process for production of flexible large area polymer solar cells entirely using screen printing—First public demonstration, Solar Energy Materials and Solar Cells. 93 (2009) 422–441.

[185] S. Cornelius, M. Vinnichenko, N. Shevchenko, a. Rogozin, a. Kolitsch, W. Möller, Achieving high free electron mobility in ZnO:Al thin films grown by reactive pulsed magnetron sputtering, Applied Physics Letters. 94 (2009) 042103.

[186] U. Ozgur, Y.I. Alivov, C. Liu, a. Teke, M. a. Reshchikov, S. Dogan, et al., A comprehensive review of ZnO materials and devices, Journal of Applied Physics. 98 (2005) 041301.

[187] A. Manor, E. a. Katz, T. Tromholt, F.C. Krebs, Enhancing functionality of ZnO hole blocking layer in organic photovoltaics, Solar Energy Materials and Solar Cells. 98 (2012) 491–493.

[188] H. Cheun, C. Fuentes-hernandez, Y. Zhou, W.J. Potscavage, S. Kim, J. Shim, et al., Electrical and Optical Properties of ZnO Processed by Atomic Layer Deposition in Inverted Polymer Solar Cells †, (2010) 20713–20718.

[189] M.S. White, D.C. Olson, S.E. Shaheen, N. Kopidakis, D.S. Ginley, Inverted bulk-heterojunction organic photovoltaic device using a solution-derived ZnO underlayer, Applied Physics Letters. 89 (2006) 143517.

[190] F. Verbakel, S.C.J. Meskers, R. a. J. Janssen, Electronic memory effects in diodes from a zinc oxide nanoparticle-polystyrene hybrid material, Applied Physics Letters. 89 (2006) 102103.

[191] M.R. Lilliedal, A.J. Medford, M. V. Madsen, K. Norrman, F.C. Krebs, The effect of post-processing treatments on inflection points in current–voltage curves of roll-to-roll processed polymer photovoltaics, Solar Energy Materials and Solar Cells. 94 (2010) 2018–2031.

Bibliography

157

[192] I. Litzov, H. Azimi, G. Matt, P. Kubis, T. Stubhan, G. Popov, et al., Accelerated degradation of Al3+ doped ZnO thin films using damp heat test, Organic Electronics. 15 (2014) 569–576.

[193] N. Grossiord, J.M. Kroon, R. Andriessen, P.W.M. Blom, Degradation mechanisms in organic photovoltaic devices, Organic Electronics. 13 (2012) 432–456.

[194] M. Corazza, F.C. Krebs, S.A. Gevorgyan, Predicting, categorizing and intercomparing the lifetime of OPVs for different ageing tests, Solar Energy Materials and Solar Cells. 130 (2014) 99–106.

[195] M.O. Reese, S. a. Gevorgyan, M. Jørgensen, E. Bundgaard, S.R. Kurtz, D.S. Ginley, et al., Consensus stability testing protocols for organic photovoltaic materials and devices, Solar Energy Materials and Solar Cells. 95 (2011) 1253–1267.

[196] C.H. Peters, I.T. Sachs-Quintana, J.P. Kastrop, S. Beaupré, M. Leclerc, M.D. McGehee, High Efficiency Polymer Solar Cells with Long Operating Lifetimes, Advanced Energy Materials. 1 (2011) 491–494.

[197] U. Hoyer, T. Swonke, Concept for a Real AM1.5 Simulator Based on LED-Technology and Survey on Different Types of Solar Simulators, in: EU-PVSEC, 2009: pp. 3377–3379.

[198] J. Adams, G.D. Spyropoulos, M. Salvador, N. Li, S. Strohm, L. Lucera, et al., Air-processed organic tandem solar cells on glass: toward competitive operating lifetimes, Energy Environ. Sci. (2014) 1–26.

[199] http://www.keithley.de, (n.d.) www.keithley.de.

[200] T. Hierl, W. Gross, H. Scheuerpflug, A. Lutz, U. Schirl, M.J. Schulz, Infrared imaging of buried heat sources and material nonuniformities, in: B.F. Andresen, M. Strojnik (Eds.), 1997: pp. 943–953.

[201] C. Busse, Lock-in thermography for nondestructive evaluation of materials, Revue Générale de Thermique. 37 (1998) 693–703.

[202] H.S. Carslaw, J.C. Jaeger, Conduction of Heat in Solids, 2nd ed., Oxford University Press Inc., Oxford, 1959.

