university of groningen polymer tandem solar cells

University of Groningen

Polymer tandem solar cellsHadipour, Afshin

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite fromit. Please check the document version below.

Document VersionPublisher's PDF, also known as Version of record

Publication date:2007

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):Hadipour, A. (2007). Polymer tandem solar cells. s.n.

CopyrightOther than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of theauthor(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license.More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne-amendment.

Take-down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons thenumber of authors shown on this cover page is limited to 10 maximum.

Download date: 07-06-2022

https://research.rug.nl/en/publications/93ce8872-7571-4934-a16b-4c81e1fe3406

Polymer Tandem Solar Cells

Afshin Hadipour


Afshin Hadipour

PhD thesis

University of Groningen, The Netherlands

The Zernike institute for Advance Materials PhD thesis series 2007-11

ISSN 1570-1530

ISBN 978-90-367-3205-5

The research described in this thesis was performed in the research group

Molecular Electronic: Physics of Organic Semiconductors of the Zernike Institute

for Advanced Materials, University of Groningen, The Netherlands. The project was

financially supported by the Zernike Institute for Advanced Materials.

RIJKSUNIVERSITEIT GRONINGEN


Proefschrift

ter verkrijging van het doctoraat in de Wiskunde en Natuurwetenschappen aan de Rijksuniversiteit Groningen

op gezag van de Rector Magnificus, dr. F. Zwarts, in het openbaar te verdedigen op

vrijdag 7 december 2007 om 13.15 uur

door

Afshin Hadipour geboren op 14 oktober 1968

te Abadeh, Iran

Promotor : Prof. dr. ir. P. W. M. Blom

Copromotor : Dr. B. de Boer Beoordelingscommissie : Prof. dr. P. Heremans Prof. dr. J. C. Hummelen Prof. dr. ir. R. A. J. Janssen ISBN 978-90-367-3205-5

POLYMER TANDEM SOLAR CELLS............................................................................1

CHAPTER 1 FROM LIGHT TO ELECTRICITY ......................................................9 1.1 INTRODUCTION............................................................................................................10 1.2 CONJUGATED POLYMERS ............................................................................................11 1.3 CONCEPT OF A HETEROJUNCTION SOLAR CELL ............................................................13 1.4 FABRICATION ASPECTS................................................................................................13 1.5 ELECTRICAL CONSIDERATIONS....................................................................................15 1.6 OPTICAL ABSORPTION.................................................................................................17 1.7 OUTLINE OF THIS THESIS .............................................................................................18 1.8 ABBREVIATIONS..........................................................................................................20

CHAPTER 2 A REVIEW ON TANDEM (MULTI-JUNCTION) ORGANIC SOLAR CELL ....................................................................................................................23

2.1 INTRODUCTION............................................................................................................24 2.2 A. TANDEM AND MULTILAYER ORGANIC SOLAR CELLS BASED ON LOW MOLECULAR WEIGHT MOLECULES 27 2.3 B. HYBRID TANDEM ORGANIC SOLAR CELL................................................................34 2.4 C. SOLUTION-PROCESSED TANDEM ORGANIC SOLAR CELLS ......................................37

2.4.1 Tandem Organic Solar Cell Processed on Separated Substrates........................38 2.4.2 Tandem Organic Solar Cell with ITO as Separating Layer ................................39 2.4.3 Multiple Organic Solar Cell with Solution-processed ZnO Interlayer................41 2.4.4 Tandem Organic Solar Cell with Solution-processed TiOx Interlayer ................44

2.5 SUMMARY...................................................................................................................46 CHAPTER 3 TANDEM POLYMER SOLAR CELL USING A COMPOSITE METAL INTERLAYER....................................................................................................47

3.1 INTRODUCTION............................................................................................................48 3.2 THE MIDDLE ELECTRODE.............................................................................................50 3.3 OPTICAL MATCHING....................................................................................................51 3.4 CURRENT MATCHING...................................................................................................55 3.5 OPTIMUM TANDEM PHOTOVOLTAIC DEVICE................................................................58 3.7 FABRICATION ASPECTS OF THE DEVICE .......................................................................61 3.5 SUMMARY...................................................................................................................63

CHAPTER 4...................................................................................................................65

POLYMER TANDEM SOLAR CELLS WITH OPTICAL SPACER AS INTERLAYER ...................................................................................................................65

4.1 INTRODUCTION............................................................................................................66 4.2 ELECTRODES OF THE DEVICE.......................................................................................67 4.3 OPTIMUM SUB CELLS...................................................................................................72 4.4 OPTICAL CONSIDERATIONS..........................................................................................74 4.5 ELECTRICAL CONSIDERATIONS....................................................................................77 4.6 DEVICE FABRICATION .................................................................................................80 4.7 SUMMARY...................................................................................................................82

CHAPTER 5...................................................................................................................83

6


CURRENT-VOLTAGE CHARACTERISTICS OF ORGANIC TANDEM CELLS..83 5.1 SERIES CONFIGURATION..............................................................................................84 5.2 PARALLEL CONFIGURATION ........................................................................................89 5.3 COMPARISON WITH EXPERIMENT ................................................................................92 5.5 SUMMARY.................................................................................................................103

CHAPTER 6.................................................................................................................105

DOWN-CONVERSION OF HIGH ENERGY PHOTONS..........................................105 6.1 INTRODUCTION..........................................................................................................106 6.2 OPTIMUM SEMITRANSPARENT CATHODE ...................................................................107 6.3 DOWN-CONVERSION EFFECT .....................................................................................111 6.4 DEVICE FABRICATION ...............................................................................................114 6.5 PERFORMANCE IMPROVEMENT..................................................................................115 6.6 SUMMARY.................................................................................................................119

REFERENCES.................................................................................................................121 PUBLICATIONS……………………………………………………………………….129

SUMMARY………………………………………………………………………………131

SAMENVATTING………………………………………………………………………135

DANKWOORD………………………………………………………………………….139

8


9

Chapter 1 From Light to Electricity

Abstract

In order to realize tandem (or multi-junctions) organic solar cells, we have to

understand the single-junction organic solar cells in advance. Since all experiments

in this thesis are made of polymers as electron-donor materials and the fullerene

derivative PCBM as electron-acceptor material, this Chapter deals with basic

aspects of a single-junction bulk heterojunction (BHJ) solar cell. The BHJ solar cell

is based on blending of the donor (polymers) and the acceptor (PCBM). First, some

introduction is given about the materials and conduction mechanisms of the bulk

heterojunction solar cell. Second, the processing steps are discussed for a single-

junction cell, which form the basis for the more advanced processing of the tandem

cells. Afterwards, current-voltage measurements under standard conditions are

mentioned, as well as important parameters that can be extracted from such an

experiment. Finally, two short introductory sections are given in this Chapter about

electrical and optical aspects in general that are discussed in more details in the

following Chapters.

10


1.1 Introduction Mankind needs energy for a living. Modern living on Earth leads to a continuous

expansion of the global demand for energy. It is clear that by 2025 almost the

double amount of energy is needed as compared to today.[1] In addition, it is also

clear that combustion of fossil fuels will have serious consequences for our climate

by increasing the concentration of carbon dioxide in our atmosphere.[2,3] Therefore,

realization of green power has become a must! The most clean, renewable and

sustainable energy known to humankind is solar energy. With a photovoltaic

device, the solar light (photons) can be converted into electricity. The electrical

energy is one of the most useful forms of energy, since it can be used for almost

any application. The ‘photovoltaic effect’ is the mechanism in which a solar cell

converts photons from the solar light into electricity. While the photovoltaic effect

was first observed in 1839 in an electrochemical process by French physicist

Alexander-Edmond Becquerel, the first well-performing solid-state solar cell was

built by Charles Fritts in 1883. He coated the semiconductor Selenium with an very

thin layer of gold to form a junction that had an efficiency of 1%.[4,5] Modern

generation of solar cells was born in 1953 when at Bell Laboratories (New Jersey,

USA) the first silicon solar cell was developed with a power conversion efficiency of

6%.[6] After that, many different technologies and materials were developed in order

to improve the performance of the device and lower their production cost. In order

to achieve this goal, organic materials provide us with a variety of possibilities.

Especially semiconducting polymers combine the favorable opto-electronic

properties of organic materials, such as high absorption coefficients, with the

excellent processing and mechanical properties of plastic materials. This implies

that an organic solar cell can be processed from solution at room temperature onto

a (flexible) substrate using simple and, therefore, much cheaper methods such as

spin (or blade) coating and inkjet printing.[7,8] In this chapter, the physical aspects

and different realized device structures of organic single bulk heterojunction solar

cells are explained. The optical consideration of the single bulk heterojunction solar

cell demonstrates the need for the multiple structures in which the absorption is

11

improved by using materials with non-overlapping absorption spectra, as explained

in the further chapters.

1.2 Conjugated polymers

In the 1950s photoconductivity has been reported for the first time in organic

crystals such as anthracene and naphthalene.[9,10,11] However, they could not be

used for applications because of their poor processability. Shirakawa, MacDiarmid

and Heeger made their breakthrough in 1977, demonstrating that the conductivity

of conjugated polymers can be tuned by doping.[12] Since then, many type of

semiconducting polymers were developed and used in organic devices such as

light emitting diodes (LEDs),[13] field-effect transistors (FETs)[14,15] and solar cells

(PVs).[16,17,18] The origin behind the semiconducting properties of these polymers is

the conjugation. This means that the bonds between carbon atoms building the

backbone of the polymer are alternating single or double. In general, the single C-

C bonds, formed by the overlap of two sp2 orbitals, are σ-bonds which all lie in a

plane. However, the additional bond in a double C=C bond is formed by sideways

overlap of the remaining pz orbital on each carbon atom. This kind of overlap

results in the formation of a different type orbitals, forming π-bonds. This is

demonstrated in Figure 1.1. The conjugation between neighbouring double bonds

leads to delocalization of the electrons above and below the plane of the

conjugated polymer. Both the Lowest Unoccupied Molecular Orbital (LUMO) and

the Highest Occupied Molecular Orbital (HOMO) of such systems are empty and

filled molecular orbitals, respectively, that are totally made up from the π-system.

12


Figure 1.1. Schematic representation of the bonds of carbon in a conjugated polymer. The electrons in

π-molecular orbitals are delocalized.

The energy difference between the LUMO and HOMO level defines the (optical)

band gap of the polymer. Optical measurements demonstrate that the band gap of

most known conjugated polymers ranges from 1 to 4 eV.[19] The polymers used in

this thesis serve as donor materials, which implies that upon optical excitation

these materials can donate an electron to an accepting moiety. More specifically:

poly(2-methoxy-5-(3’,7’-dimethyl octyloxy)-1,4-phenylene vinylene) (MDMO-PPV),

region-regular poly(3-hexylthiophene) (RR-P3HT) and poly((2,7-(9,9-dioctyl)-

fluorene)-alt-5,5-(4',7'-di-2-thienyl-2',1',3'-benzothiadiazole) (PFDTBT), which all

have large band gaps of about 2.1 eV. In contrast, poly5,7-di-2-thienyl-2,3-bis(3,5-

di(2-ethylhexyloxy)phenyl)-thieno[3,4-b]pyrazine (PTBEHT) has a small band gap

of about 1.2 eV. The fullerene derivative [6,6]-phenyl-C61-butyric acid methyl ester

(PCBM) is used as the acceptor material in all experiments performed in this

thesis. The chemical structures and relevant references of the organic materials

used in organic solar cells are given in the next chapter.

13

1.3 Concept of a heterojunction solar cell

The conversion of solar light into electrical power requires the generation of both

negative (electron) and positive (hole) charges and a driving force (a potential) that

extracts these charges to an external circuit. In organic semiconductors, absorption

of photons results in the formation of excitons. An exciton can be considered as of

a strongly bound electron-hole pair with a binding energy of about 0.4 eV caused

by Coulomb interaction. These excitons, carrying energy but no net charge, have

to diffuse to the dissociation sites where their charges can be separated.[20,21,22]

The interface between suited donor (D) and acceptor (A) materials provides

dissociation sites. At the D/A interface, the energy difference in the electron

affinities and the ionization potentials of those two materials is large enough to

overcome the exciton binding energy. The holes reside in the material with the

lower ionization potential or HOMO, and electrons are captured in the material with

the higher electron affinity or lowest LUMO. After this charge transfer process the

electrons and holes are still bound by Coulomb interaction across the D/A

interface. After breaking the binding energy of this bound electron-hole (e-h) pair,

meaning the electron in the acceptor and the hole in the donor, the then free

electrons and holes travel through the acceptor and donor phase, respectively, to

the electrodes of the device.[23,24,17]To obtain efficient photon to charge conversion,

many different device architectures have been developed such as single layer

cells,[25] double layer cells[26] and bulk heterojunction blend cells. The bulk

heterojunction (BHJ) solar cells based on the intimate mixing of conjugated

polymers and fullerene derivatives are mostly used today. Because of the larger

D/A interface, which provides the exciton dissociation sites, a BHJ cell is more

efficient than the other above-mentioned structures.[27,28,29,30,31,32,33,34,35,36]

Therefore, state-of-the-art organic solar cells are based on so-called

donor/acceptor (D/A) bulk heterojunction.

1.4 Fabrication aspects In general, the electrodes of an organic solar cell must be conductive with suited

work function. This means that for extraction of the holes the anode should have

the same work function (or deeper in energy) as the HOMO level of the polymer

14


used in the active layer. For this purpose a layer of poly(3,4-ethylene

dioxythiophene) : polystryrenesulfonic acid (PEDOT:PSS, H. C. Starck) is spin

coated (about 50 nm) onto the anode that is generally based on indium tin oxide

(ITO) or gold (Au). The layer of PEDOT:PSS stabilizes the work function of anode

to about 5.1 eV. Furthermore, the PEDOT: PSS improves the wetting between the

electrode and the solution (blend of donor and acceptor that are dissolved into

chloroform or chlorobenzene) during processing of the device. Additionally,

PEDOT: PSS flattens the rough surface of the ITO anode, which decreases the

chance of shorting (anode in direct electrical contact with the cathode) of the

device. One of the two electrodes of the solar cell has to be optically as transparent

as possible, since the active layer is processed between them. Therefore, indium

tin oxide (ITO) is mostly used because it is highly transparent and has a high sheet

conductivity for transporting the holes extracted from the device under illumination (

different electrodes are considered in the following chapters). For the cathode of

the solar cell, a closed (~100 nm thick) layer of metals such as aluminum (Al) and

silver (Ag) is usually used. Electrically, the metallic cathode serves for transport of

the electrons extracted from the device under illumination. Optically, the metallic

cathode serves as a perfect mirror, which leads to more light trapping (therefore

more light absorption) inside the device. A very thin interlayer of materials such as

lithium fluoride (LiF) is normally used for two reasons; First, the combination of LiF

with above-mentioned metals has a low work function of about 3.7 eV, which is

nearly equal to the LUMO of the acceptor (PCBM) used. This means that the

cathode makes an Ohmic contact with the active layer in order to extract the

electrons from the device. The LiF also protects the soft polymer surface (the

surface of the blend) against the hot vapor-deposited metallic atoms that can

penetrate into the bulk film. This is important, because metallic particles (Al, Ag,)

that penetrate inside the active layer may serve as exciton quenching sites or may

react with the organic materials. Consequently, the efficiency and lifetime of the

device is lowered. In Figure 1.3 the device structure is depicted, based on a blend

of MDMO-PPV and PCBM.

15

Figure 1.3. The schematic structure and operation of an organic bulk heterojunction solar cell .

It is clear that the standard structure of a bulk heterojunction single solar cell

cannot be used for multiple structures processed onto one substrate, since the

structure is not (semi)transparent. For tandem (multiple) structures, both electrodes

of the first sub cell (bottom cell) have to be as transparent as possible in order to

maximize the light intensity (and therefore the performance) at the following sub

cells. In the next chapter different kinds of (semi)transparent electrodes are

reviewed.

1.5 Electrical considerations

In order to measure the electrical performance of an organic solar cell , the current-

density J vs. voltage V (J–V) characteristic has to be defined in dark and under

illumination by standard test condition (STC). The STC is 1000 W/m2 intensity,

AM1.5 simulated solar spectrum and the substrate temperature equals 25 °C.

From such a measurement, the following parameters can be extracted:

16


Open-circuit voltage (VOC): the maximum voltage that a cell can produce

under illumination. At VOC the current-density (J) is zero.

Short-circuit current (JSC): the current-density that a cell generates at zero

applied voltage under illumination.

Photo-current (Jph): the difference between currents of the device in dark

and under illumination.

Maximum Power (Pmax) : The maximum power that a cell can produce is

when the product of the current-density (J) and voltage (V) is maximum.

maxmaxmax JVP ×= (1.1)

The Fill Factor (FF) : a quantity that is defined as:

SCOC JV

JVFF

××

= maxmax (1.2)

The power conversion efficiency (η) : defined as the total power extracted

from the device under illumination divided by the total power of light:

light

SCOC

lightin

out

PJV

FFPP

PP ×

×=== maxη (1.3)

A typical result of a J–V measurement on a bulk heterojunction solar cell is

shown in Figure 1.4. The measurements are done under almost standard test

condition (STC). It has to be noted here that, in all experiments mentioned in

this thesis, the temperature of the substrates were not precise 25 ° but close to

this value (25 ° ± 3 °). The single solar cell was based on a 250 nm thick RR-

P3HT : PCBM (1:1) blend.

17

Figure 1.4. The current-voltage (J–V) measurement of an organic bulk heterojunction solar cell based

on a 250 nm thick blend of RR-P3HT : PCBM (1:1) in dark and under illumination. The open-circuit

voltage (VOC), short-circuit current (JSC), photocurrent (Jph), maximum power area (Pmax) are pointed

out.

In an organic bulk heterojunction solar cell, the quantities fill factor (FF), open-

circuit voltage (VOC), and short-circuit current-density (JSC), and therefore the

efficiency (η) of the device are all dependent on morphology,[17] light intensity,[37,38]

thickness of the active layer,[39] electrodes[40] and temperature.[41] A complete

physical description of single organic bulk heterojunction solar cells is given in

references [42,43,44].

1.6 Optical absorption

In this thesis, all bulk heterojunction organic solar cells are based on blends of

conjugated polymers (MDMO-PPV, PFDTBT, RR-P3HT and PTBEHT) with the

fullerene derivative PCBM. However, since the donor materials absorb most of the

sunlight the operation range of these device is mostly governed by the absorption

18


spectra of the polymers used in the blends. The absorption spectra of all donor

materials mentioned in this thesis are plotted in Figure 1.5.

Figure 1.5. The absorption spectra of the donor materials in this thesis. The donor materials RR-P3HT,

PFDTBT and MDMO-PPV absorbs around 550 nm. The PTBEHT has low band gap (1.2 eV) and

absorb mainly at 850 nm.

In addition, the active layer of the device optically serves as a dielectric,

sandwiched between two electrodes. This means that the transmitted light through

the anode and active layer of the solar cell interferes with the light reflected by the

cathode, leading to optical cavity effects inside the device. In the following

chapters, we shall see that this optical cavity effect can be used for the optical

matching of the multilayer structures.

1.7 Outline of this thesis

From the absorption spectra of all common donor materials (Figure 1.5), it is clear

that only a small part of the solar spectrum is used when solar cells are processed

from these materials. Because of the narrow absorption properties of most organic

19

materials used in solar cells, tandem (or multiple) structures must be considered.

When in a tandem cell two donor materials with non-overlapping absorption

spectra are used, the absorption of the device improves. Such a tandem cell can

cover the whole visible light range and also part of the infrared region of the solar

irradiance. In the next chapter, as introduction to multiple organic solar cells, all

kinds of tandem (multiple) organic solar cells are mentioned, which are realized by

different groups until recently. Chapter 3 deals with solution-processed tandem

solar cells having a common middle electrode. In this device, the sub cells (bottom

and top cell) are connected in series by a mainly metallic interlayer that leads to a

high open-circuit voltage. In Chapter 4, a solution-processed tandem solar cell is

realized with four electrodes. An additional solution-processed insulating optical

spacer is used in this 4-terminal tandem solar cell. The sub cells of this device can

be connected in series or parallel outside the device. In chapter 5 the J-V

characteristics of both series- and parallel connected solar cells are quantitatively

discussed. In the last Chapter, an additional application is mentioned for

semitransparent organic solar cells using the same electrodes and processing

methods as the tandem cells mentioned before.

1.8 Abbreviations

Table 1.1 summarizes the most common abbreviations used throughout this thesis.

