10 Complete

7/31/2019 10 Complete

http://slidepdf.com/reader/full/10-complete 1/108

lenes thesis final2.docx2

EFFICIENCY ENHANCEMENT OF

POLYMER FULLERENE SOLAR CELLS

Martijn Lenes

7/31/2019 10 Complete


Efficiency enhancement of polymer fullerene solar cellsMartijn LenesPhD thesisUniversity of Groningen, The Netherlands

Zernike Institute PhD thesis series 2009-13ISSN 1570-1530ISBN 978-90-367-4016-6ISBN 978-90-367-4057-9 (digital version)

This research forms part of the research programme of the DutchPolymer Institute (DPI), Technology Area Functional PolymerSystems, DPI project #524

7/31/2019 10 Complete


lenes thesis final2.docx

RIJKSUNIVERSITEIT GRONINGEN

EFFICIENCY ENHANCEMENT OF

POLYMER FULLERENE SOLAR CELLS

Proefschrift

ter verkrijging van het doctoraat in de Wiskunde en Natuurwetenschappen

aan de Rijksuniversiteit Groningenop gezag van de

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

vrijdag 23 oktober 2009om 16:15 uur

door

Martijn Lenes

geboren op 12 augustus 1977te Hoogezand-Sappemeer

7/31/2019 10 Complete


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

Beoordelingscommissie: Prof. dr. J. C. HummelenProf. dr. L.D.A. SiebbelesProf. dr. ir. P. Heremans

7/31/2019 10 Complete



Dedicated to the memory of

Bert de Boer (1973 – 2009)

7/31/2019 10 Complete


CONTENTS

6


Table of Contents

CHAPTER 1..................................................................................................................... 9 1.1 Introduction: Solar Energy ............................................................................ . 10 1.2 Organic Semiconductors ................................................................................ 12 1.3 Charge generation in organic solar cells .......................................................... 13 1.4 Device fabrication and characterization techniques ......................................... 15

1.4.1 Solar cell fabrication .................................................................................. 15 1.4.2 Solar cell Characterization ......................................................................... 15 1.4.3 Characterization techniques ........................................................................ 17 1.4.4 Photocurrent measurements and modeling .................................................. 20

1.5 Optimization of energy levels in a donor acceptor system ............................... 22 1.5.1 Multijunction solar cells ............................................................................. 24

1.6 Aim and scope of this thesis ........................................................................... 26 REFERENCES ............................................................................................................... 28

CHAPTER 2................................................................................................................... 31 2.1 Introduction ................................ ................................................................... 32 2.2 Space-charge limited photocurrents ................................................................ 33 2.3 Device Fabrication and measurements ............................................................ 35 2.4 Device Simulations ........................................................................................ 40 2.5 Optical Considerations ................................................................................... 43 2.6 Conclusions ................................................................................................... 45 REFERENCES ............................................................................................................... 46

CHAPTER 3................................................................................................................... 49 3.1 Introduction ................................ ................................................................... 50 3.2 Parameters governing the charge dissociation ................................................. 51 3.3 Single Carrier Devices.................................................................................... 53 3.4 PEO-PPV:PCB-EH Solar Cells ...................................................................... 55 3.5 Device Simulations ........................................................................................ 56 3.6 Conclusions ................................................................................................... 59 REFERENCES ............................................................................................................... 60

CHAPTER 4................................................................................................................... 63 4.1 Introduction ................................ ................................................................... 64 4.2 Charge transport in pristine PCPDTBT films .................................................. 65 4.3 Charge transport in PCPDTBT:PCBM blends ................................................. 66 4.4 PCPDTBT:PCBM Solar Cells ........................................................................ 68 4.5 Device Simulations and Discussion ................................................................ 72 4.6 Conclusions ................................................................................................... 76 REFERENCES ............................................................................................................... 77

7/31/2019 10 Complete


7


CHAPTER 5................................................................................................................... 79 5.1 Introduction ................................ ................................................................... 80 5.2 The bisadduct analogue of PCBM .................................................................. 81

5.2.1 P3HT:bisPCBM solar cells......................................................................... 82 5.3 Other higher adduct fullerenes ........................................................................ 85

5.3.1 Solar cells based on higher adduct fullerenes .............................................. 89 5.3.2 Device simulations using charge trapping ................................................... 93

5.4 Conclusions ................................................................................................... 96

PUBLICATIONS .................................................................................................................... 97

SUMMARY ............................................................................................................................ 99

SAMENVATTING ................................................................................................................ 102

ACKNOWLEDGMENTS ....................................................................................................... 105

7/31/2019 10 Complete


8


7/31/2019 10 Complete



Chapter 1

Introduction to organic solar cells

Abstract

Renewable energy sources are heavily sought after these days, both from an

environmental and from a cost point of view. Even though all forms of renewable

energy are indirectly powered by the sun, using the photovoltaic effect is probably

the most straightforward way of harnessing energy from the incoming solar

irradiation. Next to the traditional inorganic silicon-based photovoltaics, organic

solar cells are considered as a viable candidate for a large-area, flexible, and more importantly, low-cost energy source. In this introductory chapter the electro-

optical processes and current status of this type of devices is discussed. After

assessing critical loss mechanisms, several strategies for improving the efficiency

are discussed. Finally an overview of this thesis is given in which some of the

discussed strategies are investigated.

REFERENCES

R. Kroon, M. Lenes, J. C. Hummelen

P.W.M. Blom and B. de BoerPolymer Review, 2008, 3, 531.

7/31/2019 10 Complete


CHAPTER 1

10


1.1 Introduction: Solar Energy

When searching for an alternative source of energy, the vast amount of energy

(1.75 1017

J) the earth receives from the sun is guaranteed to draw attention. In

fact, with the total energy usage in 2003 being 4.4 1020

J per year, less then onehour is needed to fulfill this demand.

1Nevertheless, harnessing this source of

energy in a cost effective way is not an easy task. Several technologies can be

employed; first of all sunlight can be converted to thermal energy which can

subsequently be used for hot water, heating or conversion into electrical energy.

Alternatively, sunlight can be converted directly to electrical energy using the

photovoltaic effect. The field of photovoltaics is at the moment dominated by

inorganic, silicon based, solar cells. Due to the large availability of the used

material and extensive knowledge from the microelectronics industry, crystallinesilicon solar cells currently have a 90% market share.

2The main drawback of this

type of devices is the high purity needed for proper device operation. The energy,

and thus costs, needed in the fabrication process limits its usefulness as an

alternative energy source. Nevertheless silicon based solar cells are already almost

at par with consumer electric grid prices in southern Europe.3Next generation solar

cell technology is focusing on two directions; on one side high efficiency, or on the

other side a moderate efficiency combined with low cost.4

For high efficienciesmultilayer cells are investigated, often in combination with solar concentrators in

order to get the most out of these very efficient but also expensive solar cells, oftenbased on III-V semiconductors. For low cost cells, thin film technologies are

employed in order to reduce material usage, which can be either amorphous and

microcrystalline silicon as well as other inorganic compounds such as cadmium

telluride (CdTe) or copper indium gallium selenide (CIGS).

Organic solar cells are a relatively new route towards a large area, low cost

energy source.5

Organic materials can be solution processable allowing for low

cost deposition techniques such as spin coating, doctor blading, inkjet printing and

ultimately roll to roll fabrication.6

Due to the almost infinite modification of the

molecular structure it is possible to tune the chemical and physical properties of the

materials allowing for great flexibility in design. The high absorption coefficient of

organic materials allows organic solar cells to absorb most of the light in extremely

(~100 nm) thin layers reducing material usage significantly. On the down sideorganic materials are often highly disordered, limiting charge carrier transport and

device efficiency. Organic solar cells exist in roughly 3 types; dye-sensitized, small

molecule-, and polymer- based devices.The dye-sensitized solar cell was first introduced by O‟Reagan and Grätzel in

1991 and consists of a nanoporous titanium oxide (TiO2) layer.7

The TiO2 material

is covered with a Ruthenium dye which, after the absorption of light, injects an

electron into the TiO2 after which the electrons can be collected at an electrode. An

electrolyte regenerates the dye and is responsible for the hole transportation to a

7/31/2019 10 Complete


POLYMER:FULLERENE SOLAR CELLS

11


counter electrode. The main disadvantage of DSC‟s is the use of the liquid

electrolyte, which causes stability problems. Replacing the liquid electrolyte with asolid up to now results in lower efficiencies yet considerable progress is made in

this direction.8

Small molecule solar cells are fabricated by thermal evaporation of a donor and

acceptor material in either a double layer structure9

or a bulk heterojunction similarto polymer solar cells.