[203] T.R. Andersen, H.F. Dam, M. Hösel, M. Helgesen, J.E. Carlé, T.T. Larsen-Olsen, et al., Scalable, ambient atmosphere roll-to-roll manufacture of encapsulated large area, flexible organic tandem solar cell modules, Energy & Environmental Science. 7 (2014) 2925.

[204] V. Gupta, A.K.K. Kyaw, D.H. Wang, S. Chand, G.C. Bazan, A.J. Heeger, Barium: an efficient cathode layer for bulk-heterojunction solar cells., Scientific Reports. 3 (2013) 1965.

Bibliography

158

[205] L.P. Lu, D. Kabra, R.H. Friend, Barium Hydroxide as an Interlayer Between Zinc Oxide and a Luminescent Conjugated Polymer for Light-Emitting Diodes, Advanced Functional Materials. 22 (2012) 4165–4171.

[206] J. Gilot, M.M. Wienk, R.A.J. Janssen, Double and triple junction polymer solar cells processed from solution, Applied Physics Letters. 90 (2007) 143512.

[207] B.Y. Finck, B.J. Schwartz, Understanding the origin of the S-curve in conjugated polymer/fullerene photovoltaics from drift-diffusion simulations, Applied Physics Letters. 103 (2013) 053306.

[208] S.K. Hau, H.-L. Yip, H. Ma, A.K.-Y. Jen, High performance ambient processed inverted polymer solar cells through interfacial modification with a fullerene self-assembled monolayer, Applied Physics Letters. 93 (2008) 233304.

[209] Y. Sun, C.J. Takacs, S.R. Cowan, J.H. Seo, X. Gong, A. Roy, et al., Efficient, air-stable bulk heterojunction polymer solar cells using MoO(x) as the anode interfacial layer., Advanced Materials (Deerfield Beach, Fla.). 23 (2011) 2226–30.

[210] Y.-C. Tseng, A.U. Mane, J.W. Elam, S.B. Darling, Ultrathin molybdenum oxide anode buffer layer for organic photovoltaic cells formed using atomic layer deposition, Solar Energy Materials and Solar Cells. 99 (2012) 235–239.

[211] H. Zhang, T. Stubhan, N. Li, M. Turbiez, G.J. Matt, T. Ameri, et al., A solution-processed barium hydroxide modified aluminum doped zinc oxide layer for highly efficient inverted organic solar cells, J. Mater. Chem. A. 2 (2014) 18917–18923.

[212] C.H. Peters, I.T. Sachs-Quintana, W.R. Mateker, T. Heumueller, J. Rivnay, R. Noriega, et al., The mechanism of burn-in loss in a high efficiency polymer solar cell., Advanced Materials (Deerfield Beach, Fla.). 24 (2012) 663–8.

[213] N. Li, D. Baran, G.D. Spyropoulos, H. Zhang, S. Berny, M. Turbiez, et al., Environmentally Printing Efficient Organic Tandem Solar Cells with High Fill Factors: A Guideline Towards 20% Power Conversion Efficiency, Advanced Energy Materials. (2014) 140084 (1 of 7).

[214] Y. Liu, J. Zhao, Z. Li, C. Mu, W. Ma, H. Hu, et al., Aggregation and morphology control enables multiple cases of high-efficiency polymer solar cells., Nature Communications. 5 (2014) 5293.

[215] G. Shao, G.E. Rayermann, E.M. Smith, D.S. Ginger, Morphology-dependent trap formation in bulk heterojunction photodiodes., The Journal of Physical Chemistry. B. 117 (2013) 4654–60.

Bibliography

159

[216] M. Hermenau, S. Schubert, H. Klumbies, J. Fahlteich, L. Müller-Meskamp, K. Leo, et al., The effect of barrier performance on the lifetime of small-molecule organic solar cells, Solar Energy Materials and Solar Cells. 97 (2012) 102–108.

[217] F. Yan, J. Noble, J. Peltola, S. Wicks, S. Balasubramanian, Semitransparent OPV modules pass environmental chamber test requirements, Solar Energy Materials and Solar Cells. 114 (2013) 214–218.

[218] S. Cros, R. de Bettignies, S. Berson, S. Bailly, P. Maisse, N. Lemaitre, et al., Definition of encapsulation barrier requirements: A method applied to organic solar cells, Solar Energy Materials and Solar Cells. 95 (2011) S65–S69.

[219] H. Straube, O. Breitenstein, Infrared lock-in thermography through glass substrates, Solar Energy Materials and Solar Cells. 95 (2011) 2768–2771.