AM1.5 Air Mass 1.5

MDMO-PPV Poly(2-methoxy-5-(3',7'-dimethyloctyloxy)-1,4-phenylene vinylene)

rr-P3HT Regioregular poly(3-hexylthiophene)

PFDTBT Poly((2,7-(9,9-dioctyl)-fluorene)-alt-5,5-(4',7'-di-2-thienyl-2',1',3'-

benzothiadiazole)

PTBEHT Poly5,7-di-2-thienyl-2,3-bis(3,5-di(2-ethylhexyloxy)phenyl)-

thieno[3,4-b]pyrazine

PCBM The fullerene derivative [6,6]-Phenyl-C61-butyric acid methyl ester

PEDOT:PSS Poly(3,4-ethylene dioxythiophene) : polystyrenesulfonic acid

PTrFE Poly(trifluoroethylene)

FF Fill Factor

Ge-h Exciton (electron-hole pairs) Generation Rate

BHJ Bulk Heterojunction

HOMO Highest Occupied Molecular Orbital

LUMO Lowest Unoccupied Molecular Orbital

JD Current density in dark

JSC Short-circuit current

Jph Net photocurrent

VOC Open-circuit voltage

Vint The internal voltage drop across the device

Vbi Build-in voltage

21

ITO Indium-Tin-Oxide

MEK Methyl Ethyl Ketone

EBL Exciton Blocking Layer

HTL Hole Transport Layer

ELT Electron Transport Layer

Au Gold

Ag Silver

Al Aluminum

Sm Samarium

Chapter 2 A Review on Tandem (multi-junction) Organic Solar Cellsa

Abstract

Many methods and device structures have been realized in the last decade for

organic tandem or multi-junction solar cells. Because of this great variety a general

review of all organic tandem and multi-junction solar cells reported by different

groups from 1990 until recent is given. The device structures are divided into three

different classes; First, the tandem (or multi-junction) solar cells are mentioned that

are based on small organic molecules. The second group is that of hybrid solar

cells, in which the sub cells are fabricated from both small weight organic

molecules and blend of polymers with PCBM. As final group, the tandem (or multi-

junction) solar cells are discussed that are fully based on polymers and PCBM.

a The main results of this Chapter have been accepted for publication as: A. Hadipour, B. de Boer, P. M. W. Blom, Feature Article in Adv. Funct. Mater. submitted (2007).

24


2.1 Introduction In the last decades, solar cells (converting the sunlight into electricity) have

attracted much attention as a proper candidate for the main energy source in the

future. In order to produce low-cost and large-area solar cells, many new device

structures and materials are being developed. For the first time in the 1950s

photoconductivity was reported in organic crystals, such as anthracene.[9,45,10,11]

The first highly conductive polymer was reported in 1977, chemically doped

polyacetylene.[12] Since then, organic molecules and semiconducting polymers are

being increasingly used in (opto)electronic devices. In the last years, the use of

organic materials as active layer in photovoltaic devices has attracted more and

more attention and the total conversion efficiency of those cells has increased

rapidly.[26,46,29,47] First, small (low molecular weight) organic molecules[48,49] and

later also semi-conducting polymers[28,50] were incorporated into solar cells. In

general, illumination of an organic semiconductor leads to the creation of excitons [51] with a binding energy of about 0.4 eV [20,21,22], instead of free charges. The

exciton can be separated when it reaches the interface between suited donor (D)

and acceptor (A) materials, where the difference in the electron affinities and the

ionization potentials between those two (A & D materials) are sufficiently large to

overcome the exciton binding energy. The hole and electron are still Coulombically

bound across the D/A interface, even though the opposing charges reside in

different organic materials. After breaking the Coulomb binding between the

electron in the acceptor and the hole in the donor, the electrons and holes are

subsequently transported through the acceptor and donor phase, respectively, to

the electrodes of the device. However, the low mobility of electrons and holes,

together with relatively narrow absorption spectra of the organic materials, lead to

a relatively low performance of the organic solar cells. The efficiency typically

amounts to 4–5%, which limits them for practical applications. Excellent reviews

have appeared in literature regarding these so-called single active layer or mono

junction solar cells.[47,52,53,54,55,56] To improve the absorption of the solar radiation

by organic solar cells, materials with a broad absorption band have to be designed

and produced, or different narrow band absorbers have to be stacked or mixed in

25

multiple junctions.[49] When two (or more) donor materials with partially non-

overlapping absorption spectra are used in a tandem (or multi-junction) solar cell, a

broader range of the solar spectrum (whole visible and part of the IR range) can be

covered. There are several approaches for organic tandem (multiple) cells reported

in the last years, depending on the materials used for the active layer and the

proper separation or recombination layer(s). All layers can be different in each

architecture or approach. In general, the multiple organic solar cells, reported until

today, can be divided in three classes;

A) Tandem (or multi-junction) organic solar cells in which low molecular weight

molecules are used for both the bottom (front) and the top (back) cells,

B) Hybrid tandem organic solar cells in which the bottom cell is processed from

polymers by solution-processing, while the top cell is made of vacuum-deposited

low molecular weight molecules,

C) Fully solution-processed tandem or multi-junction organic solar cells in which

both the bottom and the top cells are made of polymers.

An overview of the chemical structures of organic materials presently used in

organic tandem and multi-junction solar cells is given in Figure 2.1. Depending on

which kinds of materials are being used for the active layers, different separating

layers are fabricated and reported. In the following sections, the different types of

organic tandem and multi-junction photovoltaic cells are described and recent

results obtained by various groups are presented.

26


27

Figure 2.1. The chemical structures of the donor and the acceptor materials used in organic tandem

and multi-junction solar cells. The molecules above the solid line are frequently used in multi-junction

photovoltaic cells fabricated by vapor deposition, whereas the polymers and molecules depicted below

the solid line are used in solution-processed multi-junction solar cells.

2.2 A. Tandem and multilayer organic solar cells based on low molecular weight molecules The main advantage of using low molecular weight or small molecules for tandem

structures is that different layers of donor and acceptor (or mixed layers) materials

can be evaporated (or co-evaporated) with sharp interfaces on top of each other,

without affecting the already existing layers. A disadvantage of such structures is the

relatively low evaporation rate of active materials, which limits the processing speed

for large-area applications. The first organic tandem (double junction) cell was

realized by Hiramoto, Suezaki and Yokoyama and was constructed from two

identical bilayers.[57] Each bilayer consisted of H2-phthalocyanine (50 nm) and a

perylene tetracarboxylic derivative (70 nm), and the two bilayer cells were separated

by a thin interstitial layer (2 nm) of Au. This first organic tandem solar cell resulted in

almost a doubling of the open-circuit voltage (VOC) to 0.78 V.

Yakimov et al. presented the first multiple-heterojunction solar cells by stacking

two, three or five vacuum-deposited ultrathin organic bilayer photovoltaic cells in

series.[58] All single thin heterojunction cells[59] (bottom and top cell) were made of

Cu-phthalocyanine (CuPc) as donor and 3,4,9,10 perylenetetracarboxylic bis-

benzimidazole (PTCBI) as an acceptor. The device was processed on an indium-

tin-oxide (ITO) substrate covered by 30 nm poly(3,4-ethylene dioxythiophene) :

polystryrenesulfonic acid (PEDOT:PSS), which served as anode of the device. The

sub cells were deposited by thermal evaporation in vacuum of ~10-6 Torr, starting

with the donor material (CuPc) and followed by the acceptor material (PTCBI). A

thin layer of silver clusters (0.5 nm) was deposited between the two sub cells as

recombination layer and finally 80 nm Ag was thermally deposited for the cathode

of the device. Generally, the built-in voltage of a single heterojunction cell is given

by the difference in the Fermi levels of the donor and the acceptor materials

used.[26] This means that for a well-performing tandem cell, consisting of two (or

more) series-connected heterojunction single cells, the open-circuit voltage of the

tandem cell should be equal to the sum of the built-in voltages of the individual

28


cells. However, deposition of the sub cells in series without a separation layer in

between them will cause the formation of an inverse heterojunction between the

donor layer of the top (back) cell and the acceptor of the bottom (front) cell.

Yakimov et al. have placed a very thin and discontinuous layer of silver (0.5 nm

Ag) between each sub cell. The Ag clusters then serve as charge recombination

sites. The electrons and the holes arriving from the bottom and the top cell,

respectively, can recombine at this separating layer. The structure of the small

molecule tandem cell is shown in Figure 2.2.

Figure 2.2. The structure of an organic tandem solar cell based on the small molecules CuPc (donor)

and PTCBI (acceptor) as described in ref. [58]. The 0.5 nm Ag separation layer provides recombination

sites for the free charges arriving from the bottom and top cell.

After absorption of the incident light in such a structure, excitons are created in

both the donor (D) and acceptor (A) material. Only excitons that are created very

close (in range of the exciton diffusion length, ~10 nm)[60] to the D/A interface can

be separated into charges. Then, the electrons and the holes travel through the

acceptor (PTCBI) and donor (CuPc) layers, respectively, to the contacts of the

device. The holes of the bottom cell and the electrons of the top cell are extracted

from the device, whereas the electrons of the bottom cell recombine with the holes

of the top cell at the metallic interlayer (Ag nanoclusters). It was demonstrated that

for a thickness of about 11 nm for the donor and also 11 nm for the acceptor layer

(22 nm heterojunction) the open-circuit voltage (VOC) and short-circuit current (JSC)

29

are maximized.[58] The VOC and JSC of tandem cells with two stacked sub cells and

varying thickness for the metallic interlayer (Ag) are summarized in Table 2.1.

Ag thickness [nm] 0 0.5 1.5 3

VOC [Volt] 0.45 0.9 0.9 0.9

JSC [A/m2] 26 63 52 39

Table 2.1. Data extracted from reference [58]: For a Ag interlayer with a thickness of 0.5 nm, the

tandem cell has the highest VOC and JSC. For a thinner Ag interlayer, both open-circuit voltage and short-

circuit current drop. For a thicker Ag interlayer the VOC has the same values while JSC drops.

As the data in Table 2.1 show, the use of a thicker Ag interlayer leads to a much

lower photocurrent, while the VOC is the same. The high absorption coefficient of

the silver interlayer leads to a reduction of the light intensity at the top cell and

therefore the top cell generates a lower current for a thicker Ag interlayer. The sub

cells were connected in series, meaning that the current of the tandem cell is

limited by the lower current of the two sub cells, which is the top cell. With the

same donor and acceptor materials, Yakimov et al. demonstrated also triple and

fivefold heterojunction solar cells measured at different light intensities. A summary

of their results of VOC and power conversion efficiency (η) of different cells under 1

and 10 sun illumination is given in Table 2.2.

30


Stack cells: VOC [volt]

(1 sun)

VOC [volt]

(10 sun)

η [%]

(1 sun)

η [%]

(10 sun)

Single cell 0.45 0.43 1.0 0.5

Tandem cell 0.90 0.90 2.5 1.8

Triple cell 1.20 1.40 2.3 2.6

Fivefold cell 1.23 1.70 1.0 1.4

Table 2.2. The open-circuit voltage (VOC) and power conversion efficiency (η) of single and multilayer

cells under different light intensities are extracted from reference [58]. Under 1 sun illumination

condition, the dual structure (tandem cell) has the best performance. At higher light intensity of 10 sun,

the triple cell reaches the highest efficiency.

In order to explain the results mentioned above we have to note that the total current

extracted from the tandem structure is directly dependent on how efficient charges

can recombine at the metallic interlayer (0.5 nm Ag). All sub cells have to generate

the same amount of photocurrent (current matching), since they are connected in

series. If one of the individual cells generates much more current than the other(s),

charges will pile-up at the very thin Ag interlayer such that the effective bias of the

sub cells will be far off from their best performance. In addition, in a single bilayer

heterojunction cell the excitons must be generated sufficiently close to the D/A

interface in the range of LD ~10 nm [60] such that exciton dissociation can take place.

Therefore, the performance of organic bilayer devices is limited by their exciton

diffusion length (LD ~10 nm), which is much shorter than the thickness needed to

absorb all incident light (LA ~150 nm). Stacking of cells in a tandem structure is the

way to overcome this limitation. A problem that remains when three or five identical

cells are stacked is that under 1 sun illumination the absorption of light by the first

two sub cells leads to a reduction of the photocurrent of the following cells, limiting

the total current extracted from the whole device. At higher light intensity (10 sun),

31

the triple cell (instead of the tandem cell) shows the maximum efficiency, but again

the light intensity is decreased for the following sub cells limiting the performance of

the complete 5 layer (penta junction) device. Similar experiments using CuPc and

PTCBI in combination with ultrathin Ag and Au interlayer were performed by Triyana

and co-workers to produce tandem and triple junction solar cells.[61]

To improve the efficiency of stacked solar cells, Xue and Forrest applied several

modifications to the structure of the device mentioned above.[62] First, they used C60

as acceptor since it has a longer exciton diffusion length of LD ~ 40 nm as compared

to the previous acceptor PTCBI.[49] Secondly, they incorporated a mixed donor-

acceptor layer (bulk heterojunction) sandwiched between pure donor and pure

acceptor layers. The homogenous layers of C60 and PTCBI serve as electron

transport layer (ETL) and hole transport layer (HTL), respectively. The device is

called hybrid planar-mixed heterojunction (PM-HJ) device.[62,63] Instead of a very thin

Ag interlayer for the recombination sites (separating layer between two sub cells),

they used Ag nanoclusters with a typical thickness of 0.5 nm but buried in a 5 nm

thick 4,4′,4′ ′-tris(3-methyl-phenyl-phenyl-amino)triphenylamine (m-MTDATA), which

was p-doped with 5 mol % tetrafluoro-tetracyano-quinodimethane (F4-TCNQ).[62] The

reason for using an additional layer of p-doped m-MTDATA is not explained by the

authors. Clearly, advantages are found when using doped transport layers

namely:[64,65] doping of this interlayer increasing the conductivity of this layer and

reduces Ohmic losses, quenching processes at the electrode are avoided since

excitons created in the active layer cannot enter the wide-gap transport layer, the

thickness of the highly conductive interlayer can be tuned to optimize the optical field

distribution in the solar cell, and the increase in overall thickness of the devices

allows higher stability and a lower probability for short circuiting the device. It has to

be noted that silver particles also serve as scattering centers for incident light. The

scattering of light in the middle of the device improves the optical absorption of the

active layers of the sub cells. Finally, thin layers of bathocuproine (BCP)[59] and

(PTCBI) are used as the exciton-blocking layer (EBL) in the top (back) sub cell and

bottom (front) sub cell, respectively. The structure of the whole device is depicted in

Figure 2.3.

32


Figure 2.3. The complete stacked structure of PM-HJ tandem organic solar cell as described in

reference [62].

The exciton blocking layers (EBL) serve in three different ways:

(a) for the top cell it prevents damage of the active layer caused by evaporation of a

(hot) metallic cathode (hot Ag particles),

(b) it eliminates diffusion of Ag particles, which act as exciton quenching sites, into

the active layer,

(c) the EBL layers provide an optical spacer between the active layer and the

reflecting cathode of each sub cell.

In this way, the light intensity can be tuned to be maximized at the D/A interface (or

at the mixed donor-acceptor layer), which leads to an improved optical absorption in

these devices. Because of the low thicknesses of the active layers (limited by the

short exciton diffusion length) optical interference between the transmitted light

through the active layer of the device and the reflected light at cathode of the device

has a significant effect on the total absorption of each sub cell. By using different

33

thicknesses for the active layers, the optical profiles can be tuned inside the device.

The donor material CuPc absorbs incident light between 550 nm and 750 nm

wavelengths, while the acceptor material C60 has very high absorption between 300

nm and 400 nm and relatively lower absorption between 400 nm and 600 nm with a

maximum at 550 nm. Using a thicker homogeneous CuPc layer and thinner

homogeneous C60 layer in the bottom cell as compared to the top cell (keeping the

thickness of mixed donor-acceptor layer fixed), the absorption of the bottom cell is

shifted to longer-wavelengths relatively to the absorption of the top cell. Such an

asymmetric absorption of both sub cells leads to an improved spectral overlap with

the sunlight. For current matching, the thicknesses of the homogeneous layers of

CuPc and C60 and mixed layers have to be further optimized. It is also important to

realize that molecular intermixing (mixed layer of each sub cell) leads to significantly

lower charge carrier mobilities compared to those in polycrystalline homogeneous

layers (CuPc and C60 layers), whereas the exciton separation efficiency is much

higher in the mixed layer as compared to pure CuPc or C60 layers. The power

conversion efficiencies of PM-HJ tandem cells using different thicknesses for the

active regions are summarized in Table 2.3. All thicknesses are in nm beginning with

the bottom cell followed by the top cell.

Cell CuPc CuPc:C60 C60 PTCBI CuPc CuPc:C60 C60 BCP η [%]

1 10 18 2 5 2 13 25 7.5 5.4

2 7.5 12.5 8 5 6 13 16 7.5 5.7

3 9 11 0 5 5 10 21 10 5.0

Table 2.3. Layer thicknesses in nanometers and efficiencies (%) of various PM-HJ tandem organic solar

cells from reference [62]. The effect of thickness variations for the active layers is shown. Cell 2 has the

best performance because the thicknesses used are the best tradeoff between the electrical and optical

properties.

34


2.3 B. Hybrid tandem organic solar cell

In order to improve the optical absorption of a single organic solar cell Dennler et

al. have stacked a solar cell made of the small molecules zinc phthalocyanine:C60

(ZnPc:C60) on top of a heterojunction polymer solar cell based on a mixed layer of

poly(3-hexylthiophene) (P3HT) and [6,6]-phenyl C61-butyric acid methyl ester

(PCBM).[66] For the two sub cells two different classes of materials (low molecular

weight molecules and polymer) were used, which require different processing

techniques, namely vapor deposition and spin coating. Therefore, this structure is

called a hybrid tandem organic solar cell. The tandem cell was processed onto

indium-tin-oxide (ITO) coated glass covered by a 90 nm thick layer of PEDOT:PSS.

A layer (clusters) of 1 nm gold (Au) was used as intermediate layer between the

two sub cells, which served as a recombination center. The P3HT polymer was

dissolved in chlorobenzene and PCBM was dissolved in dichloromethane. The

layer of P3HT was spin coated first followed by spin coating of the PCBM layer.

This structure is called a diffused bilayer[66] or stratified bilayer.[67] The rest of the

device was processed by thermal evaporation under vacuum (10-6 mbar). First, 10

nm C60 was evaporated serving as electron transport layer (ETL) for the bottom

cell. Then, 1 nm Au was evaporated to enhance the recombination of hole and

electrons from the top and bottom cell, respectively. The top cell was subsequently

fabricated by thermal evaporation of 10 nm of zinc phthalocyanine (ZnPc), 40 nm

of a mixture of ZnPc and C60 and 15 nm of C60. The device was completed by

evaporation of 5 nm of chromium (Cr) and 95 nm of aluminum (Al), which acts as

the cathode of the device. In this way, the bottom cell was not affected or damaged

during the processing of the top cell. The structure of the device is shown in Figure

2.4.

35

Figure 2.4. The structure of a hybrid tandem organic solar cell. The top cell, made of small molecules,

is evaporated on top of the bottom cell based on a diffused bilayer of the conjugated polymer P3HT and

the fullerene PCBM, which is processed from solution.

In order to compare the performance of the tandem cell with the single sub cells,

Dennler et al. have also processed two reference single cells. They used

ITO/P3HT:PCBM/Al processed as diffused bilayer as a reference for the bottom

cell and ITO/ZnPc/ZnPc:C60/C60/Cr/Al as a reference for the top cell. The bottom

cell has its main absorption between 375 and 630 nm, while the top cell mainly

absorbs in the range of 600 to 800 nm. As a result, the tandem cell covers the

whole visible range of the solar spectrum (from 400 to 800 nm). The measured

performances of the tandem device and reference sub cells under 1 sun simulated

light are summarized in Table 2.4.

36


Cell JSC [A/m2] VOC [volt] FF [%] η [%]

Ref. bottom 85 0.55 55 2.6

Ref. top 93 0.47 50 2.2

Tandem 48 1.02 45 2.3

Table 2.4. Performance of the hybrid tandem organic solar cell [66]. The performance of the tandem cell

is compared with the reference bottom and the reference top cells.

The results in Table 2.4 confirm the successful coupling of the two sub cells in

series using the 1 nm Au recombination layer, since the VOC of the tandem cell was

equal to the sum of the VOC of the bottom and the top cell. However, the

performance of the tandem cell was limited by the lower photocurrent and fill factor

(FF) extracted from the device. The thicknesses of the active layers of the sub cells

were not yet optimized. Also the photocurrent of the tandem, reference bottom and

reference top cells were measured under monochromatic light as a function of

wavelength. By using monochromatic incident light the J–V characteristics of each

individual sub cell can be measured and compared with the performance of the

tandem cell, as summarized in Table 2.5.

Cell JSC [mA/W]

λ = 500 nm

JSC [mA/W]

λ = 600 nm

JSC [mA/W]

λ = 700 nm

Ref. bottom 138 120 10

Ref. top 32 100 110

Tandem cell 8 51 2

37

Table 2.5. The JSC of the tandem cell, reference bottom and reference top cell measured under

monochromatic light with different wavelength [66].

Table 2.5 demonstrates that the tandem cell is only efficient when both sub cells

perform properly. If one of the individual cells generates a low current, the current

extracted from the tandem device is limited due to the series connection. As

mentioned above, the main advantage of hybrid tandem organic solar cells is the

absence of solvents for the processing of the top cell since the top cell has to be

vapor deposited. However, using both processing methods may lead to higher

costs of producing tandem (multiple) solar cells for commercial applications.