10Advantage of small molecule cells is the large control of

the deposition enabling for instance combinations of bilayer and bulk

heterojunctions. On the downside the vacuum based deposition does not comply

with the concept of a low cost and high throughput fabrication technique and is

difficult to be applied to large areas.

Polymer solar cells are based on π conjugated polymers as electron donors.Modification of the molecular structure allows one to tailor chemical and physical

properties and have resulted in a number of well performing materials with

different band gaps and energy levels such as P3HT,11

PCPDTBT,12

PF10TBT13

and

pBBTDPP2.14

As an acceptor either another semiconducting polymer,15,16

inorganic

materials17

or fullerenes18

can be used. Historically fullerenes have resulted in

superior device performances yet the other options still remain a viable alternative.

In this thesis solar cells based on a blend of a polymer and a fullerene areinvestigated.

7/31/2019 10 Complete


CHAPTER 1

12


1.2 Organic Semiconductors

The field of polymer electronics emerged in 1977 by the serendipous discoveryof a chemically doped polymer which exhibited a high conductivity,

19,20a

remarkable feature considering the traditional usage of polymers as insulating

material for metal components. While initial research focused on creating highly

conductive materials, attention was soon shifted towards the semiconducting

properties of conjugated materials. Conjugated materials, which can be either

polymers or small molecules, consist of an alternation of single and double carbon-

carbon bonds. Since each carbon atom is bound to only three neighboring atoms,one electron is left in a pz orbital. The mutual overlap between these pz orbitals

results in the formation of π bonds along the backbone. These delocalized πelectrons are the origin of the intrinsic semiconductor behavior of conjugated

materials. The charge transport in organic semiconductors is fundamentally

different from traditional inorganic physics where the concept of band conduction

and free charge carriers is used. Instead, carriers move through the material from

one localized state to another, a process called hopping.21

Due to the energetic and

spatial disorder in the materials this hopping process results in a relatively low

charge carrier mobility. Hence, organic semiconductor devices are not meant to

compete with high speed applications such as silicon computer chips but

applications should be found in combining ease of processing (and the associated

low cost) with moderate performances. Typical examples are field effect transistorsfor identification tags or sensors, light emitting diodes for large area lighting or

displays, memory devices and solar cells. Alternatively the option to process these

materials on plastic substrates offers opportunities for flexible devices such as

rolable displays.

Figure 1.1: Chemical structure of polyacelylene, the most general form of a conjugated polymer.

7/31/2019 10 Complete



13


1.3 Charge generation in organic solar cells

In contrast to inorganic photovoltaic devices, organic solar cells produce aneutral mobile excited state (exciton) after the absorption of light instead of free

charge carriers. Since the binding energy of this exciton is typically 0.2-0.8 eV,22

much higher than the thermal voltage, a single component organic solar cell will

result in extremely low efficiencies. In order to separate the excitons into free

charge carriers a donor-acceptor (D-A) system can be employed.9

When an exciton

reaches the donor/acceptor interface, the electron will transfer to the material with

the larger electron affinity and the hole will be accepted by the material with thelower ionization potential. Due to the low exciton diffusion lengths of typical 1 – 10

nm in polymeric materials23,24

a simple bilayer structure will result in low

efficiencies, since only photons absorbed within this distance from the

donor/acceptor interface will contribute to the device current.25

A drastic increase

in the generated photocurrent can be achieved by employing an interpenetrating

network of donor and acceptor materials.15,18

Ideally, in this so-called bulk

heterojunction (BHJ) all absorbed photons will be in the vicinity of a donor

acceptor interface and can contribute to the generated photocurrent.

Figure 1: Charge generation in a polymer:fullerene bulk heterojunction solar cell: a) absorption of aphoton resulting in an exciton, b) diffusion of the exciton towards the donor acceptor interface, c)electron transfer from donor to acceptor, d) dissociation of the bound electron hole pair into free

carriers, e) transport of free carriers towards the electrodes, f) collection at the electrodes. Loss

mechanisms are indicated by 1) non absorbed photons, 2) exciton decay, 3) geminate recombinationof the bound pair, 4) bimolecular recombination.

7/31/2019 10 Complete


CHAPTER 1

14


The complete process starting from an absorbed photon and ending up with

charges collected at the electrodes is depicted in Fig. 1.1: First a photon is absorbedby the donor material (a) after which an exciton is created. This exciton diffuses

towards a donor/acceptor interface (b) where the electron is transferred to the

acceptor material (c). Even though the hole and electron are now on different

materials they are still strongly bound by Coulomb interaction and need to bedissociated into free carriers (d) after which they are transported through the two

respective phases (e) and can be collected at the electrodes (f). During each of the

above-mentioned processes energy can be lost resulting in various loss

mechanisms. First of all, not all photons are absorbed by the active layer, not only

due to limitations of the bandgap but also due to the often limited thickness of the

active layer (1). Secondly, excitons will decay when created too far from the D-Ainterface (2). After electron transfer, geminate recombination of the bound electron

hole pair can occur (3) as well as bimolecular recombination (4) of free charge

carriers during transport to the electrodes.

Figure 1.2: Schematic layout of a polymer:fullerene bulk heterojunction solar cell.

7/31/2019 10 Complete



15


1.4 Device fabrication and characterization techniques

1.4.1 Solar cell fabrication

In figure 1.2 the basic structure of a solar cell fabricated in this thesis is shown.

We start with a glass substrate, prepatterned with a layer of indium tin oxide (ITO)

which acts as a transparent electrode. In order to reduce the roughness of the ITO athin layer of poly (3,4-ethylenedioxythiophene:polystyrenesulfonate (PEDOT:PSS)

is spincoated on top of the substrate. Besides reducing the roughness of the ITO the

PEDOT:PSS provides a proper (high) work function and allows for a good wettingof the active layer. After a short baking step to remove all the water from the

PEDOT:PSS layer, the substrates are moved into an inert atmosphere for further

processing. Now the blend of polymer and fullerene is spincoated on top of thesubstrate. The devices are completed by thermal evaporation of a thin low work

function material (either samarium or lithium-fluoride) and a 100 nm thick

aluminum top contact. The top contacts are deposited through a shadow mask in

order to define the active layer. After completion of the devices they are transferredin a nitrogen filled container to the measurement setup which is also under inert

atmosphere.

1.4.2 Solar cell Characterization

The solar cell performance and electrical characteristics are determined by

measuring the current density to voltage ( J-V ) characteristics, both in dark andunder illumination. Figure 1.3 shows the typical J-V characteristics of a polymer:

fullerene solar cell under illumination. From the J-V curve four parameters can be

deduced. The current density under illumination at zero applied bias is called the

short circuit current density ( J sc), when the current density under illumination is

zero the cell is at the open circuit voltage (V oc) and the fill factor (FF ) relates the

maximum power the cell can deliver, to the open circuit voltage and short circuitcurrent density:

ocsc

mppmpp

V J V J FF

The power conversion efficiency is now determined by dividing the maximum

power point (Pmpp) by the incoming light power Pin:

in

ocsc

in

mpp

P

FF V J

P

P

7/31/2019 10 Complete


CHAPTER 1

16


-0.2 0.0 0.2 0.4 0.6-120

-80

-40

0

40

80

120

160

MPP

JMPP

VMPP

JSC

Under Illumination

In Dark

J [ A / m 2 ]

V [V]

VOC

FF =

Figure 1.3: Typical current-voltage characteristics of a polymer:fullerene solar cell showing the V oc,,

J SC , FF and maximum power point ( MPP)

The power conversion efficiency has to be determined under standard test

conditions (STC) which includes the temperature of the solar cell (25 C), anillumination intensity of 1000 W/m

2and a spectral distribution of the illumination

source (air mass 1.5 or AM1.5).26

Since the spectrum of the used illumination

source is in general not the same as the AM1.5 solar spectrum the mismatch factor

( M ) for the measurement has to be determined using

27

„

where E R(λ) and E S(λ) are the AM1.5 solar spectrum and spectrum of the used

illumination source and S R(λ) and ST (λ) are the spectral responses of a reference cell

and the tested cell, respectively. For this purpose we use a silicon solar cell of

which the spectral response is determined at the Energy Centrum of the

Netherlands (ECN). In order to determine the spectral response of the tested cell

Incident photon-to-current efficiency (IPCE) measurements are performed. In thismeasurement the cell under test is illuminated by monochromatic light and the ratio

between the generated photocurrent (at short circuit conditions) and the incident

photon flux at that particular wavelength is determined (for the whole range of the

solar spectrum). Figure 1.4 shows an example of such and IPCE, also known as

External Quantum Efficiency (EQE) measurement, which besides determining the

mismatch factor of the measurement, is also very useful for analyzing loss

mechanisms in solar cells.