[220] D. Baran, N. Li, A.-C. Breton, A. Osvet, T. Ameri, M. Leclerc, et al., Qualitative analysis of bulk-heterojunction solar cells without device fabrication: an elegant and contactless method., Journal of the American Chemical Society. 136 (2014) 10949–55.

[221] H.-S. Kim, J. Huh, J. Ryu, Investigation of moisture-induced delamination failure in a semiconductor package via multi-scale mechanics, Journal of Physics D: Applied Physics. 44 (2011) 034007.

[222] B.J. Finlayson-Pitts, J.N. Pitts Jr., Upper and Lower Atmosphere: Theory, Experiments, and Applications, 6th ed., Academic Pr Inc, 1999.

[223] S. Schuller, P. Schilinsky, J. Hauch, C.J. Brabec, Determination of the degradation constant of bulk heterojunction solar cells by accelerated lifetime measurements, Applied Physics A: Materials Science & Processing. 79 (2004) 37–40.

[224] A. Kinloch, Adhesion and Adhesives: Science and Technology, 31st ed., Springer: Berlin, Heidelberg, New York, 1987.

[225] R. a. Gledhill, a. J. Kinloch, Environmental Failure of Structural Adhesive Joints, The Journal of Adhesion. 6 (1974) 315–330.

[226] K. Tvingstedt, K. Vandewal, A. Gadisa, F. Zhang, J. Manca, O. Inganäs, Electroluminescence from charge transfer states in polymer solar cells., Journal of the American Chemical Society. 131 (2009) 11819–24.

[227] B. Wozniak, J. Dera, Light Absorption in Sea Water, 33rd ed., Springer: Berlin, Heidelberg, New York, 2007.

[228] A. Maurano, R. Hamilton, C.G. Shuttle, A.M. Ballantyne, J. Nelson, B. O’Regan, et al., Recombination dynamics as a key determinant of open

Bibliography

160

circuit voltage in organic bulk heterojunction solar cells: A comparison of four different donor polymers, Advanced Materials. 22 (2010) 4987–4992.

[229] C.M. Proctor, M. Kuik, T.-Q. Nguyen, Charge carrier recombination in organic solar cells, Progress in Polymer Science. 38 (2013) 1941–1960.

[230] E.L. Ratcliff, A. Garcia, S. a. Paniagua, S.R. Cowan, A.J. Giordano, D.S. Ginley, et al., Investigating the Influence of Interfacial Contact Properties on Open Circuit Voltages in Organic Photovoltaic Performance: Work Function Versus Selectivity, Advanced Energy Materials. 3 (2013) 647–656.

[231] J.J. Michels, M. Peter, A. Salem, B. van Remoortere, J. van den Brand, A combined experimental and theoretical study on the side ingress of water into barrier adhesives for organic electronics applications, Journal of Materials Chemistry C. 2 (2014) 5759.

[232] M.A. Green, K. Emery, D.L. King, S. Igari, W. Warta, Solar cell efficiency tables, Progress in Photovoltaics: Research and Applications. 12 (2004) 55–62.

[233] F. Krebs, J. Carle, N. Cruysbagger, M. Andersen, M. Lilliedal, M. Hammond, et al., Lifetimes of organic photovoltaics: photochemistry, atmosphere effects and barrier layers in ITO-MEHPPV:PCBM-aluminium devices, Solar Energy Materials and Solar Cells. 86 (2005) 499–516.

[234] L. Lucera, P. Kubis, F.W. Fecher, C. Bronnbauer, M. Turbiez, K. Forberich, et al., Guidelines for Closing the Efficiency Gap between Hero Solar Cells and Roll-To-Roll Printed Modules, Energy Technology. (2015) n/a–n/a.

[235] F. Kessler, D. Herrmann, M. Powalla, Approaches to flexible CIGS thin-film solar cells, Thin Solid Films. 480-481 (2005) 491–498.

[236] B. Dimmler, R. Wächter, Manufacturing and application of CIS solar modules, Thin Solid Films. 515 (2007) 5973–5978.

[237] J. Poortmans, V. Arkhipov, Thin Film Solar Cells, 1st ed., John Wiley & Sons Ltd., 2006.

[238] B. Dimmler, M. Powalla, R. Schaeffler, CIS solar modules: pilot production at Wuerth Solar, IEEE Photovoltaic Specialists. (2005) 189–194.

[239] http://www.delo.de

failure analysis and long term stability of thin film ... · i want to thank felix hoga, urs...

Documents