2.4 C. Solution-processed tandem organic solar cells

Semiconducting polymers are appropriate materials for developing low-cost

technologies for large-area solar cells, since polymers can be deposited from

solution using simple methods as spin coating and ink-jet printing.[7,8] The main

problem to fabricate polymer tandem cells is the stack integrity: deposition of the

top cell might dissolve or damage the earlier deposited bottom cell, especially

when similar solvents as chlorobenzene and chloroform are used. Therefore, a

separating layer (middle contact) is required that has to be thick enough (a closed

layer) to protect the bottom cell from dissolving during spin coating (processing) of

the top cell. At the same time, the middle contact has to be as transparent as

possible to transmit light efficiently to the top cell. In order to overcome the stack

integrity problem different structures and methods have been reported recently. In

Chapter 3 the first solution-processed polymer tandem solar cell is reported in

which sub cells have different optical absorption properties. In Chapter 4 a 4-

electrode polymer tandem solar cell is considered using an optical spacer, which is

the only 4-electrode structure reported until today. Recent alternative approaches

to solution processed organic tandem cells are discussed in the following sections.

38


2.4.1 Tandem organic solar cell processed on separated substrates

One of the methods to overcome the processing difficulties was reported by

Shrotriya et al..[68] Two identical bulk heterojunction single cells were fabricated

onto different glass substrates and then positioned on top of each other. The sub

cells were connected in series or in parallel outside the device. The bottom cell has

a semitransparent cathode consisting of 1 nm thick lithium fluoride (LiF), 2.5 nm

aluminum (Al) and 12.5 nm gold (Au) with a maximum transparency of about 74 %

at 580 nm. Both sub cells were fabricated from a blend of poly(2-methoxy-5-(2′-

ethyl)-hexyloxy)-1,4-phenylene vinylene (MEH-PPV) and PCBM with a 1:4 weight

ratio. The thickness of the active layer of both the bottom and the top cell was

about 70 nm. For PPV-based solar cells it is known that an increase of the active

layer thickness does lead to an increased absorption, but not to an increased

performance [69]. The absorption increase is cancelled by increased space-charge

formation and recombination. Using two thin identical active layers for both sub

cells is an attempt to effectively increase the absorption while maintaining the

favorable electrical properties of thin devices. However, the coverage of the solar

spectrum was not improved. The structure of the device from reference [68] is

given in Figure 2.5. Furthermore, various semitransparent cathodes have been

investigated. The highest transparency and best performance was obtained with

the above-mentioned LiF/Al/Au cathode. Finally, it should be noted that the

efficiency of ~2.5% of both the parallel and series tandem cell is equal to the

efficiency of a single layer device with a thickness of 140 nm. Apparently, the

coupling of two identical thin cells did not bring advantage as compared to a single

cell with the same total active layer thickness. Shrotriya et al. have connected the

sub cells in series and in parallel. They have compared the performance of the

bottom, the top and the connected devices (series and parallel) under 1 sun

simulated solar light. A summary of their results is given in Table 2.6.

39

Figure 2.5. The separated sub cells positioned on top of each other as described in Ref [68]. The

bottom cell has semitransparent cathode.

Cell VOC [Volt] JSC [A/m2] FF [%] η [%]

Bottom 0.86 26 45 1.1

Top 0.86 32 45 1.3

Series connected 1.64 34 45 2.4

Parallel

connected 0.84 63 45 2.5

Table 2.6. Performance of stacked organic solar sell processed on separated substrates [68]. For the

measurements of the devices connected in series and in parallel, the bottom cell was positioned in front

of the top cell, allowing the light to transmit through the bottom cell before reaching the active layer of

the top cell.

2.4.2 Tandem organic solar cell with ITO as separating layer

40


Another polymer based tandem cell is recently reported by Kawano et al.. Two bulk

heterojunction solar cells were electronically coupled with a interlayer of indium-tin-

oxide (ITO)[70]. The stacked bulk heterojunction solar cells were both based on a 80

nm blend of poly[2-methoxy-5-(3′,7′-dimethyloctyloxy)-1,4-phenylene vinylene]

(MDMO-PPV) and PCBM (1:4 weight ratio). As the separating layer, Kawano et al.

deposited ITO by dc magnetron sputtering in 1 Pa of argon gas without substrate

heating. The presence of argon gas into the deposition chamber prevented

damaging of the active layer of the bottom cell during deposition of the ITO layer.

The full structure of the device is given in Figure 2.6.

Figure 2.6. The structure of tandem cell as fabricated in Ref [70] is shown. A layer of transparent ITO

separates the two sub cells.

The most important aspect of this report is the method used to sputter the ITO onto

the polymer surface without damaging it. The very high transparency of the ITO

layer enhances the light intensity reaching the top cell resulting in an increased

photocurrent of the top cell. A summary of the results achieved by the authors

using a 20 nm ITO interlayer is given in Table 2.7.

41

Cell VOC [Volt] JSC [A/m2] FF [%] η [%]

Reference 0.84 46 59 2.3

Tandem 1.34 41 56 3.1

Table 2.7. Performance of a tandem organic solar cell with 20 nm ITO as separating layer [70] is

demonstrated. The performance of the tandem cell is compared to a reference BHJ single cell (using a

ITO/PEDOT/80 nm MDMO-PPV:PCBM/Al structure).

As the data of Table 2.7 demonstrate, connection of the two sub cells in series

leads in this case to a VOC of 1.6 times higher than the reference cell, instead of the

double value. The reason is that the high work function of ITO (~ 4.8–5.0 eV) limits

the maximum VOC of the bottom cell.[43,44] Thus, the absence of an Ohmic contact

for electron extraction at the interface between the separating layer and the active

layer of the bottom cell (LUMO of the PCBM in the blend) leads to a decrease of

the VOC of the bottom cell from 0.84 V (normal value when using an Ohmic contact)

to 0.48 V, leading to a VOC of the tandem cell of 1.34 Volt. Since in this study the

same donor material (MDMO-PPV) was used for both sub cells the narrow

absorption of the active layer was not improved. However, optimizing this structure

using a transparent, low work function, metallic layer between the bottom cell and

the ITO interlayer, in combination with donor materials with partially non-

overlapping absorption band can lead to the fabrication of highly efficient tandem

cells.

2.4.3 Multiple organic solar cell with solution-processed ZnO interlayer

One of the latest approaches to solution-processed organic tandem solar cell was

reported by Gilot, Wienk and Janssen in which a solution-processed middle

electrode (separating layer) was introduced.[71] These kinds of electrodes

demonstrate the possibility to realize fully solution-processed solar cells (tandem or

multiple cells) without using vapor-deposited contacts. For the fabrication of the

42


separating layer, ZnO nanoparticles were prepared,[72,73] dissolved in acetone, and

spin coated. In order to create a high work function anode neutral pH PEDOT

(ORGACON, pH = 7, 1.2 w%, Agfa Gevaert NV) was spin coated from a water-

based suspension. The normal acidic PEDOT:PSS can not be used since the ZnO

layer dissolves in an acidic solution. The ZnO layer serves as electron transporting

layer (ETL), while the neutral PEDOT acts as hole transporting layer (HTL). The

active layers were fabricated from a chlorobenzene solution of the donor materials

MDMO-PPV or P3HT mixed with the acceptor material PCBM in a 1:4 and 1:1

ratio, respectively. The thicknesses of the active layers were 45 nm and 85 nm for

the bottom and top cell of the tandem junction, and 45 nm, 65 nm and 85 nm for

the bottom, middle, and top cell of the triple junction. The structure of this tandem

device is given in Figure 2.7.

Figure 2.7. Structure of tandem solar cell based on ZnO/PEDOT as separating layer. The MDMO-

PPV:PCBM (1:4) layers are 45 nm and 85 nm thick for the bottom and the top cell, respectively.

The recombination of charges across the interface between ZnO and neutral pH

PEDOT is normally very poor, since a large energy offset exists at the interface

between them. An Ohmic contact between ZnO and PEDOT is needed to prevent

a voltage drop across the interface and allows the holes and the electrons to

recombine efficiently. In order to increase the recombination probability, the two

43

materials have to be sufficiently doped or metallic clusters (0.5 nm Ag or Au) have

to be deposited between them. In this study doping methods have been applied:

The PEDOT is all p-doped, whereas the doping of the ZnO layer can be reached

by exposure the layer to UV light (photo-doping) for a few seconds.[74,75] A

summary of their results for tandem and triple solar cells using MDMO-PPV:PCBM

for all active layers and photo-doped ZnO/neutral pH PEDOT layer is given in

Table 2.8.

Cell VOC [Volt] JSC [A/m2] FF [%]

Single Cell 0.82 49 57

Tandem Cell 1.53 30 40

Triple Cell 1.92 24 33

Table 2.8. Performance of multi-junction organic solar cell with a solution-processed interlayer as taken

from reference [71]. The tandem cell has a 45 nm active layer for the bottom and 85 nm layer for the top

cell. The triple cell is based on 45 nm, 65 nm and 85 nm active layers for the bottom, the middle and the

top cell, respectively.

As the data of Table 2.8 demonstrate, the VOC of the multi-junction solar cells

connected in series are slightly lower than the sum of the VOC‘s of the individual

sub cells. The deviation of the VOC from estimated values (two or three times the

VOC of the single cell) is probably caused by a small voltage drop at the interface

between the ZnO and neutral pH PEDOT. The very high transparency of the

ZnO/PEDOT interlayer in combination with the possibility to fabricate the whole

multi-junction solar cell from solution makes this approach very useful and practical

for future applications.

44


2.4.4 Tandem organic solar cell with solution-processed TiOx interlayer

The most recent and efficient tandem organic solar cell based on solution

processing was reported by Kim et al.[76] In this tandem cell a highly transparent

titanium oxide (TiOx) layer was used to separate the two sub cells of the tandem

device. This transparent middle electrode demonstrates the possibility to realize

fully solution-processed solar cells (tandem or multiple cells) without using vapor-

deposited contacts. The advantages of using oxides such as the above-mentioned

ZnO or TiOx for the middle electrode is the optical transparency of the layer and the

orthogonally compatible solvents used for processing all layers. In these cases, the

separating layer does not significantly affect the light intensity. The increased light

intensity at the top cell leads to a higher photocurrent generated by the top cell.

Therefore, the efficiency of the tandem device is not limited any more by the lower

current of the top cell as observed for tandem structures with a metallic interlayer.

The cell was processed onto a glass/ITO substrate covered by a 40 nm thick layer

of PEDOT:PSS (Baytron P). The transparent TiOx interlayer was fabricated by spin

coating from methanol solution by means of sol-gel chemistry.[77] For the bottom

BHJ cell a 130 nm thick layer of PCPDTBT:PCBM (1:3.6)[78] was used that was

processed from chlorobenzene. The top BHJ cell was fabricated from a 170 nm

thick P3HT:PC70BM (1:0.7) blend, which was processed from chloroform. The two

sub cells have complementary absorption spectra, which leads to coverage of the

whole visible and part of the infrared of solar spectrum by the tandem device. On

top of the TiOx interlayer layer, the highly conductive hole transport layer poly(3,4-

ethylene dioxythiophene) : polystryrenesulfonic acid (PEDOT:PSS, Baytron

PH500) was spin coated. The structure of the device is given in Figure 2.8.

45

Figure 2.8. The structure of the tandem device from reference [76]. The two sub cells are separated by

a highly transparent layer of TiOx covered by the highly conductive PEDOT:PSS (PH 500).

The performance of the above-mentioned tandem device in the dark and under

illumination with different light intensities was measured and the results of these

measurements are summarized in Table 2.9.

Light intensity [W/m2] 200 500 1000 2000

FF [%] 0.68 0.67 0.66 0.64

JSC [A/m2] 24 38 75 150

VOC [volt] 1.15 1.2 1.25 1.28

η [%] 6.7 6.5 6.3 6.2

Table 2.9. The data extracted from reference [76]. An optimum tandem solar cell is measured under

different light intensities.

As listed in Table 2.9, this all-solution processed tandem cell has the highest

efficiency reported until now of all tandem organic solar cells. The power-

conversion efficiency reaches more than 6 % at intensities above 200 milliwatts per

46


square centimeter. The main reasons for the high efficiency in this series tandem

cell is the relatively high photocurrent of both the sub cells due to the transparent

separating interlayer.

2.5 Summary

In order to achieve organic solar cells with higher performance, the narrowness of

the optical absorption of the active layer has to be improved. By using two (or

more) donor materials with partially non-overlapping absorption spectra in a

tandem or multi-junction structure, the whole visible range of the solar light can be

absorbed and even extended into the near IR. In addition to covering a larger part

of the spectrum, tandem solar cells offer the distinct advantage that the photon

energy is used more efficiently, because the voltage at which charges are collected

in each sub cell is closer to the energy of the photons absorbed in that cell. In the

various tandem structures the layer that separates both sub cells plays a very

important role. The final efficiency of the tandem cell depends directly on the

electrical (Ohmic contact and proper conductivity) and optical (transparency)

properties of this middle electrode. In the series configuration, the middle contact

serves as recombination site only and therefore does not have to be highly

conductive. In the parallel configuration, the sheet-conductivity has to be high since

the contact is also used to extract the charges from the device. This overview

demonstrates that recently developed multi-junction photovoltaic cells with

optimized materials and thicknesses lead eventually to higher efficiencies as

compared to single layer solar cells. Recently, power-conversion efficiencies of

more than 6 % have been reported using solution-processed organic bulk

heterojunction tandem solar cells. The realization of working organic tandem and

multi-junction photovoltaic cells is an important step forward to the improvement

and finally commercialization of large-area organic solar cells.

47

Chapter 3 Tandem Polymer Solar Cell Using a Composite Metal Interlayerb Abstract

This Chapter deals with a solution-processed polymer tandem cell by stacking two

single cells in series. The two bulk heterojunction sub cells have complementary

absorption maximal at λmax ~ 850 nm and λmax ~ 550 nm, respectively. A composite

middle electrode is applied that serves both as charge recombination center and as

a protecting layer for the first cell during spin coating of the second cell. The sub

cells are electronically coupled in series, which leads to a high open-circuit voltage

of 1.4 V, equal to the sum of each sub cell. The layer thickness of the first (bottom)

cell is tuned to maximize the optical absorption of the second (top) cell (optical

matching). In the end a tandem cell with an unbalanced performance of the sub

cells is compared with an optimized tandem cell.. It becomes clear that the low light

intensity due to high optical absorption of the middle contact leads to a low photo-

generated current by the top cell. In series configuration, this low current of the top

cell of the device limits the current extracted from the tandem device.

b The main results of this Chapter have been published as: A. Hadipour, B. de Boer, J. Wildeman, F. B. Kooistra, J. C. Hummelen, M. G. R. Turbiez, M. M. Wienk, R. A. J. Janssen, and P. W. M. Blom, Advanced Functional Materials 2006, 16, 1897-1903.

48


3.1 Introduction Most solution processable semiconducting polymers exhibit band gaps larger than

1.2 eV and narrow absorption bands, which results in a limited overlap of their

absorption spectrum with the solar spectrum. Since a very broad absorption

spectrum is not feasible in one single polymeric material, a tandem cell, consisting

of multiple layers each with their specific absorption maximum and width, can

overcome this limitation. In addition to covering a larger part of the spectrum,

tandem solar cells offer the distinct advantage that the photon energy is used more

efficiently, because the voltage at which charges are collected in each sub cell is

closer to the energy of the photons absorbed in that cell. This was demonstrated

for organic tandem photovoltaic cells processed by co-evaporation of small organic

molecules.[57,59,62,79,58,80,61] In this method, layers of organic molecules are thermally

deposited on top of each other. If the tandem cell is stacked from two photovoltaic

sub cells that were fabricated from the same conjugated polymer [70,68], the

absorption spectra of both cells are identical and, consequently, can only be used

for very thin organic films that do not absorb the intense solar flux (even with an

aluminum mirror as top electrode). In such a tandem cell the coverage of the solar

spectrum is not improved but more light of the same wavelength is absorbed. In

this Chapter a solution-processed tandem solar cell is discussed consisting of two

bulk heterojunction sub cells with complementary absorption spectra. In order to

harvest as many photons of the solar spectrum as possible, the tandem cell

(Figure 3.1C) consists of two cells with complementary absorption spectra (Figure

3.2). A composite middle electrode separates the two sub cells. The middle

electrode serves two different purposes; as a charge recombination center, and as

a protecting layer for the first cell during spin coating of the second cell. In general,

because the sub cells have completely different absorption spectra, high energy

photons are absorbed in the bottom cell and low energy photons are absorbed in

the top cell. Consequently, the whole visible part of the solar spectrum can be

covered. The sub cells are electronically coupled together in series, which results

in an open-circuit voltage (VOC) of the tandem cell that equals the sum of the VOC of

each sub cell. Attenuation of incident light on the top cell, caused by absorption

49

and the optical cavity of the bottom cell, leads to the lower photocurrent of the top

cell, which limits the current of the tandem cell. The structures of reference bottom

(Figure 3.1A), reference top cell (Figure 3.1B) and the tandem cell (Figure 3.1C) is

shown in figure 3.1.

Figure 3.1. A) Schematic structure of the reference bottom cell to determine the optical out coupling. B)

The schematic structure of the reference top cell, which is the same as the top cell in tandem

configuration but processed directly on glass. C) The schematic structure of the tandem cell. Two bulk

heterojunction cells (bottom and top cell) are stacked in series. The absorption spectra of the two

semiconducting polymers PFDTBT and PTBEHT are complementary.

50


3.2 The middle electrode

The two organic solar cells are electronically linked together by a composite

electrode. Since the two sub cells are connected in series, this composite middle

electrode provides a site for recombination of electrons and holes approaching

from bottom and top cell, respectively. Therefore, the middle electrode does not

have to be highly conductive, since, in the tandem cell, no current is extracted from

this electrode. Furthermore, to efficiently extract the electrons from the bottom cell

and the holes from the top cell (in order to recombine at the middle electrode), the

middle electrode has to provide an Ohmic contact for both sub cells, since an

insufficient recombination results in limited currents in both devices. The middle

electrode has to act as the cathode (with a low work function to align with the

LUMO of the acceptor) for the bottom cell and, at the same time, as the anode

(with a high work function to align with the HOMO of the donor) for the top cell.

With a gold middle electrode, an Ohmic contact for both holes and electrons can

not be obtained and, therefore, we added a very thin layer of 0.5 nm LiF and 0.5

nm Al on top of the bottom cell. The Ohmic contact provided by LiF/Al with the

LUMO level of PCBM will extract the electrons generated in the bottom cell. Next,

we evaporated 15 nm Au and topped it off by spin coating 60 nm PEDOT:PSS.

The latter provides a stable and Ohmic contact for hole extraction from the top cell.

The extracted electrons and holes then recombine in the middle electrode. A very

thin layer of gold (as thin as 0.5 nm) provides already enough recombination sites for the holes and electrons that arrive at the middle electrode.[79] However, in our

case the gold layer has to protect the LiF/Al cathode of the bottom cell from being

dissolved and oxidized during the spin coating of the water-based PEDOT:PSS.

This means that the gold layer has to be a continuous, closed layer. On the other

hand, the middle contact has to be as thin as possible to maximize the light

transmission that is required as the input of the top cell (Section 3.3). We found

that the thinnest possible layer that is required to protect the bottom layer during

solution processing of PEDOT:PSS cell is 10–15 nm. At that thickness the gold

layer forms a closed and continuous electrode, which can protect the cathode of

51

the bottom cell from being destroyed (the thickness that is required to form a

closed layer depends on the roughness of the blend used for the bottom cell). The

total composite middle electrode (LiF/Al/Au/PEDOT:PSS) protects the bottom cell

from being dissolved during the spin coating of the active layer of the top cell.

3.3 Optical matching

Due to the relatively narrow absorption bands of the semiconducting polymers

presently used in solar cells, the small band gap polymers will not significantly

contribute to a better performance of a single cell (assuming similar transport

properties), because the absorption spectrum is only shifted under the solar

emission spectrum. Small band gap polymers can improve the overall efficiencies

more significantly if used in a tandem cell. In such a tandem cell the utilization of

two absorbing polymers with (partially) complementary absorption spectrum results

in an enhanced photon harvesting of the solar spectrum. Figure 3.2 depicts the

normalized absorption spectra of the large band gap polymers MDMO-PPV, P3HT,

and PFDTBT and the small band gap polymer PTBEHT(chemical structures are

shown in Figure 2.1) that are used in this study. Obviously, two large band-gap

polymers are complementary with PTBEHT and could, in principle, be used in a

tandem cell.

52


Figure 3.2. Absorption spectra of the large band gap polymers PFDTBT, P3HT, and MDMO-PPV and of

the small band gap polymer PTBEHT with its maximum at 850 nm.