)()(

)()(

)()(

)()(

T R

T S

RS

R R

S E

S E

S E

S E M

7/31/2019 10 Complete



17


400 500 600 700 800 900 10000

10

20

30

40

50

60

70

80

90

100

E . Q . E .

[ % ]

Wavelength [nm]

Figure 1.4: External Quantum Efficiency (EQE), also known as Incident Photon-to-Current

Efficiency (IPCE) of a polymer:fullerene solar cell, measured at short circuit conditions.

1.4.3 Characterization techniques

Above it is shown that the efficiency of a solar cell is represented by its maximum

power point ( MPP). By itself the MPP does not contain much information on theworking of a solar cell, but it can be expressed as the product of short circuit

current ( J sc), open circuit voltage (V oc) and fill factor (FF ). Very generally, the V oc

is governed by the HOMO of the donor and the LUMO of the acceptor,28,29

the Jsc

depends on the photon absorption of the active layer and the FF is determined bythe (balanced) charge transport and recombination properties of the materials.

30In

reality these guidelines will only apply for optimized devices and even then are

only first approximations. Below some considerations are given when analyzing

the J-V curves of a solar cell and which experimental techniques can be used to

asses the potential of a solar cell.

The V oc of a polymer fullerene solar cell can be described by the following

relationship:31

in which q is the elementary charge, P the dissociation probability of a boundelectron – hole pair into free charge carriers, G M the generation rate of bound

electron-hole pairs, γ the Langevin recombination constant, N c the effective density

of states, k the Boltzmann constant, and T is the temperature. However, this

relation is only valid when the electrodes form ohmic contact with the HOMO of

the donor and the LUMO of the acceptor. If this is not the case the V oc will be

M

c

ocPG

N P

q

kT --V

2)1(lnLUMO(A)HOMO(D)

7/31/2019 10 Complete


CHAPTER 1

18


limited to the difference in workfunction of the electrodes. This is often observed

when aluminium is used as cathode without a lithium fluoride (LiF) or other lowworkfunction interlayer, reducing the voltage significantly. On the other hand an

ill-defined cathode interface can give rise to a distinct S shape of the J-V curve

around the open circuit voltage resulting in low fill factor.32

Furthermore, a low

amount of photogenerated charges as well as recombination with charge traps, willresult in a lower V oc.

33

The current generated by a solar cell is ultimately governed by the amount of

absorbed photons. Because of the low exciton diffusion lengths in the donors23,24

a

bulk heterojunction is employed to harvest all the excitons.15,18

The domain size of

donor and acceptor thus plays a very important role in the actual short circuit

current measured in a device. In fact, the control of this morphology is the mostdifficult and most investigated part of the solar cell fabrication. Typically a large

range of solvents, polymer:fullerene ratios, annealing effects and additives are

required to induce the correct morphology. When domain sizes are too large,

excitons will be lost due to exciton decay. Photophysical studies can be employed

to see whether all excitons are able to reach an interface. However, too small

domain sizes can induce an enhanced recombination of charge carriers. Also the

donor and acceptor domains need to have a percolated pathway towards anode andcathode, respectively, in order for charges to be collected. A range of morphology

imaging tools including transmission electron microscopy (TEM), selected area

electron diffraction (SAED), scanning electron microscopy (SEM), scanning probe

microscopy (SPM) and atomic force microscopy (AFM) can be used for

characterization of the active layer morphology.34 Even when all of the generated excitons reach an interface this does not

automatically imply that all charges are actually converted into free charge carriers.Due to the low dielectric constant of the polymer and fullerene, the electron and

hole are coulombically bound at the interface and need to be dissociated into free

carriers by an electric field. Plotting the photocurrent as a function of effective field

can be used to determine the dissociation efficiency of a device.35,36

For

MDMO:PPV:PCBM blends the dissociation efficiency was shown to be only 60%

at short circuit conditions.35

Spectroscopic evidence for the occurrence of the (field

dependent) dissociation of a bound state at the donor-acceptor interface has been

found in both polymer:polymer and polymer:fullerene blends.37

The fill factor of a device depends in a complicated way on charge

dissociation, charge carrier transport and recombination processes.30

A good hole

and electron transport capability is of vital importance for proper device operation.

When hole and electron transport are unbalanced a build up of space-charge results

in a square root dependence of the photocurrent on voltage, resulting in low fill

factors. Even a difference in hole and electron mobility of only one order of

magnitude can influence the device performance, which imposes limitations on the

active layer thickness in order to avoid space-charge problems. Light intensity

dependent measurements can provide information on which type of recombination,geminate or bimolecular, is dominant and whether space-charge problems play a

7/31/2019 10 Complete



19


role. Several type of experimental test beds can be employed to determine the

charge carrier mobilities in the materials such as field-effect transistor (FET),space-charge limited current (SCLC) measurements, photo induced charge carrier

extraction in a linearly increasing voltage (photo-CELIV), or time-of-flight

measurements (TOF). For solar cells, field-effect transistor measurements are less

useful due to the much higher charge carrier density in these types of devices,which strongly influences the mobility.

38Due to the large variety of measurement

techniques the comparison of values reported is often difficult. Furthermore, it is of

importance to determine the charge carrier properties in the polymer:fullerene BHJ

as it is fabricated in the actual solar cell, since blending a polymer with a fullerene

can have very different effects on the charge carrier properties. For instance, in

MDMO:PPV a 200 fold increase of the hole mobility is observed when blended ina 1:4 weight ratio with PCBM whereas for P3HT a decrease in mobility is observed

upon blending, only to be recovered by thermal or solvent annealing. In this thesis,

besides the standard J-V measurements, the determination of the charge carrier

mobilities, and modelling of the photocurrent are used as main tools for analysing

polymer:fullerene solar cells. These characterisation techniques will be discussed

in further detail below.

1.1.1 Single carrier devices

As explained above, in order to understand the device operation of polymer

solar cells it is crucial to determine the charge carrier mobility of the individual

components of the solar cell. In this thesis charge carrier mobilities are determinedusing the SCLC method. The advantages of this method are that the charge carrier

mobilities are determined in an identical geometry as the solar cell itself, and thatprocessing conditions are almost identical. Furthermore, both pristine materials as

well as the blends actually used in the solar cell can be investigated. Blending a

polymer with a fullerene can have very different effects on the charge carrier

properties.

In order to measure the charge carrier mobility of holes or electrons the

transport of the other charge carrier needs to be suppressed. This is done by

choosing appropriate electrode materials. One can either fabricate a double carrier

device similar to the solar cell, make a hole-only device by suppressing the electron

injection using a palladium (Pd) cathode, or suppress the hole injection by using an

aluminum oxide anode and thus make an electron-only device. Since the light

emission from solar cells is very low it is difficult to determine whether only one

carrier is injected. From experiments on the pure materials it is found that injection

from the “wrong” contact can be ruled out. Furthermore no difference in extractioncapabilities was observed. In its most basic form the SCLC is now given by

3

2

int0

8

9

L

V J r

7/31/2019 10 Complete


CHAPTER 1

20


where 0 r is the dielectric constant of the material, L is the device layer thickness,V int is the internal field of the device (applied voltage V A corrected for the built in

voltage V bi and voltage drop over the connections V rs) and is the mobility of the

material. Previous studies have shown the charge carrier mobility in organic

materials to be dependent on both the electric field and the charge carrier density in

the device. In this thesis only the field dependence of the charge mobility is taken

into account using a stretched exponential dependence:

))(exp()(),( 0 E T T T E

where 0(T) is the zero field mobility and (T) describes the field activation.

Justification for neglecting the density dependence are the relatively low chargecarrier concentrations during device operation and the relatively low amount of

disorder in the used materials compared to for instance organic light-emitting

diodes.

1.4.4 Photocurrent measurements and modeling

In section 1.3.1 the basic characterization of an organic solar cell wasintroduced. From the fourth quadrant of the J-V measurement the characteristic

values J sc, V oc, FF and the maximum power point can be determined. For analyzingthe physical processes inside the solar cell however this representation is not very

adequate. Much more information can be obtained by plotting the photocurrent as a

function of the electric field inside the device.35

For this the photocurrent

J ph= J L− J D, where J L and J D are the current density under illumination and in dark,respectively, is plotted as a function of effective applied voltage V 0-V A. Here V 0 is

the compensation voltage defined as J ph (V 0) =0 and V A is the applied bias. Figure

1.5 shows a typical example of the photocurrent of a polymer:fullerene solar cell.

Two different regimes can be identified; for small effective voltages thephotocurrent increases linearly, which has been shown to be caused by a

competition between drift and diffusion currents.39

At higher voltages the

photocurrent gradually increases until it saturates to a maximum value. This

gradual increase has been attributed to the field- and temperature dependentdissociation of bound electron hole pairs at the donor-acceptor interface. The Braun

Onsager model of geminate recombination has been shown to adequately describe

this dissociation of bound electron hole pairs in a number of polymer:fullerenesystems. A typical fit of the photocurrent using this model is also shown in figure

1.5 showing the nice agreement of the model at high voltages. In order to fully

describe the photocurrent and the standard J-V characteristics, a numerical program

developed by Jan Anton Koster is used.36

This program, which includes drift and

diffusion of charge carriers, the effect of space-charge on the electric field,

7/31/2019 10 Complete



21


bimolecular recombination, and a field- and temperature dependent generation

rates of free charge carriers, has been shown to consistently describe a number of polymer:fullerene systems, including their temperature and illumination

dependence. A fit of the complete voltage regime of the photocurrent using the

simulation program is shown in figure 1.5.

Besides the increase of the photocurrent at reverse bias, spectroscopicexperiments have indicated the existence of a charge transfer state at the donor-

acceptor interface. There exists however a discrepancy between the lifetime of this

state, as determined from the Braun model used here (micro to milli-seconds), and

the spectroscopically lifetime (nanoseconds). Recently is was suggested to take

into account the local mobility of charge carriers at the interface, instead of

macroscopic determined mobilities when using the Braun model. Decay ratesdetermined in this way do agree with spectroscopic decay rates.

0.1 1 10

10

100

Data

Simulations

Braun model

J

p h

[ A / m 2 ]

V0-V [V]

Figure 1.5: Photocurrent density Jph as a function of effective applied voltage V0-V. The symbolsrepresent the experimental data, the dotted line a fit of the photocurrent at high reverse bias using the

Braun model and the solid line a fit over the whole voltage regime using the numerical simulationprogram.

7/31/2019 10 Complete


CHAPTER 1

22


1.5 Optimization of energy levels in a donor acceptor system

Above we have discussed the use of a donor-acceptor (D-A) system in order toseparate excitons into free carriers. Unfortunately, during the transfer of the

electron from the LUMO of the donor to the LUMO of the acceptor, energy is

inevitably lost. This loss in energy is manifested in the low open circuit voltage of

a D-A BHJ solar cell compared to the bandgap of the absorber. The open circuit

voltage is ultimately limited by the difference between the HOMO of the donor and

the LUMO of the acceptor. This means that the energy offset between donor and

acceptor LUMO enables electron transfer but also, inevitably, results in a loss of V oc.

Figure 1.3: Energy diagram of a P3HT:PCBM solar cell (a) and strategies to reduce the loss of energy

during electron transfer by (b) reducing the LUMO of the donor (c) reducing the LUMO and HOMO

of the donor and d) raising the LUMO of the acceptor.

In Figure 1.3 the energy diagram for the P3HT:PCBM system is shown. What

is striking is the LUMO-LUMO offset which is much larger than the 0.3 – 0.5 eV

necessary for electron transfer to occur. This results in P3HT:PCBM cells havingan open circuit voltage of only 0.6 V, much smaller compared to the bandgap of

P3HT of 2 eV. It is the reduction of this excess LUMO-LUMO offset where a large

increase in device efficiency can be obtained.1 40

What is assumed here is that reducing the LUMO-LUMO offset does not reduce the

charge separation at the donor-acceptor interface. Recent experiments have hinted at thefact that the large offset in P3HT:PCBM is necessary for the efficient generation of free

charge carriers.

7/31/2019 10 Complete



23


To reduce the offset three strategies can be employed. Firstly, the LUMO of the

donor can be lowered resulting in so called small (or sometimes called low)bandgap donors. By decreasing the bandgap of the donor the absorption of the

donor material is extended towards higher wavelengths. In figure 1.6 the solar

spectrum is shown together with the integrated photon flux versus wavelength.

Since the absorption of P3HT is limited to 650 nm it is clear that a reduction in thedonor bandgap can result in a large increase of the amount of absorbed photons and

hence device current.

Figure 1.6: Photon flux as function of wavelength. The percentage of the total photon flux and thecorresponding maximum current is displayed at the x-axis.

Alternatively, both LUMO and HOMO level of the donor can be lowered. In

this case the bandgap of the donor remains constant and the device gains in

efficiency due to a larger V oc. An advantage of this strategy is the expected air

stability once the HOMO of the polymer is lower then 5.4eV.

41

One has to take intoaccount however, that for a very deep HOMO the workfunction of PEDOT:PSS

might not be adequate anymore, limiting the V oc.

As a third option, the LUMO-LUMO offset can be reduced by raising the

LUMO of the acceptor, were again the device gains in efficiency due to an

enhancement of the open circuit voltage. All these three strategies are illustrated in

Figure 1.5.

Which of the above-mentioned strategies, and thus which donor bandgap, is

optimal is still under debate and depends on the models used to predict efficiencies

and the restrictions made on the materials used. When calculating the increase in

7/31/2019 10 Complete


CHAPTER 1

24


current generated by narrowing the bandgap of the donor for instance, one has to

take into account that not all photons with larger energy than the donor bandgap areabsorbed. Koster et al have calculated the increase in absorption by taking the

absorption profile of P3HT and shifting this absorption in energy to account for a

narrowing of the bandgap. Combined with realistic values for the fill factor and

charge dissociation, Koster et al predict an efficiency of 6.6% for a donor bandgapof 1.5 eV at which point the LUMO-LUMO offset is reduced to 0.5 eV (in

combination with [60]PCBM as the acceptor). Further narrowing of the bandgap

will have to be realized by raising the HOMO of the polymer which will result in a

lower V oc and no increase in efficiency is expected. Note, however, that very small

bandgaps may still have their use in infrared photo detectors and tandem or multi-

junction solar cells. As a second step Koster et al calculated the increase inefficiency when P3HT is taken as donor and the LUMO of the acceptor is raised up

to the 0.5 eV offset. For this strategy the predicted maximum efficiency was found

to be more than 8%, showing the great potential of energy level alignment at the

acceptor side. If one now allows both donor and acceptor LUMO to vary the

optimal bandgap can be determined. It was shown that this optimal bandgap is in

fact not small, as is usually asumed, but reaches a maximum around 1.9 eV.

1.5.1 Multijunction solar cells

In the discussion above only single layer cells are considered. When this

constraint is lifted even higher efficiencies can be obtained as is demonstrated inthe field of inorganic photovoltaics. Here multijunction solar cells have shown

efficiencies of over 40 percent,42

higher than the theoretic limit for single junctions

of 31 percent.43

For single junction solar cells there exists an optimal bandgap dueto a competition between a high voltage and a high current. When enlarging the

bandgap the voltage is raised yet the amount of photons which can be absorbed

decreases and vica versa. Multiple junctions enable photons with high energy to be

absorbed by a wide bandgap material and photons with low energy by a small

bandgap material. In this way the thermalisation losses, due to photons having a

larger energy compared to the bandgap of the absorber, can be diminished.

For polymer solar cells the realization of a tandem or multijunction

architecture is difficult to achieve, especially concerning the middle electrode(s).44

Since polymer solar cells are made of intrinsic semiconductors the work function

of this electrode should match the energy level of the acceptor on one side and that

of the donor on the other. Furthermore, the electrode should function as a good

recombination layer, needs to be optically transparent, and last but not least has to

protect each active layer from the deposition of the next one. Neverthelessconsiderable progress is made in the fabrication and the design of tandem and

multijunction cells. The fact that more well-performing materials, with proper

energy level alignments, are needed for both single and mulitjunction cells, is

reflected in the best performing polymer multijunction cell so far. In this tandem

7/31/2019 10 Complete



25


cell the sub cell which absorbs the low wavelength part of the spectrum generates a

lower open circuit voltage compared to the cell which absorbs the high wavelength,exactly opposite to the general design rule for multijunction cells.