It should be noted that the blend of the large band gap polymer and PCBM acts as

a dielectric sandwiched between two thin semi-transparent layers of gold (bottom

and middle electrode). Light transmitted through the active layer of the bottom cell

interferes with the reflected light at the middle electrode, leading to optical

cavities.[39] As discussed in Section 3.2, a middle electrode with a layer thickness of

about 15 nm of gold is required to prevent the cathode from being destroyed by

spin coating of the water-based PEDOT:PSS. This inevitably reduces the

illumination intensity of the top cell. [38,37] Since such a semi-transparent middle

electrode creates, together with the bottom electrode, optical cavities, the optical

transmission through the first (bottom) cell needs to be optimized for the optical

absorption of the second (top) cell. The optical cavity, and therefore the

transmission, is affected by the thickness of the active layer, as demonstrated in

Figure 3.3. Given the fact that the small band gap polymer (PTBEHT) absorbs

between 700 and 950 nm, the optical cavities (or layer thickness) of the bottom cell

have to be optimized in order to transmit at this wavelength range. Therefore, we

53

processed blends of large band gap polymers (MDMO-PPV, P3HT and PFDTBT

with PCBM) in a similar sandwich as is used in the tandem cell (Figure 3.1C). The

experimentally determined optimal layer thickness of the bottom cell (reference

bottom cell in Figure 3.1A) can be derived from Figure 3.3 and amounts to 330 nm

for a 1:4 MDMO-PPV:PCBM blend (Figure 3.3A), 120 nm for a 1:1 P3HT:PCBM

blend (Figure 3.3B), and 110 nm for a 1:4 PFDTBT:PCBM blend (Figure 3.3C).

Based on this transmission a bottom cell processed from PFDTBT would result in

the most intense irradiation of the top cell (48 % transmission), since both MDMO-

PPV and P3HT have transmission intensities of 34 and 37 %, respectively, in the

wavelength range that is absorbed by the small band gap polymer PTBEHT.

Furthermore, the width of transmitted intensity for MDMO-PPV is narrow, which

also reduces the amount of photons that can be absorbed by the top cell. The third

reason to process a bottom cell from PFDTBT is the current matching, which will

be discussed in the following Section. Figure 3.4 demonstrates that a thickness of

110 nm for the bottom cell based on PFDTBT:PCBM (1:4) results in a perfect

optical coupling between the light output of the bottom cell and the absorption

spectrum of the top cell, since both have their maximum intensity at 850 nm.

54


55

Figure 3.3. Transmission of incident light through the bottom cell (or light input of the top cell). Due to

optical cavities, the spectrum of the transmitted light shifts for different thicknesses of the active layer of

the bottom cell a) for the blend of MDMO-PPV and PCBM (1:4), b) for the blend of P3HT and PCBM

(1:1) and c) for the blend of PFDTBT and PCBM (1:4). The effect of the optical cavities allows for tuning

the transmission optically.

Figure 3.4. The light output of the bottom cell consisting of glass/1 nm Cr/15 nm Au/60 nm

PEDOT:PSS/110 nm PFDTBT:PCBM (1:4)/0.5 nm LiF/0.5 nm Al/15 nm Au (circles) and the absorption

of the small band gap polymer (PTBEHT) of the top cell.

3.4 Current matching

Because the two BHJ sub cells are stacked in series, the current that can be

extracted out of the tandem cell follows the lowest of the currents generated in the

bottom and top cell. As demonstrated in Section 3.3, the top cell receives the light

that is not absorbed by the bottom cell and is illuminated under lower light

intensities. Consequently, the top cell produces lower photocurrents. This means

that in a tandem structure the extracted current is almost the same as the

photocurrent of the cell that generates the lowest current. When the bottom cell

generates much more current (e.g., at short-circuit conditions), the excess

electrons cannot recombine with the holes from the top cell and will charge the

middle electrode. This charging will partially compensate the build-in voltage

across the bottom cell until the current of the bottom cell matches the current of the

top cell. At the same time, the charged middle contact leads to reverse-biasing of

the top cell, which leads to a small improvement in performance of the top cell in a

56


tandem structure. The device operation of unbalanced tandem cells will be

explained in more detail in Chapter 5. In the structure mentioned here, the

matching of the photocurrent of the sub cells is presented as follows: There are

many large band gap polymers (MDMO-PPV, P3HT, PFDTBT) suitable for the

bottom cell. While their absorption spectra are almost the same, the charge carrier

mobilities are different. Also treatments such as annealing at different temperatures

in case of P3HT leads to different mobilities. [81] Using materials with higher

mobilities causes a higher generated photocurrent in the device. In this way, using

proper material and processing methods, the photocurrent of the bottom cell can

be tuned. To determine the photo current that can be generated in the top cell, a

reference top cell (Figure 3.1B) was fabricated that mimics the top cell in the

tandem geometry. This reference top cell was illuminated under various light

intensities (100–1000 W/m2, AM1.5), using neutral density filters, and the J–V

characteristics are measured and from this the JSC was extracted [38,37] (Figure 3.5).

The maximum light intensity that reaches the top cell is estimated from Figure 3.3

and amounts to 34, 37 and 48 % for blends based on MDMO-PPV, P3HT and

PFDTBT, respectively. Consequently, a top cell that is processed on a bottom cell

based on MDMO-PPV:PCBM (1:4) or on P3HT:PCBM (1:1) will generate a short-

circuit current of ~ 6 A/m2, whereas a top cell fabricated on a blend of

PFDTBT:PCBM (1:4) will generate a JSC ~ 10 A/m2 at 48 % light intensity (or 480

W/m2) . Since the top cell is already deprived of photons due to the absorption and

optical cavity of the bottom cell, processing with PFDTBT:PCBM as the bottom cell

is advantageous for the performance of the tandem cell due to the higher

transmission intensity at the relevant wavelengths (Section 3.3) as well as the

higher VOC of the bottom cell.

57

Figure 3.5. The short-circuit current density of the reference top cell under illumination with different

light intensities (AM1.5 solar spectrum).

As described above, the current densities of the two cells, bottom and top, have to

match quantitatively. Therefore, the J–V characteristics of all the reference bottom

cells from Figure 3.3 were measured and compared to the characteristics of the top

cell under various illumination intensities. The important parameters of the bottom

cell are the transmitted light intensity, JSC, FF, and VOC. In Figure 3.6 the current

density versus voltage is plotted for the bottom cells with the correct optical cavities

(layer thickness), as determined from Figure 3.3, which are: 110 nm

PFDTBT:PCBM (1:4), 120 nm P3HT:PCBM (1:1), and 330 nm MDMO-PPV:PCBM

(1:4). Figure 3.6 clearly demonstrates that the bottom cell based on

PFDTBT:PCBM (1:4) generates a JSC of ~ 10 A/m2, which is required to match the

top cell. Bottom cells based on P3HT: PCBM (1:1) and MDMO-PPV:PCBM (1:4)

with an active layer optimized for the optical matching (Section 3.3) generate JSC

that is higher and does not match the photo current generated in the top cell

(Figure 3.5). Therefore, the combination of PFDTBT/PCBM (1:4) for the bottom cell

(large band gap) and PTBEHT/PCBM (1:4) for the top cell (small band gap), as

shown in Figure 3.4, results in an optimized optical (Section 3.3) and electronic

coupling (this Section) for the tandem cell in series.

58


Figure 3.6. The current density (J)–voltage (V) measurements in dark and under illumination with 1000

W/m2, AM 1.5 solar spectrum. The photocurrent of the bottom cell using P3HT:PCBM(1:1) and MDMO-

PPV:PCBM (1:4) is much higher than the photocurrent of the bottom cell using PFDTBT:PCBM (1:4) for

the active layer.

3.5 Optimum tandem photovoltaic device

In the tandem geometry, the bottom and top BHJ solar cells are stacked in series,

which implies that, for a well performing tandem cell, the open-circuit voltage (VOC)

of the tandem cell is equal to the sum of the VOC of both individual cells. The short-

circuit current (JSC) of the tandem cell is limited by the lowest short-circuit current

(JSC) of the two individual cells. For the maximum performance of the tandem cell

with series configuration, the JSC of each sub cell has to be matched (Section 3.4).

In the tandem cell geometry, the bottom and top cell can be measured individually

by contacting the bottom (anode) and middle (cathode) electrode for the bottom

cell, and the middle (anode) and top (cathode) electrode for the top cell. As

explained above, the combination of 110 nm PFDTBT/PCBM (1:4) for the bottom

cell (large band gap) and PTBEHT/PCBM (1:4) for the top cell (small band gap),

results in an optimized optical (Section 3.3) and electronic coupling (Section 3.4)

for the tandem cell in series. The thickness of the active layer (PTBEHT:PCBM,

1:4) of the top cell is optimized for its performance and amounts to 90 nm. The

59

structure of the tandem and reference top cell is shown in Figure 3.1C and Figure

3.1B, respectively. The current density–voltage (J–V) measurements of the

individual top and bottom cell, and of the tandem cell are depicted in Figure 3.7.

The values of the VOC, JSC, and the fill factor (FF) extracted from the measurements

are summarized in Table 3.1.

Figure 3.7. The current density–voltage characteristics measured separately of the bottom, top and

tandem cell in dark and under illumination (AM1.5, 1000 W/m2).

Cells VOC

[V]

JSC

[A/m2]

FF

[%]

η

[%]

Bottom 0.9 ~10 50 0.35

Top 0.5 ~9 64 0.23

Tandem 1.4 ~9 55 0.57

Table 3.1. The quantitative values extracted from the J–V measurements of the individual cells.

Illumination conditions: 1000 W/m2 , AM1.5 simulated solar spectrum.

60


A high open-circuit voltage (VOC) of 1.4 Volt is achieved which amounts to the sum

of the open-circuit voltages of the bottom cell (0.9 Volt) and the top cell (0.5 Volt).

Figure 3.7 and Table 3.1 demonstrate that the tandem structure improves the

performance of the individual sub cells (bottom and top cell), since the efficiency of

the tandem cell is 1.6 times higher than the bottom cell and 2.5 times higher than

the top cell.

The top cell generates a lower current than the bottom cell and limits the

performance of the tandem cell in series configuration. To demonstrate that the

bottom and top cell are optically coupled in a proper manner, the reference top cell,

that is processed directly on glass, was illuminated under 400 and 600 W/m2 by

using neutral density filters. Those two intensities were close to the 500 W/m2

already predicted by measuring the transmission through the bottom cell (Figure

3.3). The J–V characteristics of the reference top cell illuminated under 400 and

600 W/m2 were compared with the illuminated top cell (top cell in Figure 3.7) in a

tandem geometry, and plotted in Figure 3.8. From the short-circuit currents in

Figure 3.8, the light intensity that is transmitted through the bottom cell and that

illuminates the top cell is estimated to be about 500 W/m2, which is in perfect

agreement with the measurements in Figure 3.3 and 3.5.

61

Figure 3.8. The J–V characteristics of the PTBEHT/PCBM (1:4) cell in dark and under illumination with

intensities of 400 (squares) and 600 (circles) W/m2 (AM1.5 solar spectrum), and the J–V characteristics

of only the top cell (top cell in Figure 7) as measured in the tandem configuration (triangles) in the dark

and under 1000 W/m2, AM1.5 solar spectrum.

3.7 Fabrication aspects of the device

The bottom cells were fabricated by using blends in a 1:4 ratio of poly((2,7-(9,9-

dioctyl)-fluorene)-alt-5,5-(4',7'-di-2-thienyl-2',1',3'-benzothiadiazole) PFDTBT (see

inset Figure 3.3C) and the fullerene derivative [6,6]-Phenyl-C61-butyric acid methyl

ester (PCBM) or poly(2-methoxy-5-(3',7'-dimethyloctyloxy)-1,4-phenylene

vinylene) (MDMO-PPV, see inset Fig 3.3A) and PCBM. Also bottom cells based

on blends in a 1:1 ratio of regioregular poly(3-hexylthiophene) (P3HT, see Figure

3.3C)) and PCBM were fabricated. The top cell was processed from a CHCl3

solution of the poly(di-2-thienylthienopyrazine) derivative PTBEHT (see inset

Figure 3.2) and PCBM in a 1:4 ratio. Poly5,7-di-2-thienyl-2,3-bis(3,5-di(2-

ethylhexyloxy)phenyl)-thieno[3,4-b]pyrazine (PTBEHT) was synthesized from the

corresponding dibrominated monomer via a condensation polymerization using

bis(1,5-cyclooctadiene)nickel(0) [Ni(COD)2]. The PTBEHT was extensively purified

by removing Ni with EDTA disodium salt, Soxhlet extraction using different

solvents, and finally a BioBeads GPC column to remove the low molecular weight

fraction. The molecular weights as determined with size exclusion chromatography

62


are Mn = 52,000 g/mol and Mw = 160,000 g/mol (PDI =3.1). PTBEHT is soluble in

most common organic solvents. The devices were fabricated on cleaned glass

substrates. For the bottom contact, by using a shadow mask, 1 nm Cr (as adhesion

layer) and 15 nm Au were vapor deposited (at 10-7 mbar) on the substrate. Both Cr

and Au layers were deposited with very low evaporation-rate ( ~ 0.02 nm/s) in

order to increase the flatness of the electrode (Anode). A 60 nm thick poly(3,4-

ethylene dioxythiophene) : polystryrenesulfonic acid (PEDOT:PSS, Bayer AG) was

spin coated and dried in the oven at 140 °C for 15 minutes. All conjugated

polymers and PCBM were dissolved in chloroform and mixed in the designated

weight ratio. A proper spin program was chosen to tune the thickness of the layers.

For the middle electrode, using a shadow mask, 0.5 nm LiF + 0.5 nm Al and 15 nm

Au were vapor deposited (at 10-7 mbar) without breaking the vacuum. The

combination of LiF and Al used in the middle contact were deposited with very low

rate ( ~ 0.01 nm/s) to reach a very smooth interlayer. The gold (Au) used in the

middle contact was deposited with a very high evaporation rate ( ~ 2.2 nm/s) in

order to prevent penetration of the gold atoms into the polymer surface. This is

important since experience shows that the integrity of the layers is necessary to

achieve a well-performing tandem device. The high evaporation rate (or cooling the

substrate during the evaporation processes) leads to creation of bigger gold

islands at the middle contact and a sharper interface with the active layer of the

bottom cell. Therefore, such a layer better protects the bottom cell against

dissolving, during the processing of the top cell. A 60 nm thick PEDOT:PSS was

spin coated onto the middle electrode. This PEDOT layer was dried in a vacuum

chamber at 10-2 mbar (1 Pa) for 30 minutes. For the active layer of the top cell, the

polymer PTBEHT (as electron donor) and PCBM (as electron acceptor) were used

in a 1:4 ratio. Both were dissolved in chloroform. A proper spin program was

chosen to tune the thickness of the layers. The reference top cell was fabricated on

a glass substrate that consisted of the same electrodes and active layer as the top

cell in the tandem configuration. For the top contact, using a shadow mask, 1 nm

LiF and 100 nm aluminum (Al) were thermally evaporated. The thicknesses of the

different layers were measured by a Dektak 6M profilometer. All optical

measurements were performed on a Perkin–Elmer Lamda 900 Spectrometer. For

63

optical measurements under different light intensities, neutral density filters were

used. The J–V measurements (in dark and illuminated) were done by using a

Keithley 2400 source meter. The illumination was done using a AM1.5 simulated

solar spectrum from a Steuernagel SolarConstant 1200 light source with an

intensity of 1000 W/m2 (unless noted otherwise). The processing was done in a

glove box under nitrogen and at room temperature.

3.5 Summary

A solution-processed organic tandem cell has been fabricated, using the same

solvent (chloroform) for both active layers. Dissolving of the first layer was

prevented by using a composite middle electrode that is semitransparent. The layer

thickness of the bottom cell was optimized in order to create an optical cavity that

efficiently transmits the required wavelength for the top cell. The top cell is

processed from a small band gap polymer that allows the collection of a broader

range of the solar spectrum. The bottom and top cell are complementary with

respect to their absorption spectra. Since the bottom and top cell are electronically

stacked in series, the open-circuit voltages of the tandem cell equals the sum of

the open-circuits voltages of both sub cells, and the short-circuit current is limited

by the lowest value. In this way an open-circuit voltage of 1.4 V is achieved. For

highly efficient tandem cells, new materials with higher mobilities and proper optical

absorption spectra are needed. The high optical absorption at longer wavelengths

of the metallic electrodes (Au) limited the light intensity at the top cell.

64


65

Chapter 4

Polymer Tandem Solar Cells with Optical Spacer as Interlayerc Abstract

In this Chapter a solution-processed polymer tandem solar cell is described in

which the two photoactive single cells are separated by an optical spacer. The use

of an optical spacer allows for an independent optimization of both the electronic

and optical properties of the tandem cell. The optical transmission window of the

bottom cell is optimized to match the optical absorption of the top cell by varying

the layer thickness of the optical spacer. The two bulk heterojunction sub cells

have complementary absorption maximal at λmax ~ 850 nm for the top cell and λmax

~ 550 nm for the bottom cell. The sub cells are electronically coupled in series or in

parallel using four electrical contacts. The series configuration leads to an open-

circuit voltage of >1 V, which is equal to the sum of both sub cells. The parallel

configuration leads to a high short-circuit current of 92 A/m2, which is equal to the

sum of both sub cells. The parallel configuration results in a higher efficiency

compared to the series configuration.

c The main results of this Chapter have been published as: A. Hadipour, B. de Boer, and P. W. M. Blom, J. Appl. Phys. 2007, 102, 074506.

66


4.1 Introduction

In Chapter 3 a solution-processed organic tandem solar cell was reported that was

fabricated from conjugated polymers with complementary absorption spectra and

separated by a conducting composite middle contact. In this tandem cell, high

energy photons are absorbed in the bottom cell and low energy photons are

absorbed in the top cell. In such a stacked geometry, the middle electrode serves

two different purposes; as a charge recombination centre, and as a protecting layer

for the bottom cell during spin coating of the top cell. The sub cells are

electronically coupled together in series, which results in an open-circuit voltage

(VOC) of the tandem cell that equals the sum of the VOC of each sub cell. The layer

thickness of the bottom cell had to be optimized to in such a way that the optical

out coupling is adapted to the absorption of the top cell. A disadvantage of this

approach is that the optimum thickness of the bottom cell, required to match the

optical output, is not necessarily equal to the thickness where the bottom cell

reaches its optimum performance. If, for example, a thickness of 300 nm is

required for optical out coupling the occurrence of unbalanced transport may lead

to the formation of space-charges, giving rise to a reduced fill factor and

performance.[69,38,37] In order to improve the first generation solution-processed

tandem cells as described in Chapter 3, we introduce here an additional solution-

processable, transparent and insulating layer which serves as an optical spacer

and leads to the fabrication of a 4-contact tandem cell. In this second-generation

tandem cell, the thickness of the bottom cell is optimized for its electrical

performance and the optical out coupling is tuned by varying the thickness of the

optical spacer on top of the bottom cell. The transmitted light through the complete

stack (bottom solar cell with optical spacer) is matched with the absorption

spectrum of the top cell. Since the optical spacer is an insulator, the fabrication of

tandem cells with four electrodes is feasible and, consequently, the two sub cells

can be coupled electronically (external) in parallel or in series.

67

Figure 4.1. A) Schematic structure of the previously mentioned tandem cell in Chapter 3 in which the

bottom cell acts as an optical spacer. B) Schematic structure of the second-generation tandem cell in

which an additional optical spacer is used that allows for independent tuning of the optical cavity. Two

bulk heterojunction cells (bottom and top cell) are stacked in series or parallel outside the device (four

terminal devices). The absorption spectra of the two semiconducting polymers P3HT and PTBEHT are

complementary.

The optical out coupling and electronic performance of the tandem cells are

addressed in the following Sections. The materials and thickness of the electrodes

are optimized to have a good optical transmission and a low resistance.

Furthermore, the layer thicknesses of both sub cells are optimized for their

electronic transport. The bottom cell with the optical spacer processed on top is

optimized for its optical transmission in order to harvest the maximum amount of

photons on the (infra)red edge of the visible spectrum. Finally, all important

parameters are used to design and optimize the 4-electrode tandem solar cell with

an imbedded optical spacer.

4.2 Electrodes of the device

The two organic solar cells are linked together by an insulating, solution-

processable and transparent layer of poly(trifluoroethylene) (PTrFE) dissolved in

methyl ethyl ketone (MEK). This spin coated layer acts as an optical spacer. Due to

the presence of the insulating layer of PTrFE the fabrication of the tandem cell now

requires four electrodes. The electrodes of the bottom and the top cell can be

68


connected externally in series or parallel. Consequently, the contacts have to be

highly conductive in order to extract and transport the charges from each sub cell.

Two of the four contacts used here in the tandem configuration are known to be

efficient in BHJ solar cells, namely, 140 nm ITO/50 nm PEDOT:PSS as anode for

the bottom cell and 1 nm LiF/80 nm Al as cathode of the top cell. Now, the semi-

transparent cathode of the bottom cell and the semi-transparent anode of the top

cell have to be defined. The requirements for these electrodes are:

a) as thin as possible to obtain maximum transparency,

b) conductive enough to extract charge carriers,

c) stable during spin coating of the PTrFE optical spacer.

For the cathode of the bottom cell 3 nm of Samarium (Sm) topped with 12 nm Au

was found to be very transparent, highly conductive and stable during spin coating

of the PTrFE layer. The low work function of Sm (~2.7 eV [82]) provides an Ohmic

contact with the bottom cell in order to maximize the open-circuit voltage. The 12

nm Au layer has a sufficiently low resistance for transporting the charge carriers

towards the external electrodes. Figure 4.2 demonstrates the difference between

two BHJ single cells using ITO/PEDOT:PSS/250 nm P3HT:PCBM (1:1) with

different cathodes. Both cells shows the same FF (~64%) and Voc (~0.6 Volt) but

the cell with 3 nm Sm/12 nm Au cathode yields a lower photocurrent since this

electrode is semi-transparent, resulting in a reduction of the amount of absorbed

light. On the other hand, the reflection of a closed layer of 80 nm Al (perfect mirror)

is ~ 100 %, which results in a higher photocurrent for a solar cell topped with 80 nm

Al.