45

7/31/2019 10 Complete


CHAPTER 1

26


1.6 Aim and scope of this thesis

In the last decade the field of polymer:fullerene solar cells has seen an

impressive improvement of both device efficiency and understanding of the

physical processes governing these type of devices. The switch to chlorobenzene as

a solvent for spincoating MDMO-PPV:PCBM layers led to the first reasonable

performing device, achieving an efficiency of 2.5%. Due to the relatively largeavailability of the used materials there was now a model system for the community

which made a thorough investigation of the device physics possible. The MDMO-

PPV:PCBM system turned out to be a rather peculiar system in which the addition

of a fullerene to the polymer results in an increase of the hole mobility of the

polymer by a factor 200, which turned out to be essential for device operation and

explains the high amount of fullerene used in the optimized cell. Furthermore, thedissociation of the bound electron hole pair at the donor acceptor interface was

shown to be a significant limiting factor for PPV based devices.

Devices based on MDMO-PPV:PCBM blends typically have an active layer

thickness of 100 nm at which still a significant portion of light is not absorbed. Inchapter 2 of this thesis the origin of the need for such a relatively thin active layer

is investigated. It is shown that the decrease in fill factor which, from a device

point of view, is the origin for the decreasing efficiency upon increasing the active

layer thickness, is due to a combination of space-charge effects, a decreasing

dissociation efficiency and charge recombination.In chapter 3 a new glycol substituted PPV is investigated which has a higher

permittivity compared to normal PPVs. The aim here is to increase the abovementioned low dissociation efficiency of PPV based devices, which is strongly

dependent on the average permittivity of the active layer blend. Due to a significant

lower hole mobility of the polymer and morphology problems devices based on

this new polymer did not show improved power conversion efficiencies compared

to the model MDMO-PPV system. Nevertheless, an increase in dissociation

efficiency from 60 to 72% was observed for the enhanced permittivity polymerindicating the importance of the average permittivity in polymer:fullerene devices.

After the initial success of MDMO-PPV, poly(3 -hexylthiophene) emerged as a

new model material for polymer:fullerene devices. Thermal and solvent annealing

steps improved the charge mobility of the hole transport considerably, resulting ina balanced transport of both holes and electrons. Due to this balanced transport

much thicker active layers can be used without sacrifice of the fill factor, resulting

in a significantly enhanced device performance. Again, wide (commercial)

availability of the polymer allowed extensive research into the device physics and

optimization of the material system. With internal quantum efficiencies as high as

90% it is clear new materials are needed in order to further increase the device

efficiency. Using the knowledge from the MDMO-PPV and P3HT system a more

focused approach can now be taken when searching for new systems, especially

concerning the energy level optimization as discussed in the previous section.

7/31/2019 10 Complete



27


One of the approaches towards higher efficiencies is reducing the bandgap of

the donor polymer in order to increase the light harvesting of polymer:fullerenesolar cells. One of the most promising materials following this route is PCPDTBT.

In chapter 4 the charge transport and photogeneration of this material blended with

PCBM is investigated. Despite an almost balanced transport the photocurrent

shows a square root dependence on effective voltage. It is shown that this squareroot dependence does not stem from an unbalance in mobilities, as seen in other

polymer solar cells, but from an enhanced recombination of the bound electron

hole pair. This enhanced recombination is likely due to a too close intermixing of

polymer and fullerene.

In chapter 5 another route towards efficiency enhancement is investigated.

Instead of lowering the LUMO of the donor now the LUMO of the acceptor israised allowing for a very direct enhancement of the efficiency due to a larger open

circuit voltage. In order to achieve this raising of the LUMO level, the bisadduct

analog of PCBM was used. The additional functionalisation of the fullerene cage

leads to the saturation of one more double bond raising the LUMO level of the

molecule significantly. It is shown that despite the additional functionalisation and

the increased disorder introduced (due to having a multitude of isomers of the

molecule), replacing PCBM with bisPCBM results in only a very slightlydecreased photogeneration and transport properties. Combined with a significantly

enhanced open circuit voltage a power conversion efficiency of 4.5% was

achieved, which is among the highest reported for polymer:fullerene solar cells. In

the second part of chapter 5, bis- and trisadduct analogs of other fullerenes are

investigated. It is shown that the existence of multiple isomers leads to shallowtrapping for single carriers devices which do not affect the device operation of the

solar cells itself.

7/31/2019 10 Complete


CHAPTER 1

28


REFERENCES

[1] N. S. Lewis, MRS Bull. 2007, 32, 808.\

[2] A. Slaoui, R. T. Collins, MRS Bull. 2007, 32, 211.

[3] M. Šúri, T. A. Huld, E. D. Dunlop, H. A. Ossenbrink, Solar Energy, 2007, 81,1295.

[4] M. A. Green, Physica E: Low-dimensional Systems and Nanostructures, 2002 14,65.

[5] C. J. Brabec, N. S. Sariciftci, J. C. Hummelen, Adv. Funct. Mater. 2001, 11, 15.

[6] C. J. Brabec, J.D Durrant, MRS Bul. 2008, 33, 607.

[7] B.O‟reagan, M. Gratzel, Nature, 1991, 353, 737.

[8] U. Bach, D. Lupo, P. Comte, J. E. Moser, F. Weissörtel, J. Salbeck, H. Spreitzer,M. Grätzel, Nature 1998, 395, 583.

[9] C. W. Tang, Appl. Phys. Lett . 1986, 48, 183

[10] J. Xue, B. P. Rand, S. Uchida, S. R. Forrest, Adv. Mater . 2005, 17 , 66

[11] F. Padinger, R. S. Rittberger, N. S. Sariciftci, Adv. Funct . Mater . 2003, 13, 85.

[12] D. Muhlbacher, M. Scharber, M. Morana, Z. Zhu, D. Waller, R. Gaudiana, C.Brabec, Adv. Mater. 2006, 18, 2884.

[13] L. H. Slooff, S. C. Veenstra, J. M. Kroon, D. J. D. Moet, J. S. Sweelssen, M. M.Koetse, Appl. Phys. Lett. 2007, 90, 143506.

[14] M. M. Wienk, M. Turbiez, J. Gilot, R. A. J. Janssen, Adv. Mater . 2008 20, 2556

[15] J. J. M. Halls, C. A. Walsh, N. C. Greenham, E. A. Marseglia, R. H. Friend, S. C.Moratti, A. B. Holmes, Nature 1995, 376, 498.

[16] C. R. McNeill, A. Abrusci, J. Zaumseil, R. Wilson, M. J. McKiernan, J. H.Burroughes, J. J. M. Halls, N. C. Greenham, R. H. Friend, Appl. Phys. Lett . 2007,

90, 193506.

[17] D. J. D. Moet, L. J. A. Koster, B. de Boer, P. W. M. Blom, Chemistry of Materials,2007, 19, 5856.

7/31/2019 10 Complete



29


[18] G. Yu, J. Gao, J. C. Hummelen, F. Wudl, A. J. Heeger, Science 1995, 270, 1789.

[19] C. K. Chiang, C. R. Fisher, Y. W. Park, A. J. Heeger, H. Shirakawa, E. J. Louis, S.

C. Gau, A. G. MacDiarmid, Phys. Rev. Lett . 1977, 39, 1098.

[20] H. Shirakawa, E. J. Louis, A. G. MacDiarmid, C. K. Chiang, A. J. Heeger J. Chem.Soc. Chem. Commun. 1977, 578.

[21] D. Hertel, H. Bassler, ChemPhysChem 2008, 9, 666.

[22] V.I. Arkhipov, H. Bassler, Phys. Status Solidi A 2004, 201, 1152

[23] D. E. Markov, E. Amsterdam,P. W. M. Blom, A. B. Sieval, J. C. Hummelen, J.

Phys. Chem. A 2005, 109, 5266.

[24] D. E. Markov, C Tanase, P. W. M. Blom, J Wildeman, Phys. Rev. B. 2005, 72,

045217.

[25] M. Granstrom, K. Petritsch, A. C. Arias, A. Lux, M. R. Andersson, R. H. Friend, Nature 1998, 395, 257.