69

Figure 4.2. Comparison between a semi-transparent cathode made of 3 nm Sm/12 nm Au and a

‘perfect’ mirror of 1 nm LiF/80 nm Al on top of the bottom cell consisting of a BHJ of 250 nm

P3HT:PCBM (1:1). The fill factor and open-circuit voltage of both cells are the same but the photo-

generated current of the cell with a semi-transparent cathode is significantly lower.

On top of the PTrFE optical spacer a bilayer of 20 nm Au and 50 nm PEDOT:PSS

were chosen for the anode of the top cell. The high and stable work function (~5.2

eV) of PEDOT:PSS forms an Ohmic contact for the top cell to extract the holes. In

addition, PEDOT:PSS improves the wetting of the anode for the processing of the

top cell. The 20 nm Au layer serves again as conducting layer to extract the holes.

This Au layer needed to be optimized for its sheets resistance and optical

transparency. To determine its optimal thickness (sheets resistivity versus optical

transparency) a reference top cell was fabricated on top of a 250 nm thick layer of

PTrFE that was spin coated onto glass substrate. The reference top cell consists of

x nm Au/50 nm PEDOT:PSS/100 nm MDMO-PPV:PCBM (1:4)/1 nm LiF/80 nm Al

as shown in Figure 4.3.

70


Figure 4.3. The structure of the reference top cell. By varying the thickness of gold (X nm) used in the

anode of the device, the thinnest, well-conducting layer can be found.

As Figure 4.4 demonstrates, at least 20 nm Au is needed to create a well-

performing device. A thinner layer of Au (15 nm) leads to a very high sheet

resistance of the anode and, therefore, to a very poor performance of the cell. This

is attributed to the surface roughness of the spin coated optical spacer of PTrFE,

leading to the formation of a semi-continuous Au layer.

71

Figure 4.4. Current density–voltage characteristics of a reference top cell structure with different gold

thickness at the anode on 250 nm PTrFE. On top of the gold layer a cell is processed based on 50 nm

PEDOT:PSS/100 nm MDMO-PPV:PCBM (1:4)/1 nm LiF/80 nm Al. At a thickness of 15 nm the sheet

resistance of the contact is too high which leads to a very poor performance of the cell.

The J–V performances of two reference top cells are compared in Figure 4.5 with

20 nm and 30 nm thick gold (Au) layers used in the anode of the device. As Figure

4.5 shows, increasing the thickness of the gold layer does not improve the

electrical performance of the device.

Figure 4.5. Two reference top cells with different anodes are compared. Increasing the thickness of the

gold layer used in the anode has little effect on the performance of the device.

72


Therefore the combination of 20 nm Au + 50 nm PEDOT:PSS was used for the

fabrication of the anode of the top cell as most transparent and well-conducting

layer.

4.3 Optimum sub cells

Now that we have optimized and defined all anodes and cathodes of the four

terminal tandem solar cell, we can further optimize the layer thicknesses of the

photoactive layers of the tandem solar cell. These layers can be independently

optimized since the optical out coupling of the bottom cell can be tuned by the layer

thickness of the PTrFE optical spacer (Figure 4.1). Therefore, blends of

P3HT:PCBM (1:1) were spin coated with various layer thicknesses onto the

substrates covered with 140 nm ITO/50 nm PEDOT:PSS. The device was topped

off with 3 nm Sm/80 nm Al. The structure used for those measurements is shown in

Figure 4.6.

Figure 4.6. Standard device structure for optimizing the bottom cell. By varying the thickness of the

active layer, the bottom cell can be optimized for its best electrical performance. Using chloroform as

solvent, a 250 nm thick film has the best performance.

The data in Figure 4.7 demonstrate that a cell with a layer thickness of 250 nm of

P3HT:PCBM (1:1) performs best. When the active layer is thinner (120 nm, defined

by the optical matching of the device in Chapter 3), the amount of absorbed

photons is much lower and, consequently, a lower photocurrent is obtained. When

73

the layer thickness is further increased to 450 nm, the amount of absorbed photons

is only slightly increased, but the unbalanced transport of the charge carriers

becomes a limitation; The build up of space-charge results in a significant

decrease of the fill factor, which in turn leads to a lower total performance.[69]

Figure 4.7. Current–voltage characteristics of single BHJ bottom cells with various thicknesses of

P3HT:PCBM (1:1). A layer thickness of 250 nm shows the best performance.

By using a standard structure for a single bulk heterojunction solar cell (Figure 4.6),

the performance of the top cell can be compared with the performance of the

bottom cell. For the top cell we used the conjugated polymer poly5,7-di-2-thienyl-

2,3-bis(3,5-di(2-ethylhexyloxy)phenyl)-thieno[3,4-b]pyrazine (PTBEHT). The

thickness of the top cell made of PTBEHT:PCBM (1:4) was kept to 90 nm, which

was found to give the best performance (the chemical structure of PTBEHT is

given in Figure 2.1). The J–V characteristics of the 250 nm P3HT:PCBM (1:1) and

90 nm PTBEHT:PCBM (1:1) solar cells are plotted in dark and under illumination

(1000 W/m2 , AM1.5) in Figure 4.8.

74


Figure 4.8. The J–V characteristic of the active layers used in the bottom (250 nm P3HT:PCBM (1:1))

and top (90 nm PTBEHT:PCBM(1:4)) cells. These reference single cells are measured in dark and

under illumination with 1000 W/m2, AM1.5 solar spectrum.

4.4 Optical considerations

The transparent layer of PTrFE is sandwiched between the semi-transparent

cathode of the bottom cell and the semi-transparent anode of the top cell. This

allows for the light transmitted through the bottom cell to interfere with the reflected

light at the anode of the top cell. Therefore, the thickness of the PTrFE layer affects

the total light output through the bottom stack before it reaches the top cell. After

optimization of the layer thickness of the bottom cell, the wavelength of the

transmitted light output can now be tuned to match the absorption spectrum of the

top cell by varying the thickness of the optical spacer (PTrFE layer). The

combination of Sm and Au as cathode for the bottom cell is inert for processing of

the PTrFE optical spacer, which is spin coated from methyl ethyl ketone (MEK).

The advantage of choosing PTrFE is that it can be dissolved in a polar solvent like

MEK very fast, but very slowly in solvents like chloroform, chlorobenzene or

dichlorobenzene. As a result the solvents used for processing of the active layers

of the bottom and the top cell (chloroform or chlorobenzene) are orthogonally

compatible with the polar solvent (MEK) used for processing of the optical spacer.

75

The layer thickness of this transparent PTrFE layer can be easily varied between

several tens of nanometers to micrometers. By using donor materials with non-

overlapping absorption spectra for the sub cells in a tandem configuration, the

utilization of two absorbing polymers with complementary absorption spectrum

results in an enhanced photon harvesting of the solar spectrum. Figure 4.9 depicts

the normalized absorption spectra of the large band gap polymer P3HT and the

small band gap polymer PTBEHT (chemical structure PTBEHT and P3HT inset

Figure 4.9) that are used in this study.

Figure 4.9. Absorption spectra of the large band gap polymer P3HT and of the small band gap polymer

PTBEHT with its maximum at 850 nm are shown. The absorption spectra are complementary.

We have performed optical measurements for various layer thicknesses of PTrFE

to evaluate the optical transmission of the combined layers. Both the

ITO/PEDOT:PSS/P3HT:PCBM/3 nm Sm/12 nm Au stack and the PTrFE

sandwiched between 3 nm Sm/12 nm Au and 20 nm Au act as dielectric layers

between semi-transparent electrodes. The structure used for the optical

measurements is demonstrated in Figure 4.10.

76


Figure 4.10. Device structure used for the optical matching of the tandem device. The transmitted light

through the device has to have the maximum at about 850 nm wavelength where the top cell absorption

is maximum..

As demonstrated in Figure 4.11(circles), the transmitted light through the bottom

cell consisting of ITO/PEDOT:PSS/250 nm P3HT:PCBM/3 nm Sm/12 nm Au is not

yet matched with the absorption spectrum of the top cell. Given the fact that the

small band gap polymer (PTBEHT) absorbs between 700 and 950 nm, the optical

cavity (or layer thickness) of the optical spacer has to be optimized in order to

transmit in this wavelength range. Figure 4.11 (triangles) demonstrates that a layer

thickness of the optical spacer of 250 nm, in combination with 250 nm

P3HT:PCBM/3 nm Sm/12 nm Au, results in an optical out coupling of the bottom

stack that matches the absorption of the small band gap polymer in the top cell.

The bottom cell processed from 250 nm P3HT:PCBM (1:1) topped with an optical

77

spacer (PTrFE) of 250 nm transmits about 50 % in the wavelength range that is

absorbed by the small band gap polymer PTBEHT (Figure 4.11 dotted line).

Figure 4.11. Transmission of incident light through the bottom cell consisting of glass/140 nm ITO/50

nm PEDOT:PSS/250 nm P3HT:PCBM (1:1)/3 nm Sm/12 nm Au/x nm PTrFE/20 nm Au/PEDOT:PSS.

For a blend of P3HT:PCBM (1:1) with a layer thickness of 250 nm (and no optical spacer), the light

output is not yet matched with the absorption of the top cell (700–950 nm, spectrum given by the dotted

line). By spin coating an additional transparent layer of PTrFE, the transmitted light can be tuned in

order to match the absorption of the top cell.

4.5 Electrical considerations

In this second-generation tandem geometry, the bottom and the top cells can be

electronically coupled either in series or in parallel. When the bottom and top cells

are connected in series, the open-circuit voltage (VOC) of the tandem cell is equal to

the sum of the open-circuit voltages of both individual cells. The photocurrent

density of the tandem cell becomes limited by the lowest photocurrent density of

the two sub cells (the top cell in Chapter 3). As shown from the single reference

cells (Figure 4.8) the cells based on PTBEHT generate less photocurrent as

compared to the P3HT based cells. Combined with the fact that the top cell is also

illuminated with lower light intensity (~50%), the photocurrent of the top cell is

substantially lower as compared to the bottom cell. The J–V measurements of the

individual bottom and top cells, and of the tandem cell connected in series are

depicted in Figure 4.12A. As expected, the top cell strongly limits the photocurrent

of the tandem cell. In this four terminal tandem geometry, the two sub cells can

78


also be easily connected in parallel. Since both cells produce the same VOC the

tandem solar cell connected parallel is expected to have the same VOC as both sub

cells (no losses). In contrast, the current that can be extracted from the parallel

tandem cell is now the sum of the photocurrent of the bottom and top cell. This

implies that the lower current of the top cell does not limit the tandem cell in parallel

configuration. The current density–voltage characteristics for the bottom, top and

the parallel tandem cell are plotted in Figure 4.12B. Since the top cell generates a

much lower photocurrent as compared to the bottom cell the parallel tandem solar

cell leads to a higher performance as compared to the tandem solar cell connected

in series. Indeed, the open-circuit voltage of the parallel tandem cell (0.59 Volt) is

close to the VOC of the bottom (0.60 Volt) and top cells (0.51 Volt), while the current

is the sum of both photocurrents generated by the two sub cells. A direct

comparison between the series and parallel configuration is shown in Figure

4.12C; the series configuration has a high open circuit voltage (VOC = 1.03 V), while

its short circuit current (JSC = 16.3 A/m2) is limited by the lower current of the top

cell. The parallel configuration shows the same open circuit voltage as both sub

cells (VOC = 0.59 V), combined with a high short circuit current (JSC = 92.0 A/m2).

The values of the VOC, JSC, FF and efficiencies of the bottom, top, series tandem

and parallel tandem solar cells are summarized in Table 4.1. The performance of

the parallel tandem cell is much better than the series tandem cell where the

photocurrent is limited by the low photocurrent of the top cell. However, it should

be noted that in the present parallel configuration the performance hardly exceeds

the performance of the bottom cell alone (estimated efficiency is 3% for both

bottom and parallel tandem cell). Since the current of this tandem cell is equal to

the sum of the currents generated by both sub cells, the lower fill factor (FF) of the

top cell (50%) (Compared to the high FF of the bottom cell (64%)) also affects the

fill factor of the tandem cell (54%). This lower fill factor compensates the gain in

photocurrent in the parallel tandem solar cell. In the case of two sub cells having

similar fill factors, the parallel configuration will always results in a tandem solar cell

with higher performance as compared to the individual sub cells.

79

Figure 4.12. A) J–V characteristics of the bottom and top cell measured separately, and the tandem

cells connected in series in dark and under illumination with 1000 W/m2, AM1.5. The current density is

limited by the lowest current density of the two cells and the open-circuit voltage is the sum of both sub

80


cells. B) J–V characteristics of the bottom and top cell measured separately, and the tandem cells

connected parallel in dark and under illumination with 1000 W/m2, AM1.5. The current density is equal

to the sum of both sub cells and the open-circuit voltage is limited to the lowest VOC of the sub cells. C)

The J–V characteristics of the series tandem cell versus the parallel tandem cell.

Solar cell type VOC (V) JSC (A/m2) FF (%) η (%)

Bottom cell 0.60 81.6 64 ~3

Top cell 0.51 12.7 50 ~0.31

Series Tandem cell 1.03 16.3 51 ~ 0.85

Parallel Tandem cell 0.59 92.0 55 ~3

Table 4.1. The quantitative values extracted from the J–V measurements of the individual and tandem

photovoltaic cells. Illumination conditions: 1000 W/m2, AM1.5.

4.6 Device fabrication

The devices were fabricated on cleaned glass/ITO substrates. A 50 nm thick

poly(3,4-ethylene dioxythiophene) : polystryrenesulfonic acid (PEDOT:PSS, H. C.

Starck) was spin coated and dried in an oven at 140 °C for 15 minutes. The bottom

cells were based on blends in a 1:1 ratio of regioregular poly(3-hexylthiophene)

(P3HT, see Figure 6)) and the fullerene derivative [6,6]-phenyl-C61-butyric acid

methyl ester (PCBM) processed from chloroform. The P3HT:PCBM film was

annealed at 135 °C for 2 hrs using a hot plate in the glove box under nitrogen

atmosphere. The top cell was processed from a CHCl3 solution of the poly(di-2-

thienylthienopyrazine) derivative PTBEHT (see inset Figure 6) and PCBM in a 1:4

ratio. Poly5,7-di-2-thienyl-2,3-bis(3,5-di(2-ethylhexyloxy)phenyl)-thieno[3,4-

b]pyrazine (PTBEHT) was synthesized from the corresponding dibrominated

monomer via a condensation polymerization using bis(1,5-cyclooctadiene)nickel(0)

[Ni(COD)2].[24] PTBEHT was extensively purified by removing Ni with EDTA

81

disodium salt, Soxhlet extraction using different solvents, and finally a BioBeads

GPC column to remove the low molecular weight fraction. The molecular weights

as determined with size exclusion chromatography are Mn = 52,000 and Mw =

160,000 (PDI =3.1). PTBEHT is soluble in most common organic solvents. The

optical spacer was fabricated from poly(trifluoroethylene) (PTrFE) dissolved in

methyl ethyl ketone (MEK). The spin coating of the PTrFE was done with high

rpm’s for increasing the flatness of the layer. All conjugated polymers and PCBM

were dissolved in chloroform and mixed in the designated weight ratio. A proper

spin program was chosen to tune the thickness of the layers (250 nm for

P3HT:PCBM and 90 nm for PTBEHT:PCBM). For the cathode of the bottom cell,

using a shadow mask, 3 nm Sm and 12 nm Au were vapor deposited (at 10-7

mbar). The deposition-rate of the Sm has to be very slow ( ~ 0.01 nm/sec) for

creation as small as possible island of the Sm-interlayer. The Au layer has to be

deposited with high rate ( ~ 0.5 nm/sec) in order to increase the size of the islands

of gold during evaporation process. This leads to lower penetration depth of the Au

atoms into the polymer surface and sharper interface and therefore higher

performance of the bottom cell. PTrFE was dissolved in MEK and spin-coated onto

the bottom cell in various layer thicknesses. On top of the PTrFE layer, using a

shadow mask, 20 nm Au was evaporated and 50 nm PEDOT:PSS was spin

coated. Here, the gold has to be evaporated with high evaporation-rate ( ~ 0.5

nm/sec) for the same reasons as before. The complete stack was dried in a

vacuum chamber at 10-2 mbar (1 Pa) for 30 minutes in order to evaporate the water

in the PEDOT:PSS. For the active layer of the top cell, the polymer PTBEHT as

electron donor and PCBM as acceptor were used in a 1:4 ratio. Both were

dissolved in chloroform. A proper spin program was chosen to tune the thickness

of the layers (90 nm PTBEHT:PCBM (1:4) film). For the top contact, using a

shadow mask, 1 nm LiF and 80 nm aluminum (Al) were thermally evaporated. The

thicknesses of the different layers were measured by a Dektak 6M profilometer. All

optical measurements were performed on a Perkin–Elmer Lamda 900

Spectrometer. The J–V measurements (in dark and illuminated) were done by

using a Keithley 2400 source meter. The illumination was done using a AM1.5

simulated solar spectrum from a Steuernagel Solar Constant 1200 light source with

an intensity of 1000 W/m2. The processing was done in a glove box under nitrogen

and at room temperature.

82


4.7 Summary

Solution-processed organic tandem solar cells were fabricated using an electrically

insulating and solution-processable optical spacer. The optical spacer is based on

the transparent poly(trifluoro ethylene), which is spin coated from a MEK solution.

This polar solvent is orthogonally compatible with the solvents of the active layers

of the solar cells and leaves the bottom cell unaffected. By varying the thickness of

the optical spacer, the transmission through the bottom cell and spacer is tuned to

match the absorption of the top cell. This allows for an independent optimization of

the thickness of the bottom cell for its best electrical performance. Furthermore, the

tandem solar cell can electrically address in series or parallel. The performance of

the series tandem solar cell is limited by the low current of the top cell, even though

the VOC is equal the sum of both sub cells and amounts to > 1 V. The parallel

tandem cell has a much higher efficiency since the VOC of the parallel tandem cell

is equal to the VOC of both sub cells and the photocurrent density is equal to the

sum of the photocurrent densities of both sub cells. The lower FF of the top cell

limits the performance of the parallel tandem cell.

83

Chapter 5

Current-Voltage Characteristics of Organic Tandem Cells Abstract

In order to understand the electrical properties of a tandem organic solar cell, we

consider in this Chapter a tandem cell that is based on two sub cells with totally

different electrical properties. In this general case, the bottom cell generates higher

current but lower voltage compared to the top cell. It is important to understand the

physics behind the tandem devices for a given performance of the sub cells, since

theoretical predictions can strongly reduce the experimental work needed to reach

the optimum device structure. The experiments on tandem organic solar cells were

already mentioned in the Chapters 3 and 4 in which the sub cells were connected

electrically in series (Chapter 3) or in parallel (Chapter 4). In this chapter we

demonstrate how the electrical characteristics of tandem cells that are either

connected in series or parallel can be predicted from the characteristics of the sub

cells. In order to compare the calculated results with experiment, the 4-electrode

tandem cell is used that was demonstrated in Chapter 4. The use of such a device

has two advantages; First, because of the presence of 4 electrodes the J-V

characteristics of the individual bottom- and top cell as well as the tandem cell can

be measured in one single device. Second, since the sub cells are electrically

separated, both the series and parallel configuration can be measured within the

same device. In this way the test conditions (STC) are exactly the same for all

cells.

84


5.1 Series configuration

Suppose that we have two arbitrary sub cells (bottom and top cell) with different

current and voltage characteristic under illumination, as shown in Figure 5.1.

Figure 5.1. The current-voltage characteristic of two arbitrary sub cells. The bottom cell generates more photocurrent while the top cell has higher generated voltage but lower photocurrent.

The question now is how the J-V characteristic of a tandem cell based on those

two sub cells will look like when they are electrically connected in series or in

parallel. As a first step we consider the series configuration and later the parallel

tandem cell is also discussed. When the two sub cells are connected in series, the

total generated photocurrent will be constant throughout the device (conservation

of charge) in steady state. Furthermore, the voltages generated by the sub cells

will add up. As a result for each point of the J-V characteristic of the tandem

device the following relations are valid,

TopBottomTandem JJJ == (5.1)

TopBottomTandem VVV += (5.2)

85

Graphically, equation 5.1 means that we can draw an arbitrary horizontal line

through Figure 5.1, indicating a chosen constant current density that flows through

the cells. This horizontal line crosses the J-V curves under illumination of the

individual bottom and the top cell at a specific voltage for each cell. Those cross-

points are the values of the voltages with which the sub cells are effectively biased

in order to generate the chosen constant current density. Equation 2 then shows

that we have to add those two voltage values in order to determine the bias voltage

of the tandem cell in series at that constant current density. To do so, we replot

Figure 5.1 between zero and -10 A/m2 in order to enlarge the vertical axis and

choose three arbitrary current densities as shown in Figure 5.2. The horizontal line

1 is the open-circuit condition for both sub cells in which the current densities in

both of them are zero (cross-points A and B). Line 2 shows the short-circuit

condition of the top cell (cross-point C) whereas the bottom sub cell is biased by a

positive voltage (cross-point D). Line 3 is the condition in which the bottom cell is

biased by a positive voltage (cross-point F), whereas the top is biased by a

negative voltage (cross-point E).

86


Figure 5.2 A close-up of the vertical axis of Figure 5.1 between 0 and 10 A/m2. The horizontal lines 1, 2

and 3 cross the curves of the bottom (A, D and F) and top cells (B, C and E), indicating a constant

current density at different values. At each line also the energy-band diagrams are given.

Following equations 5.1 and 5.2 we can say that,

At line 1:

01 ==== JJJJ TopBottomTandem [A/m2] (5.3)

27.1)7.0()57.0( =+=+=+= BATop

OCBottom

OCTandem

OC VVVVV [Volt] (5.4)

At line 2:

68.62 −==== JJJJ TopSCBottomTandem [A/m2] (5.5)

56.0)56.0()0( =+=+=+= DCTopBottomTandem VVVVV [Volt] (5.6)

At line 3:

46.73 −==== JJJJ TopBottomTandemSC [A/m2] (5.7)

and because for this current the distance from E and F to the y-axis are equal,

0)55.0()55.0( =+−=+=+= FETopBottomTandem VVVVV [Volt] (5.8)

In this way the open-circuit voltage (equation 5.4), short-circuit current (equation

5.7) and an additional arbitrary point (at short-circuit condition of the top cell) of the

series tandem cell are predicted. In Fig. 5.2 also the energy band diagrams are

schematically depicted for these three cases. We now discuss the biasing

conditions of this series tandem cell in more detail: In a series configuration the

cathode of the bottom cell is electrically connected with the anode of the top cell. In

the tandem cell studied here the bottom cell generates much more photocurrent

than the top cell (Figure 5.1) under for example short-circuit condition. This means

that there are not enough holes arriving from the top cell to recombine with the

large amount of electrons arriving from the bottom cell. As a result, in steady-state

the excess of electrons negatively charges the connected electrodes of the sub

87

cells. This charging reduces the effective voltage across the bottom cell, and thus

also the extracted current from the bottom cell. On the other hand, the additional

electrons in the middle electrode provide a stronger voltage-drop across the top

cell (the top cell is more reversed biased) and therefore a higher current flows

through the top cell. Steady-state is reached when the lowered current in the

bottom cell is equal to the enhanced current of the top cell. This situation is

schematically demonstrated in Figure 5.3.

Figure 5.3. Schematic picture of the series tandem cell. In steady-state, the middle electrodes are

negative charged with electrons since the bottom cell generates more photocurrent than the top cell.

The charging of the electrodes changes the effective voltage across both sub cells in an opposite way.

The voltage across the top cell is enhanced, whereas the voltage across the bottom cell is decreased.

At the open-circuit voltage (line 1 in Figure 5.2) both the sub cells are biased in

such a way that the effective electric field across them is close to zero (the bias

neutralizes the built-in electric field). Current matching is then achieved since both

cells do not generate any current, they only act as two voltage sources of which the

generated voltages add up. Line 2 in Figure 5.2 shows the situation where the

effective bias across the top cell is zero (C), meaning that the field across the top

cell is now equal to its built-in electric field. Due to the negative charging of the

middle electrode the effective voltage across the bottom cell is strongly reduced

(D) in order to balance the current with the top cell. For line 3 in Figure 5.2 the

electric field across the top cell is even further enhanced by the increasing amount

of charge on the middle electrode, such that it is now reverse (negative) biased. In

88


this case, the electric field across the top cell is larger than its built-in electric field

(E). Finally, line 3 in Figure 5.2 is chosen in such a way that the negative bias

across the top cell (E) is equal to the positive bias of the bottom cell (F). As a result

the total voltage across the tandem equals zero, such that line 3 in Figure 5.2

represents the short-circuit current of the tandem cell. By choosing sufficient

horizontal lines (current levels) and extracting the voltages as mentioned above,

the whole illuminated J-V curve of the series tandem cell can be constructed. In

Figure 5.4, the given J-V characteristic of the bottom, top and constructed series

tandem cell under illumination (STC) are demonstrated.

Figure 5.4. The current-voltage characteristic of two sub cells (bottom and top) and calculated series

tandem.

Qualitatively, the constructed J-V characteristic of the tandem device is in

agreement with the experimental series tandem cells that are shown in the

Chapters 3 and 4. For these cells the extracted current in the tandem cell was very

close to, but slightly larger than, the lower current of the top cell. A quantitative

comparison will be made in the next section.

89

5.2 Parallel configuration

When the two sub cells are electrically connected in parallel, in steady-state for

each point of the J-V characteristic of the tandem device the following relations are

valid,

TopBottomTandem VVV == (5.11)

TopBottomTandem JJJ += (5.12)

Graphically, equation 5.11 means that we can now draw an arbitrary vertical line

through Figure 5.1, which indicates the chosen operating voltage for the sub cells.

This vertical line crosses the J-V characteristics under illumination of the bottom

and the top cell at a specific current for each cell. Those cross-points are the

values of the current density generated by the sub cells at the chosen operating

voltage. These two values of the current densities of the bottom and top cell then

have to be added (Equation 5.12) to calculate the current of the parallel tandem

cell for the chosen operating voltage. We now enlarge the horizontal axis of Figure

5.1 and again draw three vertical lines, as shown in Figure 5.5. The vertical line 1

is the open-circuit condition for the top cell and positive current density for the

bottom cell (cross-points K and L). Line 2 shows the condition in which the sub

cells have opposite current densities. At line 2, the bottom cell generates positive

current due to dark injection (cross-point M), whereas the top cell generates a

negative photocurrent (cross-point N). Line 3 is the short-circuit condition for all

cells in which both the bottom cell (cross-point O) and the top cell (cross-point P)

generate negative photocurrents.

90


Figure 5.5. Close-up of the horizontal axis of Figure 5.1 between -0.2 and 1 Volt. By extracting the

cross points between the vertical lines and the bottom cell (K, M and O), respectively the top cell (L, N

and P) from the graph the current density of the parallel tandem device can be constructed for a series

of chosen voltages. The energy-band diagrams in the different operation points are also given.

From equations 5.11 and 5.12 we obtain that

At line 1:

69.01 ==== VVVV TopOCBottomTandem [Volt] (5.13)

5.87)0()5.87( =+=+=+= LKTopBottomTandem JJJJJ [A/m2] (5.13)

At line 2:

58.02 ==== VVVV TopBottomTandem

OC [Volt] (5.14)

since M and N have equal distance to the x-axis,

91

0)4()4( =−+=+=+= NMTopBottomTandem JJJJJ [A/m2] (5.16)

At line 3:

03 ==== VVVV TopBottomTandem [Volt] (5.17)

75.90)1.7()65.83( −=−+−=+=+= POTopSC

BottomSC

TandemSC JJJJJ [A/m2] (5.18)

With this method, the short-circuit current, open-circuit voltage and an additional

point of the tandem J-V characteristic are determined for a parallel tandem cell

based on the sub cells mentioned before. By drawing sufficient vertical lines

through the J-V curves of the sub cells and extracting the operation points the

complete J-V characteristic of the parallel tandem device can be constructed. The

constructed J-V curve of the parallel tandem cell is shown in Figure 5.6, together

with the characteristics of the individual sub cells.

Figure 5.6 Current-voltage characteristic of the sub cells and calculated parallel tandem device.

Also shown in Fig. 5.5 are the corresponding energy band diagrams for the three

lines. In the parallel configuration the two outer electrodes are connected and show

up on an equal level in these diagrams. For line 1 in Figure 5.5 the top cell is

biased such that the electric field across the cell is close to zero (L). However,

because of the lower built-in field in the bottom cell the electric field in the bottom

cell changes sign (K) under this bias. As a result the dark injection in the bottom

92


cell is switched on and electrons now flow to the PEDOT:PSS in stead of to the

LiF/Al electrode, leading to a positive current. For the voltage corresponding to line

2 in Figure 5.5 the bottom cell is still dominated by (positive) dark current, but its

current is now of equal magnitude as the (negative) photocurrent generated by the

top cell. This voltage therefore represents the open-circuit voltage of the tandem

cell and is located in between the VOC ‘s of the individual cells. Finally, line 3 in

Figure 5.5 shows the situation when there is no bias applied across the parallel

tandem. In that case both sub cells are effectively biased by their built-in electrical

fields.

Using the procedures described in this section the current-voltage curve of any

parallel- and series connected tandem solar cell can be derived from the electrical

performance of the individual sub cells. It should be noted that this method can

also be used for the prediction of the J-V curves of multi-junction organic solar cells

with three or more active layers.

5.3 Comparison with experiment

In order to compare the calculated J-V characteristics of the tandem cells with

experimental data we can use the measurements on the tandem device as shown

in Figure 4.1B in Chapter 4. The tandem cell we now consider is based on a 250

nm P3HT:PCBM blend for the bottom cell and a 80 nm MDMO-PPV:PCBM blend

for the top cell. An optical spacer with a thickness of 190 nm was used to separate

the sub cells. For this thickness the optical spacer maximizes the transmitted light

for the wavelengths that correspond to the absorption spectrum of the MDMO-

PPV. The experimental J-V characteristics of the individual bottom- and top cell of

this structure were already shown in Figure 5.1. The complete structure of this

tandem test device is given in Figure 5.7. The two sub cells can be connected

electrically in series or in parallel using external wiring.

93

Figure 5.7. Structure of the tandem test device. The electrical performance of this device is given in

Figure 5.1. The electrodes of the device can be electrically connected in parallel or in series by

adjusting the external wires.

First, the electrodes of the device under illumination were connected in series. In

Figure 5.8 both the experimental and calculated J-V characteristic of the series

tandem cell are shown under illumination (STC). It is clear that the experiment is in

very good agreement with the characteristic that is extracted from the two

individual sub-cells, as described in the previous section.

94


Figure 5.8. The current-voltage measurements of the experimental series tandem device under

illumination (STC) and the calculated series tandem cell.

Subsequently, the electrodes of the sub cells of the above-mentioned 4-electrode

tandem device in Figure 5.7 were connected in a parallel configuration. The J-V

characteristics under illumination (STC) of both the calculated- and experimental

parallel tandem cell are shown in Figure 5.9.

95

Figure 5.9. J-V characteristic of the calculated and experimental photocurrent of a parallel tandem cell.

Again, a very good agreement between the calculated and experimental J-V

characteristics is obtained. In order to make a more quantitative comparison we

determine for the calculated- and experimental J-V curves of Figure 5. 8 (series

connection) and Figure 5.9 (parallel connection) the exact values for VOC and JSC,

the voltage Vmax and the current density Jmax at the maximum power point, and the

corresponding fill factor (FF) and efficiency η. For the series connection, the FF

and η are given by

%55)27.1()34.7()91.0()67.5(maxmax =

×−×−

==××

≡BA

VJVJ

FFOCSC

(5.19)

%5.0≅×

≡light

OCSC

PVJ

FFη . (5.20)

For the parallel configuration we find that,

%48)58.0()93.90()39.0()31.66(maxmax =

×−×−

=××

≡OCSC VJ

VJFF (5.20)

%53.2≅×

≡light

OCSC

PVJ

FFη (5.21)

The results are summarized in Table 5.1

96


Cell Jmax

[A/m2]

Vmax

[Volt]

JSC

[A/m2]

VOC

[Volt]

FF

[%]

η

[%]

Calculated

series -5.64 0.91 -7.34 1.27 55 0.50

Experimental

series -6 0.89 -7.60 1.28 54 0.48

Calculated

parallel -66.31 0.39 -91.15 0.58 48 2.53

Experimental

parallel -64.80 0.40 -89.94 0.58 49 2.56

Table 5.1. Comparison between calculated and experimental parameters of the series and parallel

tandem solar cell shown in Figure 5.8 and 5.9.

It is clear from table 5.1 that all relevant solar cell parameters for the series- and

parallel connected tandem cells can be accurately predicted from the electrical

characteristics of the individual sub cells. Verification of the predicted

characteristics with experimental data now allows us to further investigate the

effects other parameters, as the fill factor of the bottom- and top cell, on the fill

factor and performance of series- and parallel connected tandem cells.

5.4 The fill factor and efficiency of tandem solar cells

Another important question is how the fill factor of a tandem cell is affected when

one of the sub cells has a very poor FF. Will it be closer to the highest or the lowest

FF when connected in series or parallel? In order to investigate this we consider a

range of J-V characteristics as shown in Figure 5.10. These artificial J-V curves are

97

constructed in such a way that they all have the same VOC and JSC, but a large

variation in FF (from 25 to 66%). Each (artificial) solar cell has a different maximum

power point (MMP), which arises from the different maximum current (Jmax) and

maximum voltage (Vmax) for each cell, as demonstrated in Figure 5.10.

Figure 5.10. Current-voltage characteristic of 6 artificial solar cells with different fill factor. The area’s

B/A = 0.66, C/A = 0.54, D/A = 0.44, E/A = 0.39, F/A = 0.31 and G/A = 0.25 demonstrates the fill factor of

the cell 1 until the cell 6, respectively.

The relevant solar cell parameters as extracted from Figure 5.10 are given in Table

5.2.

Cell

number

VOC

[Volt]

JSC

[A/m2]

FF

[%]

η

[%]

1 0.59 95.3 66 3.71

2 0.59 95.3 55 3.09

3 0.59 95.3 46 2.58

4 0.59 95.3 39 2.19

5 0.59 95.3 34 1.91

6 0.59 95.3 25 1.40

98


Table 5.2. Electrical performance of the 6 artificial solar cells as extracted from Figure 5.10.

With these 6 cells as input we can now construct on paper a series of tandem cells:

we choose the characteristic of cell 1 (FF=66%) as bottom cell and then add all

curves 1 to 6 subsequently as top cell. For each combination of 2 curves we then

apply the method as explained in section 5.2, and construct the resulting electrical

tandem characteristics when the cells are connected either in series or parallel. In

this way, we can investigate the effect of a variation of the FF in one of the sub

cells on the FF and performance of the tandem cells. For the series connection the

resulting J-V characteristics of the various tandem cells are shown in Figure 5.11.

Figure 5.11. Calculated current-voltage characteristics for tandem solar cells in series configuration.

The bottom cell (cell 1) is combined with itself as top cell as well as with the other cells that have

different fill factors.

It is clear that the lower fill factor of the top cell also strongly decreases the fill

factor of the series tandem device. Furthermore, the fill factors of the series

99

tandem devices are higher than the fill factors of the top cells. Only when the

bottom cell (cell 1) is combined in a tandem cell with itself, the fill factor of the

tandem device equals to the fill factor of the bottom cell (cell 1, 66%). The values

extracted from Figure 5.11 are given in the table 5.3. Combining the highest (66%)

and lowest (25%) FF as sub cells in a series tandem device leads to a FF of 38%,

which is below the average value (45.5%).

Series tandem cell

number

VOC

[Volt]

JSC

[A/m2]

FF

[%]

η

[%]

1 1.18 95.3 66 7.42

2 1.18 95.3 59 6.63

3 1.18 95.3 54 6.07

4 1.18 95.3 48 5.39

5 1.18 95.3 45 5.06

6 1.18 95.3 38 4.27

Table 5.3. Electrical performance of the various series tandem cells with cell 1 as bottom and cell 1 to 6

as top cell. The data are extracted from Figure 5.14.

For the parallel configuration the results of tandem cells based on the bottom cell

(cell 1) with itself and the other 5 sub cells as top cells are given in Figure 5.12.

100


Figure 5.12. Current-voltage characteristic of parallel tandem solar cells with cell 1 as bottom cell and

cells 1 to 6 as top cell, respectively.

Again it is clear that the fill factor of the top cell limits the performance of the

parallel tandem device by lowering its fill factor, equal to the series configuration.

However, the fill factor of the parallel tandem cell is higher than the series

configuration in all cases. The solar cell parameters extracted from Figure 5.15 are

given in the table 5.4 see Table 5.4.

101

Parallel tandem

cell number

JSC

[A/m2]

VOC

[Volt]

FF

[%]

η

[%]

1 -190.6 0.59 66 7.42

2 -190.6 0.59 59 6.80

3 -190.6 0.59 58 6.28

4 -190.6 0.59 52 5.90

5 -190.6 0.59 49 5.62

6 -190.6 0.59 42 5.00

Table 5.4. Electrical performance of a series of parallel tandem cells with cell 1 as bottom cell and cells

1 to 6 as top cell. The data are extracted from Figure 5.12.

In order to demonstrate the different behavior of the series- and parallel

configuration tandem device when the fill factor of one sub cell is varied, we plot

the fill factor of the tandem devices as a function of the fill factor of the top cell, as

shown in Figure 5.13. The mathematical average, which is the sum of the fill

factors of the sub cells divided by two, is also plotted.

102


Figure 5.13. Fill factor of the tandem devices, series and parallel, as a function of the fill factor of the

top cell. The parallel configuration shows a higher fill factor as compared with the series configuration.

When the two sub cells have an equal fill factor both the series- and parallel

configuration have that same fill factor. When the top cell has a significantly lower

fill factor the parallel configuration follows the mathematical average and shows a

higher fill factor as compared to the series one. This is also the same for the

change in the power conversion efficiency of the tandem cells. The parallel tandem

has higher efficiency than the series one. The performance of all tandem cells

considered are compared with the sum of the efficiencies of the sub cells in Figure

5.14.

Figure 5.14. Efficiency of the series- and parallel tandem cells considered as well as the sum of the

efficiency of the sub cells in each case as a function of the efficiency of the top cell. When both sub cells

have similar electrical performance, both series and parallel configuration leads to nearly identical

efficiencies. If one of the sub cells (top cell here) exhibits a lower fill factor, the parallel configuration is

the better choice to fabricate.

It is demonstrated that the series tandem solar cell is affected to a larger extent by

the unbalanced electrical performance of the sub cells as compared to the parallel

configuration. The mathematical average of the fill factors of the sub cells is a good

103

approximation for the fill factor in the parallel configuration. The series

configuration has significantly lower fill factor and therefore lower efficiency.

5.5 Summary

In this chapter it has been presented how the current-voltage characteristic of any

arbitrary tandem device can be derived from the electrical performance of the sub

cells. The calculated characteristics are in very good agreement with experimental

data on both series- and parallel connected tandem devices. In general, when both

sub cells have almost the same electrical properties, series and parallel

configurations lead to tandem devices with the same performance. If there are

large differences in the electrical performances of the sub cells, the parallel

configuration is a better structure since its overall efficiency is higher than the

series configuration.

104


105

Chapter 6

Down-Conversion of High Energy Photons

Abstract

We demonstrate a simple technology to improve the total absorption of the solar

flux by semitransparent organic bulk heterojunction (BHJ) solar cells through down

conversion. The semitransparent BHJ organic solar cell is fabricated from a blend

of regioregular poly(3-hexyl thiophene) (P3HT) and the fullerene derivative [6,6]-

phenyl-C61-butyric acid methyl ester (PCBM). By employing a prototype composite

semitransparent cathode based on metals and a highly photoluminescent polymer

a maximum transmission from 20 to 60 % over the whole visible range is achieved.

The semitransparent solar cell has a power conversion efficiency of 2.7 % under

simulated AM1.5, 1000 W/m2 light. The high photoluminescence quantum

efficiency of the polymer on top of the semitransparent cathode improves the

amount of light reflected back into the device by down conversion [83,84 of the non-

absorbed high energy photons. Such a semitransparent luminescent cathode

improves the performance of the device.

106


6.1 Introduction

Solar cells have attracted much attention lately as an extremely reliable green

energy source, but it will still take time before the price of the photovoltaic systems

is low enough to be used in large scale/ large area applications. An interesting

option is to integrate solar cells into a building as roof or window materials, as

shown in Fig. 6.1 With semitransparent solar cells, windows can be built that

generate electricity while retaining their basic functionality.

Figure 6.1. Left: Solar panels serves as window of the building (CIS Tower in Manchester, U.K.). Right:

Solar panels are used as roof materials. The pictures are extracted from the Internet.

Windows with integrated semitransparent solar cells are especially interesting for

urbanized areas and business districts, where most of the earth’s surface is

covered with glass and concrete. A possible candidate for such an applications are

semitransparent solar cells based on conjugated polymers and PCBM. Because of

their solution processability they are relatively cheap and can be easily coated on

large areas as windows, they are light-weight and flexible which is convenient for

107

construction purposes. However, an inherent disadvantage of semitransparent

solar cells is that by definition not all the solar light can be absorbed, leading to

lower efficiencies as compared to conventional solar cells.

In spite of this disadvantage, advanced computational methods demonstrate that

due to optimum utilization of daylight 10-40% energy saving can be achieved,[85]

depending on the shape of the building and the climate zones. Different kinds of

semi-transparent photovoltaics (PVs) are reported based on inorganic [86,87]. In this

Chapter, a semitransparent organic bulk heterojunction solar cell is discussed

based on a blend of P3HT:PCBM (1:1). The power conversion efficiency of such a

cell can be improved by down-conversion of high-energy photons, as explained

here.