[26] IEC-904-3, IEC Standard

[27] J. M. Kroon, M. M. Wienk, W. J. H. Verhees, J. C. Hummelen, Thin Solid Films 2002, 403, 223.

[28] C. J. Brabec, C. Winder, N. S. Sariciftci, J. C. Hummelen, A. Dhanabalan, P. A.van Hal, and R. A. J. Janssen, Adv. Funct. Mater. 2002, 12, 709.

[29] L. J. A. Koster, V. D. Mihailetchi, R. Ramaker, and P. W. M. Blom, Appl. Phys.

Lett. 2005, 86 , 123509.

[30] P. W. M. Blom, V. D. Mihailetchi, L. J. A. Koster, D. E. Markov, Adv. Mater. 2007, 19, 1551.

[31] L. J. A. Koster, V. D. Mihailetchi, R. Ramaker, P. W. M. Blom, P.W.M. Appl.Phys. Lett. 2005, 86, 123509.

[32] D. Gupta, M Bag, K. S Narayana, Appl. Phys. Lett. 2008, 92, 093301.

[33] M. M. Mandoc, F. B. Kooistra, J. C. Hummelen, B. de Boer, P. W. M. Blom, Appl.

Phys. Lett. 2007, 91, 263505.

[34] X Yang, J Loos, Macromolecules 2007, 40, 1353-1362.

7/31/2019 10 Complete


CHAPTER 1

30


[35] V. D. Mihailetchi, L. J. A. Koster, J. C. Hummelen, P. W.M. Blom, Phys. Rev.

Lett. 2004, 93, 216601.

[36] L. J. A Koster, E. C. P. Smits, V. D. Mihailetchi, P. W. M. Blom, Phys Rev. B 2005, 72, 085205.

[37] D Veldman, O Ipek, S. C. J. Meskers, J. Sweelssen, M.M. Koetse, S.C. Veenstra,

J.M. Kroon, S.S. van Bavel, J. Loos, R.A.J. Janssen, . J.A.C.S., 2008, 130, 7721

[38] C. Tanase, E. J. Meijer, P. W. M. Blom, D. M. de Leeuw , Phys. Rev. Lett, 2003

,91, 216601.

[39] R. Sokel, R. C. Hughes, J. Appl. Phys. 1982, 53, 7414.

[40] T. M. Clarke, A. M. Ballantyne, J. Nelson, D. D. C. Bradley, J. R. Durrant, Adv.

Mater. 2008, 42, 4029.

[41] J. Locklin, M. M. Ling, A. Sung, M. E. Roberts, and Z. Bao: Adv. Mater . 2006, 18

2989.

[42] M. A. Green, K. Emery, Y. Hishikawa, W. Warta, Prog. Photovolt: Res. Appl. 2008, 16 , 61.

[43] W. Shockley, H. J. Queisser, J. Appl. Phys. 1961, 32, 510.

[44] A. Hadipour, B. de Boer, P. W. M. Blom, Adv. Funct. Mater. 2008, 18, 169.

[45] J. Y. Kim, K. Lee, N. E. Coates, D. Moses, T.-Q. Nguyen, M. Dante, A. J. Heeger,

Science 2007, 317, 222.

7/31/2019 10 Complete



Chapter 2

Increasing the active layer thickness of

polymer:fullerene solar cells

Abstract

Solar cells based on poly[2-methoxy-5-(3',7'-dimethyloctyloxy)-1,4-

phenylenevinylene] (MDMO-PPV) as electron donor and [6,6]-phenyl C 61 butyric

acid methyl ester (PCBM) as electron acceptor have for many years been the

working horse of the field of polymer photovoltaics. A striking feature of these

solar cells is that at a device thickness of 100 nm at the polymers absorption

maximum only 60% of the incident light is absorbed. From a light harvesting point

of view, an increase in the active layer thickness is thus expected to significantly

enhance the generated photocurrent and hence efficiency. Experimentally however,

the efficiency lowers when using active layers beyond 100 nm, due to a decrease in

fill factor. This decrease has been attributed to an increasing series resistance,

although its physical meaning is not clear for solar cells where charge carriers are

generated throughout the device. In this chapter the origin of this decreasing fill

factor is investigated. It is demonstrated that the formation of space-charge, and

charge recombination puts a limit to the active layer thickness. At the end of the

chapter the effect of optical interference effects on the results are discussed.

REFERENCES

M. Lenes, L.J.A. Koster

V.D. Mihailetchi and P.W.M. Blom

Applied Physics Letters, 88, 243502 (2006)

7/31/2019 10 Complete


CHAPTER 2

32


2.1 Introduction

Although outperformed by poly(3-hexylthiophene) (P3HT) today, bulk

heterojunction solar cells based on poly[2-methoxy-5-(3',7'-dimethyloctyloxy)-1,4-

phenylenevinylene] (MDMO-PPV) as electron donor and [6,6]-phenyl-C61-butyric

acid methyl ester (PCBM) as electron acceptor have been one of the most studied

systems in this field.1

Typically these devices have an active layer of around 100nm achieving power conversion efficiencies of up to 2.5% under AM1.5

illumination.2

At this active layer thickness at the polymers absorption maximum

only 60 percent of the incoming light is absorbed. It is evident that increasing the

active layer thickness will result in an increased absorption in the device

accommodating larger photocurrents. In spite of this increased absorption it is

found that upon increasing the active layer thickness beyond typically 100 nm theoverall power conversion efficiency does not increase, mainly due to a decreasing

fill factor. This decreasing fill factor has been attributed to an increasing series

resistance, although its physical meaning for solar cells, where charge carriers are

generated throughout the device is not clear.3,4

Furthermore it is expected thatcharge recombination will play an important role in thicker devices since the

charge carriers need to travel a larger distance to be collected at the contacts.3,5,6

Previous work by Mihailetchi et. al.7

has shown that a large unbalance in

charge transport in donor and acceptor leads to space-charge effects in polymer

bulk heterojunction solar cells. In their work a difference between hole andelectron mobility of 2 to 3 orders lead to a completely space-charge dominated

photocurrent resulting in fill factors of only 42%. For MDMO-PVV:PCBM devicesalso a difference in hole and mobilities is observed, but here the difference is only

one order of magnitude.8

Since the fill factor of a 100 nm MDMO-PPV:PCBM

device is typically 60% space-charge effects do not play a role here. The main

question is now whether this still holds for active layer thicknesses beyond 100 nm.

7/31/2019 10 Complete


POLYMER FULLERENES SOLAR CELLS

33


2.2 Space-charge limited photocurrents

Since the hole mobility in polymers is in general smaller than the electronmobility in fullerenes under illumination, holes will accumulate in the solar cell.

This will in turn result in a change of the electric field inside the device. In the

region near the anode the electric field will be increased, enhancing the extraction

of holes. It has been shown that at sufficiently high intensities and mobility

differences a space-charge limited regime governs the photocurrent described by

the following equation7,9

V qG J hr ph

25.0

0

75.0

8

9)( (2.1)

where G is the generation rate of free carriers and μ is the mobility of the slowest

carrier, holes in this case. As can be seen from equation 2.1 a space-charge limited

photocurrent is characterized by a square-root dependence on voltage and a three

quarter dependence on generation rate and thus intensity. Because the photocurrent

depends on the square-root of the applied voltage a purely space-charge limited

device will have a maximum fill factor of 42%, which is considerably lower thanthe 60% for standard MDMO-PPV:PCBM devices. Since the mobility difference in

MDMO-PPV:PCBM devices is of only one order of magnitude, space-chargeeffects do not play a role in standard devices. Here the photocurrent density at short

circuit and reverse bias is governed by10,11

LT E qG J ph ),((2.2)

where q is the elementary charge, G(E,T) is the field,- and generation dependent

rate of free carriers and L the active layer thickness. Therefore, increasing the

active layer thickness will in general result in a higher photocurrent due to the

increase in absorption. Since the space-charge limited photocurrent is independent

on device thickness it is expected that upon increasing the active layer thickness at

some point a transition will occur from a non space-charge limited to a space-charge limited device as illustrated in figure 2.1. Here, for the sake of simplicity, aconstant generation rate G is assumed. Two regimes in the photocurrent can be

identified; regime I where drift- and diffusion currents compete and which varies

linearly with voltage, and regime II where the photocurrent saturates. If now the

active layer thickness is increased the photocurrent will increase and at some point

will intersect the, thickness independent, space-charge limit. Now a new regime

appears in the photocurrent, characterized by a square-root dependence on voltage.