6.2 Optimum semitransparent cathode

Usually, in organic semitransparent solar cells a thin semitransparent layer of gold

or silver is used as cathode.[88,89] In order to create different levels of transparency

of the solar cell, different thicknesses for the metal can be used. In Figure 2 the

optical transmissions are shown for different layer thicknesses of gold (Au),

deposited onto glass substrates with a roughness of 1.5 nm. The gold layers are

applied by thermal evaporation under 10-7 mbar pressure. As Figure 6.2

demonstrates, a semitransparent cathode based on 10 nm gold allows a peak

transmission of about 60% around 500 nm and an averaged transmission between

300 to 1200 nm of about 40%.

108


Figure 6.2. Optical transmission of different layers of gold deposited on glass substrates. At a

thickness of 10 nm, the layer is semi-continuous and contains many holes. Therefore, it is more

transparent as compared to the thicker layers (20, 30 and 40 nm).

The roughness of the substrate has a significant effect on the optical and the

electrical properties of the deposited gold (or Ag) layer. To demonstrate this, we

can compare the optical transmission of gold layers with the same layer thickness,

but now thermally deposited onto a rougher (~10–15 nm) surface of a 2000 nm

thick PTrFE layer that was spin coated onto a glass substrate (the structure used

for this measurement is shown in the inset of Figure 6.4). The optical transmission

of the gold layers with different thicknesses on the PTrFE surface is shown in

Figure 6.3.

109

Figure 6.3. The optical transmission of gold layers with different thicknesses. The layers are thermally

deposited onto the surface of a 2000 nm thick layer of spin coated PTrFE on glass with roughness of

about 15 nm.

Figures 6.2 and 6.3 demonstrate that two layers with the same thickness, but

deposited onto substrates with different roughness, are different in their optical

transmissions. The gold (Au) layer deposited on the smooth surface of glass with a

roughness of about 1.5 nm is more closed and continuous, and less transparent.

When gold (Au) is deposited onto a rough PTrFE surface the optical transmission

is higher, since more holes exists in the layer (semi-continuous layer). For

example, a layer of gold with a thickness of 40 nm on the PTrFE surface

(roughness ~ 15 nm) has the same transparency as a gold (Au) layer of 20 nm on

glass (roughness ~ 1.5 nm), as shown in Figure 6.4. It is also clear from Figure 6.4

(circles) that the optical transmission measured on the PTrFE surface is affected by

the optical cavities. Therefore, the optical transmission shows modulations (circles

in Figure 6.4) due to interference between transmitted and reflected light inside the

PTrFE layer. The structure of the test devices are also given (inset Figure 6.4).

110


Figure 6.4. The optical transmissions of gold layers are compared. 40 nm gold layer on PTrFE is as

transparent as a 20 nm Au layer on glass because of differences in roughness of the substrates used.

The transmission on PTrFE surface also shows modulations caused by the optical cavity inside the

PTrFE layer. The structures which are used for the optical measurements are shown in the inset.

In order to define the best trade-off between optical and electrical properties of the

semitransparent cathode, the sheet-resistance of the cathode of the device has to

be considered next. In the same way, the roughness of the substrate affects the

sheet-conductivity of the deposited gold layer. On the smooth glass substrate, the

sheet-resistance of the Au cathode with a thickness of 20 nm is already low

enough (1 Ω/cm2) since in the final PV cell the anode used (140 nm ITO + 50 nm

PEDOT:PSS) in the device has a sheet-resistance of about 35 Ω/cm2. However,

the same layer thermally deposited onto rough PTrFE substrates has a sheet-

resistance of about 40 Ω/cm2, which is 40 times higher compared to the smooth

layer on glass! This is demonstrated in Figure 6.5.

In the remainder of this Chapter, for the processing of the cathode of the

semitransparent BHJ solar cell, a 3 nm layer of Samarium (Sm) is vapor-deposited

and topped with 12 nm of gold.

111

Figure 6.5. Sheet-resistance of a thin layer of vapor-deposited Au on bare glass and on a spin coated

layer of PTrFE is shown as measured by the 4-probe resistivity method. When the layer of gold is semi-

continuous (for example 20 nm of Au on PTrFE), higher roughness leads to higher electrical sheet-

resistance.

6.3 Down-conversion effect

Using a thin semitransparent metallic layer as the cathode of the device leads to a

lower reflection at the cathode, and therefore to a smaller amount of light that is

absorbed inside the active layer of the solar cell. The reduction of light intensity

inside the active layer of device, when using a semitransparent cathode, leads to a

lower performance (lower generated photocurrent) of the device.[38,37] This effect is

demonstrated in Figure 6.6. Two bulk heterojunction (BHJ) solar cells are shown

with a layer thickness of 250 nm consisting of a blend of P3HT:PCBM (1:1) as the

active layer. Both BHJ cells are processed on the standard substrate

ITO/PEDOT:PSS, and the only difference between the two cells is the cathode of

the device.

112


Figure 6.6. A standard (not transparent) BHJ cell made of 250 nm P3HT:PCBM blend (circles) is

compared with a semitransparent solar cell with the same active layer thickness and composition

(triangles). The transparency of the cathode leads to a reduction of the light intensity inside of the

device and, therefore, to a lower photocurrent.

From Figure 6.6 (triangles) it is clear that the combination of 3 nm Sm and 12 nm

Au leads to an Ohmic contact with a sheet conductance of the cathode that is high

enough to extract the electrons of the device. The lower generated photocurrent of

the semitransparent cell is due to the lower amount of absorbed light inside the

active layer of the device.[43,44] In order to improve the performance of the cell

without losing the transparency of the device in most part of the visible range of the

spectrum, we propose to down convert the high energy photons (blue, between

380 and 480 nm) that are not absorbed by the P3HT:PCBM layer to photons that

can be absorbed by the P3HT:PCBM (1:1) active layer (green/red, between 520

and 670 nm). In order to measure the optical properties of materials used here, a

layer of blend of P3HT:PCBM (1:1) and a layer of the poly(p-phenylene vinylene )

derivative super yellow-PPV (SY-PPV) were spin coated onto cleaned glass

substrates. The results of these measurements are shown in Figure 6.7.

113

Figure 6.7. The absorption of the P3HT:PCBM (1:1) blend (squares) is compared with absorption

(triangles) and photoluminescence (circles) of the polymer super yellow-PPV. A part of the solar

irradiance AM1.5 is also given . Part of the light with wavelengths between 380 nm and 480 nm is

transmitted through the P3HT:PCBM (1:1) device and is absorbed by the super yellow-PPV that is spin

coated on top of the semitransparent cathode. Super yellow emits light back in all direction. The emitted

light back inside the device with wavelength betweens 520 nm and 670 nm, can be absorbed by the

active photovoltaic layer.

As demonstrated in Fig. 6.7, part of the solar spectrum (380-480 nm) is not

absorbed by the P3HT:PCBM (1:1) active layer and is transmitted through the

device. This light is absorbed (shaded area 1) by the SY-PPV. Subsequently, the

photoluminescence of the super yellow PPV (520-670 nm) has quite some overlap

with the absorption of the active layer (shaded area 2). In this way, part of the solar

spectrum outside the absorption region of the active layer is transferred to longer

wavelengths, which can be absorbed by the solar cell. As a result the total

absorption of the photovoltaic cell can be improved. On the other hand, the

additional photoluminescent layer will also affect the transparency of the solar cell.

114


6.4 Device fabrication

The cells were fabricated on cleaned glass substrates. For the anode of the

devices, 140 nm ITO on glass is used, followed by a spin coated layer of 50 nm

poly(3,4-ethylene dioxythiophene) : polystryrenesulfonic acid (PEDOT:PSS, Bayer

AG). The active layers of all cells are 250 nm and consist of a blend of regioregular

poly(3-hexylthiophene) (P3HT) and the fullerene derivative [6,6]-phenyl-C61-butyric

acid methyl ester (PCBM) (1:1) is used. The active layer is annealed at 135 °C for

15 minutes. For the semitransparent cathode, by using shadow mask, 3 nm

Samarium (Sm) (work function of Sm ~ 2.7 eV)[82] and 12 nm Gold (Au) were

evaporated at 10-7 mbar. For the standard (not transparent) solar cell, 3 nm thick

Sm and 80 nm thick Aluminum were used for the cathode of the device which

optically serves as a mirror. For the down-conversion an additional 100 nm layer of

SY-PPV was spin coated onto the Sm/Au semitransparent cathode. The SY-PPV

was processed from toluene. The Sm/Au cathode of the device also protects the

active layer of the cell during the spin coating of the super yellow, as already

mentioned in Chapter 4. The structures of all cells are given in Figure 6.8.

115

Figure 6.8. Structure of a ‘standard’ P3HT:PCBM bulk heterojunction solar cell with A) a non-

transparent cathode (3 nm Sm + 80 nm Al), B) Structure of a semitransparent solar cell using 3 nm Sm

+ 12 nm Au for the semitransparent cathode of the device, C) Structure of a semitransparent solar cell

with an additional luminescent layer.

6.5 Performance improvement

In order to investigate the effect of the additional luminescent SY-PPV layer, the

electrical and optical properties of the device shown in Figure 6.8.C have to be

compared with the reference device shown in Figure 6.8.B. Therefore, the three

cells mentioned in Figure 6.8 were fabricated from the same solution, with identical

processing steps resulting in a 250 nm P3HT:PCBM (1:1) BHJ layer. The J–V

characteristics of all cells were measured under standard test condition using

simulated solar light. The results are shown in Figure 6.9.

116


Figure 6.9. The J–V characteristic of three BHJ solar cells are compared in dark and under 1000 W/m2,

AM1.5 simulated solar spectrum. The maximum photocurrent is achieved when a reflecting, closed Al

layer is used for the cathode (squares). When the cell is semitransparent (triangles), about 30% less

current is generated under illumination of the device. By using a layer of photoluminescent polymer, on

top of the semitransparent cathode, the photocurrent is improved by about 11%.

All-important parameters of the measurements above are extracted and

summarized in Table 6.1.

Cathode VOC (V) JSC (A/m2) FF (%) η (%)

3 nm Sm/ 80 nm Al 0.59 96 68 3.8

3 nm Sm/ 12 nm Au 0.57 70 61 2.4

3 nm Sm/ 12 nm Au/ 100 nm SY 0.56 80 61 2.7

Table 6.1. The values extracted from Figure 6.9.

Clearly, the additional layer of SY-PPV slightly improves the performance of the

semitransparent solar cell. The effect is not so large because the absorption

117

spectrum of the SY-PPV does not differ that much from the absorption spectrum of

the blend. As a result a relatively low fraction of the light that is not absorbed by the

250 nm P3HT:PCBM (1:1) blend is additionally absorbed in the SY-PPV.

Subsequently, the optics of the devices show in Figure 6.8b and 6.8c are

compared: The optical transmission was measured between 300 nm and 1200 nm.

The results are shown in Figure 6.10.

Figure 6.10. The optical transmission of a conventional semitransparent solar cell (triangle) as

compared to the optical transmission of the same cell but with a photoluminescent cathode (circle). The

transmissions of both cells are similar above 480 nm.

It is demonstrated that the cell with the photoluminescent SY-PPV layer has a

lower optical transmission between 380 and 480 nm, where the polymer has a high

absorption. In the remainder of the visible range of the spectrum both devices have

the same transmission. As a result, the total absorption of the solar cell is improved

by the photoluminescent layer without losing the transparency of the device for a

large part of the visible and infrared range of the spectrum. To demonstrate that

the improved efficiency is not due to interference effects of the polymer layer on

top of the cathode, one additional semitransparent device was processed. In this

experiment, the semitransparent PV cell is topped with a 100 nm thick blend of

super yellow and 3 w-% PCBM (Figure 6.12). The mixing of a very small amount of

PCBM (3 wt-%) with super yellow leads to complete photoluminescent quenching

of the layer, since PCBM serves as an efficient exciton quencher. The effective

quenching is shown in Figure 6.11. The present of such a small amount of PCBM

118


mixed with the SY-PPV layer has no significant effect on the optical absorption or

reflection of the layer.

Figure 6.11. Photoluminescence of the pure polymer SY-PPV after exciting the solid film on a glass

substrate with light of 460 nm. The blend of SY-PPV and 3 wt-% PCBM shows no photoluminescence at

all after exciting the film with the same intensity and wavelength.

Figure 6.12. A semitransparent BHJ solar cell with a blend of SY-PPV and PCBM spin coated onto the

cathode of the device.

The J–V characteristic of the above-mentioned device was measured in dark and

under illumination. As Figure 6.13 demonstrates, the performance of this device is

identical to the semitransparent cell in which only metals (3 nm Sm + 12 nm Au)

119

are used for the fabrication of the semitransparent cathode (data extracted from

Figure 6.8, triangles). This means that the presence of the quenching layer based

on the blend of super yellow-PPV and PCBM has no significant effect on the

performance of the device. This demonstrates that is it is indeed the

photoluminescence of the SY-PPV that is responsible for the improved

performance.

Figure 6.13. The J–V characteristics of the semitransparent PV cell with an additional layer of a blend

of ‘super yellow’ (SY) and PCBM onto the cathode (circles) compared to the performance of a

semitransparent PV cell without any polymer layer spin coated on top (triangles, data extracted from

Figure 5.8).

6.6 Summary

In this Chapter, a semitransparent BHJ organic solar cell with a maximum optical

transmission of 60 %, and a power conversion efficiency of ηp= 2.7 % under

simulated 1000 W/m2 , AM1.5 solar spectrum is discussed. The performance of a

semitransparent BHJ solar cell was increased by 11 % due to the addition of a

cathode with photoluminescent properties. The semitransparent cathode itself

consists of 3 nm Sm and 12 nm Au. This cathode proved to be stable against spin

coating of a toluene solution. The part of the solar irradiance between 380 and 480

nm, that is not absorbed by the active layer of the device (P3HT:PCBM (1:1)),

excites the SY-PPV polymer that is spin coated on top of the semitransparent

120


cathode of the device. The red-shifted photoluminescence of SY-PPV emits light

between 520 and 670 nm, which can be absorbed by the active layer

(P3HT:PCBM) of the device. Due to this processes, the photons with high energy

are converted into photons with lower energies. The down conversion matches the

photons to the absorption spectrum of the active layer and increases the

absorption of the active layer of the device.

121

References 1 The U.S. Department of Energy’s International Energy Outlook 2005.

2 United Nations Environment Programme (UNEP), Global environment outlook

(GEO yearbook 2004/5), web site: www.unep.org/geo/yearbook.

3 Intergovernmental Panel on Climate Change (IPCC), Third assessment report –

Climate Change 2001, web site: www.meto.gov.uk.

4 D. L. Pulfrey, Photovoltaic Power Generation, Van Nostrand Reinhold, New York 1978. 5 W. D. Johnston, Jr., Solar Voltaic Cells, Dekker, New York 1980. 6 D. M. Chapin, C. S. Fuller, and G. L. Pearson, J. Appl. Phys. 1954, 25, 676.

7 S. E. Shaheen, R. Radspinner, N. Peyghambarian, G. E. Jabbour, Appl. Phys.

Lett. 2001, 79, 2996.

8 G. Gustafsson, Y. Cao, G. M. Treacy, F. Klavetter, N. Colaneri, A. J. Heeger,

Nature 1992, 357, 477.

9 H. Mette, Z. Physik 1953, 134, 566.

10 R. G. Kepler, Phys. Rev. 1960, 119, 1226.

11 M. Pope, H. P. Kallmann, P. Magnante, J. Chem. Phys. 1963, 38, 2042.

12 C. K. Chiang, C. R. Fincher Jr., Y. W. Park, A. J. Heeger, H. Shirakawa, E. J.

Louis, S. C. Gau, A. G. MacDiarmid, Phys. Rev. Lett. 1977, 39, 1098.

13 J. H. Burroughes, D. D. C. Bradley, A. R. Brown, R. N. Marks, K. Mackay, R. H.

Friend, P. L. Burns, A. B. Holmes, Nature 1990, 347, 539.

14 H. Koezuka, A. Tsumura, Y. Ando, Synth. Met. 1987, 18, 699.

15 A. R. Brown, A. Pomp, C. M. Hart, D. M. de Leeuw, Science 1995, 270, 972.

16 G. Yu, K. Pakbaz, A. J. Heeger, Appl. Phys. Lett. 1994, 64, 3422.

122


17 S. E. Shaheen, C. J. Brabec, N. S. Sariciftci, F. Padinger, T. Fromherz, J. C.

Hummelen, Appl. Phys. Lett. 2001, 78, 841.

18 P. Schilinsky, C. Waldauf, C. J. Brabec, Appl. Phys. Lett. 2002, 81, 3885.

19 R. Hoffmann, C. Janiak, C. Kollmar, Macromolecules 1991, 24, 3725.

20 S. Barth, H. Bassler, Phys. Rev. Lett. 1997, 79, 4445.

21 P. G. Dacosta, E. M. Conwell, Phys. Rev. B 1993, 48, 1993.

22 R. N. Marks, J. J. M. Halls, D. C. Bradley, R. H. Friend, A. B. Holmes, J. Phys.

Cond. Mat. 1994, 6, 1379.

23 C. L. Braun, J. Chem. Phys. 1984, 80, 4157.

24 E. Goliber, J. H. Pettstein, J. Chem. Phys. 1984, 80, 4162.

25 H, Kallmann, M. Pope, J. Chem. Phys. 1959, 30, 585.

26 C. W. Tang, Appl. Phys. Lett. 1986, 48, 183.

27 J. J. M. Halls, C. A. Walsh, N. C. Greenham, E. A. Marseglia, R. H. Friend, S. C.

Morahi, A. B. Holmes, Nature 1995, 376, 498.

28 G. Yu, J. Gao, J. C. Hummelen, F. Wudl, A. J. Heeger, Science 1995, 270, 1789.

29 C. J. Brabec, N. S. Sariciftci, J. C. Hummelen, Adv. Funct. Mater. 2001, 11, 15.

30 J. C. Hummelen, B. W. Knight, F. Lepeq, F. Wudle, J. Yao, C. L. Wilkins, J. Org.

Chem. 1995, 60, 532.

31 V. D. Mihailetchi, J. K. J. van Duren, P. W. M. Blom, J. C. Hummelen, R. A. J.

Janssen, J. M. Kroon, M. T. Rispens, W. J. H. Verhees, M. M. Wienk, Adv. Funct.

Mater. 2003, 13, 43.

32 V. D. Mihailetchi, P. W. M. Blom, J. C. Hummelen, M. T. Rispens, J. Appl. Phys.

2003, 94, 6849.

33 V. D. Mihailetchi, L. J. A. Koster, J. C. Hummelen, P. W. M. Blom, Phys. Rev.

Lett. 2004, 93, 216601.

123

34 V. D. Mihailetchi, L. J. A. Koster, P. W. M. Blom, Appl. Phys. Lett. 2004, 85, 970.

35 C. Melzer, E. J. Koop, V. D. Mihailetchi, P. W. M. Blom, Adv. Funct. Mater. 2004,

14, 865.

36 V. D. Mihailetchi, J. Wildeman, P. W. M. Blom, Phys. Rev. Lett. 2005, 94,

126602.

37 L. J. A. Koster, V. D. Mihailetchi, H. X. Xie, P. W. M. Blom, Appl. Phys. Lett.

2005, 87, 203502.

38 L. J. A. Koster, V. D. Mihailetchi, R. Ramaker, P. W. M. Blom, Appl. Phys. Lett.

2005, 86, 123509.

39 N, Hoppe, N. Arnold, N. S. Sariciftci, D. Meissner, Solar Energy Mater. & Solar

Cells 2003, 80, 105.

40 C. J. Brabec, A. Gravino, D. Meissner, N. S. Sariciftci, M. T. Rispens, L.

Sanchez, J. C. Hummelen, T. Fromherz, Thin Solid Films 2002, 403, 368.

41 I. Riedel, J. Parisi, V. Dyakonov, L. Lutsen, D. Vanderzande, J. C. Hummelen,

Adv. Funct. Mater. 2004, 14, 38.

42 P. W. M. Blom, V. D. Mihailetchi, L. J. A. Koster, and D. E. Markov, Adv. Mater.

2007, 19, 1551-1566.

43 V. D. Mihailetchi, “Device Physics of Organic Bulk Heterojunction Solar Cells”

2005 PhD thesis, University of Groningen. Online:

http://irs.ub.rug.nl/ppn/288520572

44 L. J. A. Koster, “Device physics of donor/acceptor-blend solar cells” 2007 PhD

thesis, University of Groningen. Online: http://irs.ub.rug.nl/ppn/299329410

45 O. H. Le Blanc, J. Chem. Phys. 1960, 33, 626.

46 B. A. Gregg, J. Phys. Chem. B 2003, 107, 4688.

47 J. Nelson, Curr. Opin. Solid State Mater. Sci. 2002, 6, 87.

48 G. A. Chamberlain, Solar cells 1983, 8, 47.

49 a. P. Peumans, A. Yakimov, ans S. R. Forrest, J. Appl. Phys. 2003, 93, 3693.

124


b. S. M. Sze, Physics of Semiconductor Devices, JOHN WILEY & SONS (New

York) ISBN: O-471-O9837-X.

50 T. A. Skotheim, R. L. Elsenbaumer, J. R. Reynolds, Handbook of Conducting

Polymers 1998, New York.

51 M. Pope, C. E. Swenberg, Electronic Processes in Organic Crystals and

Polymers 1999, 2nd ed., Oxford University Press, New York.

52 A. Goetzberger, C. Hebling, Sol. Ener. Mater. Sol. Cells 2000, 62, 1.

53 J.-M. Nunzi, C. R. Physique 2002, 3, 523.

54 D. Wöhrle, D. Meissner, Adv. Mater. 1991, 3, 129.

55 H. Spanggaard, F. C. Krebs, Sol. Ener. Mater. Sol. Cells 2004, 83, 125.

56 K. M. Coakley, M. D. McGehee, Chem. Mater. 2004, 16, 4533.

57 M. Hiramoto, M. Suezaki, M. Yokoyama, Chem. Lett. 1990, 327.

58 A. Yakimov, S. R. Forrest, Appl. Phys. Lett. 2002, 80, 1667.

59 P. Peumans, V. Bulovic, S. R. Forrest, Appl. Phys. Lett. 2000, 76, 2650.

60 a) D. E. Markov, E. Amsterdam, P. W. M. Blom, A. B. Sieval, J. C. Hummelen,

J. Phys. Chem. A 2005, 109, 5266. b) D. E. Markov, J. C. Hummelen, P. W. M.

Blom, A. B. Sieval, Phys. Rev. B 2005, 72, 045216. c) D. E. Markov “Excitonic

processes in polymer-based optoelectronic devices”, 2006, PhD Thesis, University

of Groningen. Online:http://irs.ub.rug.nl/ppn/296023094

61 a) K. Triyana, T. Yasuda, K. Fujita, T. Tsutsui, Jp. J. Appl. Phys. 2004, 43,

2352. b) K. Triyana, T. Yasuda, K. Fujita, T. Tsutsui, Thin Solid Films 2005, 447,

198.

62 J. Xue, S. Uchida, B. P. Rand, S. R. Forrest, Appl. Phys. Lett. 2004, 85, 5757.

63 J. Xue, B. P. Rand, S. Uchida, S. R. Forrest, Adv. Mater. 2005, 17, 66.

64 B. Maenning, J. Drechsel, D. Gebeyehu, P. Simon, F. Kozlowski, A. Werner, F.

Li, D. Grundmann, S. Sonntag, M. Koch, K. Leo, M. Pfeiffer, H. Hoppe, D.

125

Meissner, N. S. Sariciftci, I. Riedel, V. Dyakonov, J. Parisi, Appl. Phys. A: Mater.

Sci. Process. 2004, 79, 1.

65 J. Drechsel, B. Männig, D. Gebeyehu, M. Pfeiffer, K. Leo, H. Hoppe, Organic

Electronics 2004, 5, 175-186.

66 G. Dennler, H.-J. Prall, R. Koeppe, M. Egginger, R. Autengruber, N. S.

Sariciftci, Appl. Phys. Lett. 2006, 89, 073502.

67 L. Chen, D. Godovsky, O. Inganäs, J. C. Hummelen, R. A. J. Janssen, M.

Svensson, M. R. Andersson, Adv. Mater. 2000, 12, 1367.

68 V. Shrotriya, E. H. Wu, G. Li, Y. Yao, Y. Yang, Appl. Phys. Lett. 2006, 88,

064104.

69 M. Lenes, L. J. A. Koster, V. D. Mihailetchi, P. W. M. Blom, Appl. Phys. Lett.

2006, 88, 243502.

70 K. Kawano, N. Ito, T. Nishimori, J. Sakai, Appl. Phys. Lett. 2006, 88, 073514.

71 J. Gilot, M. M. Wienk, R. A. J. Janssen, Appl. Phys. Lett. 2007, 90, 143512.

72 C. Pacholski, A. Kornowski, H. Weller, Angew. Chem. Int. Ed. 2002, 41, 1188.

73 W. J. E. Beek, M. M. Wienk, R. A. J. Janssen, Adv. Mater. 2004, 16, 1009.

74 W. J. E. Beek, M. M. Wienk, M. Kemerink, X. Yang, R. A. J. Janssen, J. Phys

Chem B 2005, 109, 9505.

75 F. Verbakel, S. C. J. Meskers, R. A. J. Janssen, Appl. Phys. Lett. 2006, 89,

102103.

76 J. Y. Kim, K. Lee, N. E. Coates, D. Moses, T. Nguyen, M. Dante, A. J. Heeger,

Science 2007, 317, 222.

77 J. Y. Kim, S. H. Kim, H. Lee, K. Lee, W. Ma, X. Gong, A. J. Heeger Adv. Mater.

2006, 18, 572.

78 D. Muhlbacher, M. Scharber, M. Morana, Z. Zhu, D. Waller, R. Gaudiana, C.

Brabec Adv. Mater. 2006, 18, 2884.

79 P. Peumans, S. R. Forrest, Appl. Phys. Lett. 2001, 79, 126.

126


80 J. Drechsel, B. Männig, F. Kozlowski, M. Pfeiffer, K. Leo, H. Hoppe, Appl.

Phys. Lett. 2005, 86, 244102

81 V. D. Mihailetchi, H. Xie, B. de Boer, L. J. A. Koster, P. W. M. Blom, Adv.

Funct. Mater. 2006 16, 699.

82 “Handbook of chemistry and physics”, 75th Ed., Eds. David R. Lide, CRC press

(1995). We have evaporated 20 nm samarium on glass and measured its work

function to be 2.4 eV with the Kelvin probe under N2 atmosphere.

83 T. Trupke, M. A. Green, J.Appl. Phys. 2002, 92, 3.

84 L. H. Slooff, R. Kinderman, A. R. Burgers, J. A. M. van Roosmalen, A.

Büchtemann, R. Denz, M. Schleusener, A. J. Chatten, D. Farrel, K. W. J. Barnham,

20th European Photovoltaic Solar Energy Conference and Exhibition, Barcelona

Spain, 6-10 June 2005.

85 Sullivan R, Lee ES, Selkowits S. A methode of optimizing solar control and

daylighting performance in commercial offic building. Proceeding of

ASHRAE/DOE/BTECC Conference on the Thermal Performance of the Extrerior

Envelopes of Buildings V. 1992, p. 313-9.

86 F. P. Baumgartner, D. Sutter, THE SOLAR ROOF WINDOW, 14th European

photovoltaic Solar Energy Conference, Barcelona Spain, 30 Jun-4 July 1997.

87 Satyen K. Deb, Photovoltaic-Integrated Electrochromic Device for Smart-Window

Applications, NREL / CP-590-28116, May 2000.

88 Rhonda F. Bailey-Salzman, Barry P. Rand, and Stephen R. Forrest, Appl. Phys.

Lett. 2006, 88, 233502.

89 M. G. Kang, N. Park, Y. J. Park, K. S. Ryu, and S. H. Chang, Sol. Energy Mater.

Sol. Cells 2003, 75, 475.

127

Publications 1. A. Hadipour, B. de Boer, J. Wildeman, F. B. Kooistra, J. C. Hummelen, M. G. R.

Turbiez, M. M. Wienk, R. A. J. Janssen, P. W. M. Blom, “Solution-processed

Organic Tandem Solar Cells” Adv. Funct. Mater. 2006, 16, 1897-1903.+ front cover of issue 14, Vol. 16 (2006).

2. A. Hadipour, B. de Boer, P. W. M. Blom, Solution-processed Organic Tandem

Solar Cells with embedded Optical Spacers J. of Appl. Phys. 2007, 102, 074506.

3. A. Hadipour, B. de Boer, P. W. M. Blom, “Review on Organic Tandem and

Multi-junction Solar Cells”, accepted as Feature Article in Adv. Funct. Mater. 2007.

4. A. Hadipour, B. de Boer, P. W. M. Blom, “Current-Voltage Characteristic of

Organic Tandem Cells”, Organic Electronics, submitted 2007.

5. A. Hadipour, B. de Boer, J. Wildeman, F. B. Kooistra, J. C. Hummelen, M. G. R.

Turbiez, M. M. Wienk, R. A. J. Janssen, P. W. M. Blom, “Organic multi-junction

solar cells processed from solution with sensitivity from ultraviolet to the near

infrared” Proc. SPIE Int. Soc. Opt. Eng. 2006, 6192, 61920D1-13.

6. B. de Boer, A. Hadipour, M. M. Mandoc, P. W. M. Blom, “Tuning of Metal Work

Functions with Self-assembled Monolayers” Mat. Res. Soc. Symp. Proc. 2005,

871E, 1612.

7. B. de Boer, A. Hadipour, R. Foekema, T. van Woudenbergh, M. M. Mandoc, V.

D. Mihailetchi, P. W. M. Blom, “Tuning of Metal Work Functions with Self-

Assembled Monolayers”, Proc. SPIE Int. Soc. Opt. Eng. 2004, 5464, 18-25.

8. F. P. Mena, A. B. Kuzmenko, A. Hadipour, N. A. Babushkina, D. van der Marel

“ Oxygen isotope effect in the optical conductivity of (La0.5Pr0.5)0.7Ca0.3MnO3 thin

films “ Physical Review B 2005, 72, 134422.

9. A. B. Kuzmenko, N. Tombros, F. P. Mena, A. Hadipour, H. J. A. Molegraaf, D.

van der Marel, M. Gruninger, S. Uchida “Redistribution of the c-axis spectral weight

128


and conductivity sum rules in LSCO as revealed by optical transmission” Physica

C 2004, 408-410, 330-331.

129

Summary

The global demand for energy is expanding continually. Therefore, realization of

green power sources are needed since combustion of fossil fuels will have serious

consequences for the climate on the Earth. With a photovoltaic device, the solar

light can be converted into electricity which is the most useful forms of energy. For

this reason, solar cells have attracted much attention in the last decades as most

clean, sustainable and renewable energy sources. In order to produce low-cost

and large-area solar cells, organic materials provides many possibilities. Especially

semiconducting polymers combine the favorable opto-electronic properties, such a

high absorption coefficients, of organic materials with the excellent processing and

mechanical properties of plastic materials. This implies that an organic solar cell

can be processed from solution at room temperature onto (flexible) substrate using

simple and, therefore, much cheaper methods such as spin (or blade) coating and

inkjet printing. However, the formation of a bound-electron-hole pair that needs to

be separated, the low mobility of charge carriers (or the mobility difference

between holes and electrons) together with relatively narrow absorption spectra of

the organic materials lead to relatively low performance (typically amounts to 4-

6%). To improve the absorption of the solar radiation by organic solar cells,

materials with a broad absorption band have to be designed or different narrow

band absorbers have to be stacked in multiple junctions. When two (or more)

donor materials with non-overlapping absorption spectra are used in a tandem (or

multi-junction) solar cell, a very broad range (from visible to infrared) of the solar

radiation can be absorbed by such a device. In addition to converting a larger part

of the spectrum, tandem solar cells distinct advantage that photon energy is used

more efficiently because the voltage at which chargers are collected in each sub

cell is closer to the energy of the photons absorbed in that cell. There are several

approaches organic tandem (multiple) cells reported in the last years, depending

130


on materials used for the active layer and the proper separating layer(s). The

multiple organic solar cells can be summarized in three classes;

A) Tandem (multi-junction) solar cells in which low molecular weight organic

molecules are used for all sub cells.

B) Hybrid tandem organic solar cells in which the bottom cell is processed from

polymers by solution-processing while the top cell is made of vacuum-deposited

low molecular weight organic molecules.

C) Fully solution-processed tandem or multi-junction solar cells in which all sub

cells are processed from polymers.

The main subject of this thesis is solution-processed tandem solar cells that

consists of two bulk heterojunction sub cells with complementary absorption

spectra. A composite middle electrode separated the two sub cells. This middle

contact serves two different purposes; as a charge recombination center, and as a

protecting layer for the fist cell during spin coating of the second cell. Since the

bottom and the top cell are electrically stacked in series in this tandem cell, the

open-circuit voltage of the tandem cell equals the sum of the open-circuit voltages

of both sub cells. An open-circuit voltage of 1.4 V is achieved. The short-circuit

current of this tandem cell is limited by the lowest value which is the current of the

top cell. In this cell, the layer thickness of the bottom cell has to be optimized to in

such a way that the optical out coupling is adapted to the absorption of the top cell.

A disadvantage of this approach is that the optimum thickness of the bottom cell,

required to match the optical output, is not necessarily equal to the thickness

where the bottom cell reaches its optimum performance. In order to improve the

above-mentioned structure, a additional solution-processable, transparent and

insulating layer can be used as separating layer. This layer serves as an optical

spacer and leads to the fabrication of a 4-electrode tandem cell. In this way, the

thickness of the bottom cell is optimized for its electrical performance and the

optical out coupling is tuned by varying the thickness of the optical spacer. The two

sub cells can be coupled electrically in parallel or series since this optical spacer is

131

an insulator. The parallel configuration leads to a higher performance because the

extracted current from the tandem device is not limited by the low current of the top

cell any more.

The solar cells are not cheap enough yet for large-area applications. One of the

ways to reduce the payback period (economic return-on-investment) is integrating

the solar cells into a building. This means that the solar panel itself serves as the

building materials as well as the source of electricity generation. A solar cell can be

integrated into a building in different applications such as roofing structures and

window materials. It is clear that semitransparent solar cells can not only be used

in the tandem structures but also as windows on buildings. However, the

semitransparent solar cells have lower efficiencies since less light can be reflected

inside the device by the semitransparent cathode. The performance of

semitransparent BHJ solar cells can be improved due to the addition of a proto

type semitransparent cathode with photoluminescent properties. This luminescent

cathode can be used in general for solution-processed organic semitransparent

solar cells used for PV-windows applications.

132


133

Samenvatting De vraag naar energie die nodig is voor het leven vermeerdert zich in de komende

tijd. Het gebruik van fossiele bronnen zoals olie en gas brengt veel schade aan de

natuur op de aarde. Daarom moeten we alternatieve energie bronnen zoals zonne-

energie of windenergie gaan gebruiken in de toekomst. Met een zonnecel kunnen

we de zonlicht omzetten in elektriciteit. De elektrische energie is een van de meest

nuttige vorm van energie, omdat het voor veel toepassingen gebruikt kan worden.

Om de zonnecellen toe te kunnen passen op grote schaal, moeten de

fabricatiekosten veel lager liggen dan wat ze nu zijn. Voor het maken van

goedkopere zonnecellen zijn organische materialen en in het bijzonder de

halfgeleidende polymeren een goed alternatief met veel mogelijkheden. Deze

polymeren hebben sterke lichtabsorptie en ze kunnen zoals plastic behandeld en

verwerkt worden. Dit betekent dat een organische zonnecel op een goedkope

manier kan worden gemaakt door middel van verspinnen of inkt-jet printen. Helaas

is de efficiëntie van de organische zonnecellen niet hoog genoeg om deze te

gebruiken voor commerciële toepassingen. De reden van de lage efficiëntie ligt

aan het feit dat absorptie van het zonlicht niet direct resulteert in vrije ladingsdrager

en aan de beperkte absorptie van het zonnespectrum door de huidige polymeren.

Daarnaast is er vaak een groot verschil tussen de mobiliteit van elektronen en

gaten in deze materialen, waardoor het ladingstransport uit evenwicht is. Om de

overlap van absorptie van de organische zonnecellen te verbeteren ten opzichte

van het zonlicht, moeten cellen met smallere absorptie gebieden met elkaar

gecombineerd worden in een tandem cel om zo tot een ‘volledige’ absorptie te

komen van de fotonenflux van de zon. Omdat elke cel van een tandem cel

absorbeert in een ander deel van het zonnespectrum, kan de tandem cel een

grotere deel van het zonlicht dekken. Dit proefschrift behandelt tandem

zonnecellen welke gemaakt zijn van twee enkele zonnecellen die op hun beurt

gemaakt zijn van een mengsel van een polymeer en een fullereen variant. Het

134


halfgeleidende polymeer dient om het licht te absorberen, om elektronen te

doneren aan het fullereen en om de gaten te geleiden naar de anode. De fullereen

afgeleide PCBM functioneert als elektronen accepterende eenheid en als

elektronen transporterende fase. Dit soort enkel cellen worden “bulk

heterojunction” (BHJ) zonnecellen genoemd. De twee enkel cellen zijn boven op

elkaar gefabriceerd met een scheidingslaag er tussen. Deze zogenaamde midden-

elektrode beschermt de polymeer laag van de eerste cel tijdens de fabricatie van

de tweede cel en geeft tevens de mogelijkheid voor de elektronen en gaten, die

afkomstig zijn van beide enkel cellen, om te recombineren. Op deze manier zijn

beide cellen elektrisch in serie geschakeld aan elkaar. Dat wil zeggen dat de

openklemspanning van de tandem cel gelijk is aan de som van

openklemspanningen van de twee enkel cellen. Daarmee is een hoge

openklemspanning bereikt van 1.4 volt. De totale stroom die uit deze tandem cel

gehaald kan worden wordt beperkt door de laagste stroom van één van de twee

cel. Het nadeel van deze structuur is dat de dikte van de eerste cel wordt

geoptimaliseerd voor beste optische koppeling tussen de twee enkel cellen en dat

deze dikte niet perse de optimale dikte hoeft te zijn voor de hoogst mogelijke

stroom of efficiëntie van deze onderste cel. Deze situatie kan verbeterd worden als

we een extra transparante, isoleerde tussenlaag toe voegen om de optische

koppeling te verbeteren. Dit houdt in dat de door interferentie bepaalde

intensiteitprofiel van de licht transmissie kan worden verschoven in energie door de

laagte dikte van het dielectrikum te variëren. Hierdoor kunnen beide enkel cellen

los van elkaar elektrisch geoptimaliseerd worden door de dikte van de actieve laag

te variëren. Vervolgens wordt de optische koppeling tussen beide cellen afgestemd

met behulp van deze extra transparant laag. De laagdikte van de transparante laag

tussen de twee enkel cellen kan onafhankelijk gevarieerd worden waardoor de

optische koppeling kan worden bewerkstelligd. Omdat deze transparante laag ook

elektrisch een isolator is, moet gewerkt worden met 4 elektroden om de gehele

zonnecel elektrisch uit te lezen. Dit betekent dat de twee enkel cellen zowel

parallel als in serie aan elkaar geschakeld kunnen worden. De parallelle

135

configuratie heeft een hogere efficiëntie omdat de stroom van de tandem cel niet

meer gelimiteerd is door de bovenste cel.

In dichtbevolkte steden is de ruimte om zonnecellen te plaatsen erg beperkt, omdat

de grondprijs fenomenaal hoog is, waardoor de kosten voor de opwekking van

elektriciteit via het fotovoltaische effect de pan uit rijzen. Om de prijs van

zonnecellen nog aantrekkelijker te maken, kunnen we dit soort zonnecellen

gebruiken als bouwmateriaal door bijvoorbeeld zonnecellen die semi-transparant

zijn te gebruiken als ramen. Dit soort toepassingen kan worden gezien als een

soort multi-tasking: De cellen werken als een normaal raam en op hetzelfde

moment wordt ook elektriciteit op gewekt. Een nadeel van de semi-transparante

zonnecellen is dat ze een lagere efficiëntie hebben ten opzichte van normale (niet

transparant) zonnecellen. De reden hiervoor is over duidelijk: minder licht wordt

geabsorbeerd, dus de efficiëntie is lager. Deze lage efficiëntie kan worden

verbeterd en gelijktijdig kan uniformiteit van de absorptie van het zonnespectrum

worden geoptimaliseerd. Dit kan worden bewerkstelligd door een speciale katode

te gebruiken die halfdoorzichtig is maar ook fotoluminescent eigenschappen heeft.

De fotoluminescente laag absorbeert hoog energetische fotonen (blauw licht) die

sowieso niet door de actieve fotovoltaische laag werden geabsorbeerd en zend

vervolgens laag energetische fotonen uit die wel kunnen worden geabsorbeerd

door de actieve laag. Deze speciale katode kan in het algemeen gebruikt worden

om verloren licht om te zetten in bruikbaar licht en daarbij de transparantie voor de

toepassing als raam te waarborgen.

136


137

Dankwoord Allereerste wil ik mijn promotor Paul Blom hartelijk bedanken. Paul, je was altijd

een vriendelijke leider met veel geduld en aandacht. Je gaf me op elke moment,

als het nodig was, de beste adviezen en begeleiding en uitleg. Ik heb heel veel van

je geleerd en veel genoten tijdens mijn promotie in jouw groep, hartelijk bedankt .

Ik wil graag mijn copromotor Bert de Boer bedanken maar mijn woorden komen te

kort voor dat. Bert, je stond altijd klaar voor me. Je was altijd vriendelijk en

geduldig en zorgzaam. Ik heb veel van je geleerd en genoten tijdens mijn promotie,

hartelijk bedankt. Ik wil alle heren van de lees comissie Kees Hummelen, Rene

Janssen en Paul Heremans hartelijk bedanken voor verbetering van mijn thesis.

Minte Mulder en Jan Harkema wil ik zeer veel bedanken voor technische

samenwerking. Ik heb veel van jullie geleerd en altijd met plezier uren lang in zo’n

prachtige clean room doorgebracht!

university of groningen polymer tandem solar cells

Documents