Therefore, this simplified picture predicts a strong decrease in fill factor whenincreasing the active layer thickness, even when charge recombination does not

7/31/2019 10 Complete


CHAPTER 2

34


play a role. At which point this occurs and whether the field dependence of the

generation rate and charge recombination plays a role has to be determined bymodelling of the actual photocurrents.

Figure 2.1: Illustration of the transition from a non space-charge limited (upper figure) to a space-charge limited (lower figure) device by increasing the device thickness. Since the space-charge limit

is independent on device thickness and increasing the active layer will in general result in a higherphotocurrent at some point the photocurrent will intersect the space-charge limit resulting in a square-root dependence on voltage (regime II in the lower figure). When the device is below the space-

charge limit only two regimes can be distinguished, regime I where drift- and diffusion currentscompete and which varies linearly with voltage and regime II where the photocurrent saturates. Fielddependent motilities, field dependent generation rate and charge recombination are ignored in thissimplified picture.

Jph

V

P h o

t o c u r r e n t

V0-V

Jph

=qGL

S p a c

e C h a

r g e L i m i

t

Jph

V2

Jph

V

Jph

=qGL

V0-V

P h o t o c u r r e n t

7/31/2019 10 Complete



35


2.3 Device Fabrication and measurements

Solar cells were made using the standard fabrication technique from a blend of

MDMO:PPV and PCBM in a 1:4 weight ratio spin cast from chlorobenzene. By

varying the concentration of the blend solution and spin procedure samples were

made with active layer thicknesses of 128 to 368 nm. After fabrication the devices

were measured in a nitrogen atmosphere under illumination of a white lighthalogen lamp calibrated by a silicon diode. To obtain light intensity dependent

measurements a set of neutral density filter was used, yielding an intensity

variation of two orders of magnitude

100 150 200 250 300 350 400

0.74

0.75

0.76

0.77

0.78

0.79

0.80

0.81

100 150 200 250 300 350 400

40

45

50

55

60

65

100 150 200 250 300 350 4001.5

1.6

1.7

1.8

1.9

2.0

2.1

2.2

2.3

2.4

100 150 200 250 300 350 400

40

45

50

55

60

O p e n C i r c u i t V o l t a g e [ V ]

Active Layer Thickness (nm)

F i l l F a c t o r [ % ]


E f f i c i e n c y [ % ]


S h o r t C i r c u i t C u r r e n t [ A / m 2 ]


Figure 2.2: Open circuit voltage, short circuit current, fill factor, and overall power conversionefficiency as a function of active layer thickness under 1kW/m2 illumination.

Figure 2.2 shows the open circuit voltage (V oc), short circuit current ( J sc), fillfactor (FF ) and overall power conversion efficiency (η) of the fabricated devices as

a function of active layer thickness. As expected the open circuit does not vary as a

function of active layer thickness, the short circuit increases as a result of an

7/31/2019 10 Complete


CHAPTER 2

36


increase in absorption and the fill factor decreases. Due to an increasing short

circuit current being countered by a decreasing fill factor the overall powerconversion efficiency stays approximately constant. To understand the decrease in

fill factor the photocurrents of a thin (128 nm) and thick (368 nm) have been

studied in more detail including their illumination intensity dependence.

In figure 2.3 the photocurrent density J ph= J L− J D, where J L and J D are thecurrent density under illumination and in dark, respectively of a 128 nm device is

shown as a function of effective applied voltage V 0-V A. Here, V 0 is the

compensation voltage defined as J ph (V 0) =0 and V A is the applied bias. Also shown

is the space-charge limit calculated from Equation 2.1 using µh=3×10-8

m2 /Vs and

G=1.9×1027

m-3

s-1

. This device is clearly not space-charge limited and two regimes

can be distinguished. For voltages close to V 0 the photocurrent scales linearly witheffective applied voltage due to a combination of between drift- and diffusion

currents. With increasing applied voltage (V 0-V A>0.1 V) the photocurrent saturates

to J ph=qG(E,T)L

Figure 2.3: Photocurrent density J ph= J L−J D versus effective applied voltage V 0-V A of a deviceconsisting of an active layer of 128 nm. Also shown is the predicted space-charge limit using equation1.1 with µh=3×10-8m2 /Vs and G=1.9×1027 m-3s-1. The photocurrent is below the predicted space-charge limit.

0.01 0.1 1 10

10

100

Measurement

Space Charge Limit

J p h

[ A / m 2 ]

V0-V [V]

128 nm Device

7/31/2019 10 Complete



37


Figure 2.4: Photocurrent density J ph= J L− J D versus effective applied voltage V 0-V A of a deviceconsisting of an active layer of 368 nm. Also shown is the predicted space-charge limit using equation1.1 with µh=3×10-8m2 /Vs and G=0.9×1027 m-3s-1. In this case the photocurrent density follows thespace-charge limit with a square-root dependence on applied voltage.

The photocurrent density of a thick (368 nm active layer) device is shown in

Figure 2.4 together with the predicted space-charge limited using µh=3×10-8

m2 /Vs

and G=0.9×1027

m-3

s-1

. For this active layer thickness the photocurrent intersects

the space-charge limit and three regimes appear. Like in the 128 nm device the

photocurrent is linear for small applied voltages (V 0-V A<0.1 V). For 0.3V<V 0-

V A<0.7 the photocurrent density shows a square-root behavior typical for a space-

charge limited photocurrent followed again by a saturation of the photocurrent. The

fill factor of this device will be strongly reduced due to the occurrence of the

square root dependence of the photocurrent.

0.1 1 10

10

100

Measurement

Space Charge Limit

J p h

[ A / m 2 ]

V0-V [V]

368nm Device

7/31/2019 10 Complete


CHAPTER 2

38


Figure 2.5: Intensity dependent measurements performed on a thin (128 nm top) and thick (368 nmbottom) device. For the thin device the fit of α remains close to unity whereas for the thick device αapproaches the theoretic value of ¾ in the space-charge regime.

To check whether the 368 nm device is truly space-charge limited, illumination

intensity dependent measurements were performed. Figure 2.5 shows the intensity

dependence of the photocurrent for both devices at various voltages including fits

of J ph I α. For a pure space-charge limited photocurrent one expects a value of ¾

for α whereas one expects a linear (α=1) dependence in a normal device.12

For the

100 1000

1

10

100

1000

1

10

100

1

10

100

1000

100 1000

1

10

100

S = 0.90

S = 0.95

Jph

@ V0-V= 0.1 V

Incident Light Power [W/m2]

J

p h

[ A / m 2 ]

128 nm Device

Jph

@ V0-V= 1V

368 nm Device

S = 0.94

Jph

@ V0-V= 0.2V

J p h

[ A / m 2 ]

S = 0.83

Jph

@ V0-V= 5V

Incident Light Power [W/m

2

]

7/31/2019 10 Complete



39


thin device α ranges from 0.9 in the linear regime to 0.95 in the saturated regime

indicating that space-charge effects play almost no role here. For the thick devicehowever α=0.83 at V 0-V A=0.2 V, approaching the theoretical value of ¾ for the

pure space-charge dominated regime. Again the intensity dependence becomes

almost linear in the saturated regime. Hence we can conclude that the square-root

behavior of the thick device occurs due to space-charge effects.

7/31/2019 10 Complete


CHAPTER 2

40


2.4 Device Simulations

In order to gain insights into the loss mechanisms in thick and thin

polymer:fullerene bulk heterojunction solar cells fits have been performed using

the numerical program described in chapter 1. In figure 2.6 the result of the fits of

the 128 nm and 368 nm device are shown. Figure 2.7 shows the calculated

potential through both devices in dark and under illumination at maximum powerpoint. Where for the 128 nm device both potentials are equal, for the 368 nm

device the potential under illumination is altered. As predicted for a space-charge

limited device, the electric field near the anode is increased due to a build-up of

holes and the electric field near the cathode is decreased.

Figure 2.6: Simulations of a MDMO-PPV:PCBM device made with a 128 nm (a) and 368 nm (b)

active layer.

0.01 0.1 1 10

10

100

0.01 0.1 1 10

10

Measurement

Fit

(b)

J p h

[ A / m 2 ]

V0-V

A[V]

(a)

Measurement

Fit

J

p h

[ A / m 2 ]

V0-V

A[V]

7/31/2019 10 Complete



41


Figure 2.7: Simulated potential (anode right, cathode left) through a thin (128 nm, top) and thick (368

nm, bottom) device in dark and under illumination. For the thick device the potential is clearlychanged upon illumination due to space-charge effects.

The simulations allow us to address various loss mechanisms individually.

Table 2.1 lists the current density under illumination ( J L), the dissociation

probability (<P>) and recombination losses at short circuit (V sc=0) and maximum

power point (V mpp=0.653 and 0.50 for the 128 nm and 368 nm device,

respectively). The average dissociation efficiency decreases from 61% to 45%

when going from a 128 nm to a 368 nm device and it decreases to 40% at

maximum power point. This can be attributed to the decrease in electric field due

-50 0 50 100 150 200 250 300 350 400

-0.4

-0.3

-0.2

-0.1

0.00.1

0.2

0.3

0.4

-20 0 20 40 60 80 100 120 140

-0.4

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

Under illumination

In Dark

Under illumination

In Dark

P o t e n t i a l [ V

]

x [nm]

Potential in 368 nm Device at VMpp

=0.50V

x [nm]

P o t e n t i a l [ V ]

Potential in 128 nm Device at VMpp

=0.65V

7/31/2019 10 Complete


CHAPTER 2

42


to the increase in layer thickness. Secondly, the recombination losses increase for

the 368 nm devices. At short circuit conditions losses are still small but atmaximum power point 35% of the generated charges are lost due to recombination.

Hence besides space-charge effects also the decreasing dissociation probability and

charge recombination play an important role in thick polymer:fullerene bulk

heterojunction solar cells. It should be noted however that it is not possible toexactly determine these processes individually since they are interrelated. Due to

space-charge effects the electric field is reduced in a large part of the device

leading to a lower dissociation probability and increase in carrier transit times.

Therefore the increase in recombination not only originates from an increase in the

distance carriers need to travel to reach the electrode, but is also amplified by

space-charge effects.

V [V] J L [A/m2] <P> [%] rec.

loss.[%]

128 nm device

368 nm device

V sc= 0

V mpp= 0.653V sc= 0

V mpp= 0.50

29.0

19.5

59.8

37.7

61

52

45

40

2

14

9

35

Table 2.1: An overview of voltage, current density, average dissociation probability, and relative

number of free carriers lost due to recombination at short circuit (SC) and maximum power point(MPP) for a 128 nm and a 368 nm device.

7/31/2019 10 Complete



43


2.5 Optical Considerations

In the simulations described thus far, optical effects have been ignored. Due tothe relatively thin layers compared to the wavelength of the incident light, optical

interference effects are known to play an important role in organic solar cells.13,14,15

When increasing the active layer thickness the number of absorbed photons does

not increase gradually but goes through a series of maxima and minima. In our

simulations we have accommodated for this fact by determining the saturated

photocurrent at each device thickness, which is directly related to the number of

absorbed photons in the devices using equation 2.2. In this way we do not have touse optical interference modelling to determine the total amount of absorbed

photons.

Due to optical interference, however, the absorption profile through the

active layer is inhomogeneously distributed (see figure 2.8). In the above described

modelling a constant generation profile is assumed. One can imagine that an

inhomogeneous profile will alter the operation of a solar cell, especially when the

electric field itself is non-uniform as is the case in thick MDMO-PPV:PCBM cells.

0 100 200 3000

1x1027

2x1027

3x1027

4x1027

Optical Profile

Avarage

E x c i t o n g e n e r a t i o n r a t e [ # / m 3 s ]

Distance from Al electrode [nm]

Figure. 2.8. Simulated exciton generation rate profile (solid line) and its average (dashed line) as afunction of distance x from the cathode.

To see if these effects play a role in our case an optical model has been used to

calculate the generation profile of excitons in MDMO:PPV-PCBM solar cells.16

7/31/2019 10 Complete


CHAPTER 2

44


Figure 2.9 shows simulations of the IV curve of a 368 nm device using either the

average or the generation profile shown in figure 2.8. It is clear that using anaverage profile results in a slight underestimation of the simulated IV curve due to

a majority of excitons being generated near the anode (where the field in the device

is enhanced) in the real device. For layer thicknesses up to 250 nm no significant

difference between an average or a generation profile was seen in the simulations.16

-0,2 0,0 0,2 0,4 0,6

-60

-40

-20

0

20

Avarage

Generation Profile

J L

[ A / m 2 ]

V [V]

Figure. 2.9. Simulated IV curves using an average or a generation profile. Due to inhomogenities in

both electric field and generation profile the average profile leads to a slight underestimation of thesimulated IV curve.

7/31/2019 10 Complete



45


2.6 Conclusions

Increasing the active layer thickness of MDMO-PPV:PCBM bulk heterojunction solar cells does not result in a higher power conversion efficiency,

because the increase in short circuit current is cancelled by a decrease in fill factor.

Using intensity dependent measurements and simulations of the photocurrent it is

shown that a difference of one order of magnitude in the electron and hole mobility

poses a limit on the active layer thickness of around 100 nm. When increasing the

active layer beyond that point a space-charge limited regime appears in the

photocurrent limiting the fill factor of the device. Secondly, the dissociationprobability is decreased and charge recombination is increased in thicker samples,

both by space-charge effects and by an increase in the distance carriers need to

traverse. Furthermore, optical simulations show that for relatively thin active layers

assuming a uniform generation rate is a valid approximation. The way to overcome

the limitation on the active layer thickness is to enhance the transport of the slowest

charge carriers, in this case the photogenerated holes in the MDMO-PPV. State of

the art P3HT devices have been shown to maintain high fill factors for active layers

beyond 350 nm at which almost all of the incoming light is absorbed. These high

fill factors have been attributed to a balanced charge transport in the blend.17,18

7/31/2019 10 Complete


CHAPTER 2

46


REFERENCES

[1] C.J. Brabec , N.S. Sariciftci and J.C. Hummelen, Adv. Fuct. Mater. 11, 15, (2001).

[2] S. E. Shaheen, C. J. Brabec, N. S. Sariciftci, F. Padinger, T. Fromherz, J. C.

Hummelen, Appl. Phys. Lett. 78, 84, (2001).

[3] P. Schilinsky, C. Waldauf, J. Hauch, and C. J. Brabec, J. Appl. Phys. 95, 2816,

(2004).

[4] I. Riedel, J. Parisi, V. Dyakonov, L. Lutsen, D. Vanderzande, and C. J. Brabec, ,

Adv. Funct. Mater. 14, 38, (2004).

[5] I. Riedel, V. Dyakonov, Phys. Status Solidi A. 201, 1332, (2004).

[6] W. Ma, C. Yang, X. Gong, K. Lee and A.J. Heeger, Adv. Funct. Mat. 15, 1617,(2005).

[7] V. D. Mihailetchi, J. Wildeman, and P. W. M. Blom, Phys. Rev. Lett. 94, 126602,(2005).

[8] V. D. Mihailetchi, L. J. A. Koster, P. W. M. Blom, C. Melzer, B. de Boer, J. K. J.

van Duren, R. A. J. Janssen, Adv. Funct. Mater. 15, 795, (2005).

[9] A. M. Goodman and A. Rose, J. Appl. Phys. 42, 2823, (1971).

[10] V. D. Mihailetchi, L. J. A. Koster, J. C. Hummelen, and P. W. M. Blom, Phys.Rev. Lett. 93, 216601, (2004).

[11] L. J. A. Koster, E. C. P. Smits, V. D. Mihailetchi, and P. W. M. Blom, Phys. Rev.B, 72, 085205 (2005).

[12] L. J. A. Koster, V. D. Mihailetchi, H. Xie, and P. W. M. Blom, Appl. Phys. Lett.87, 203502 (2005).

[13] H. Hoppe, N. Arnold, N. S. Sariciftci, and D. Meissner, Sol. Energy Mater.Sol. Cells, 80, 105 (2003).

[14] H. Hoppe, N. Arnold, D. Meissner, and N. S. Sariciftci, Thin Solid Films, 451-452,589 (2004).

[15] N. K. Persson, H. Arwin, and O. Inganäs, J. Appl. Phys. 97, 034503 (2005).

[16] J.D. Kotlarski,P. W. M. Blom,L. J. A. Koster, M. Lenes and L. H. Slooff, J. Appl.

Phys. 103, 084502 (2008).

7/31/2019 10 Complete



47


[17] V. D. Mihailetchi, H. Xie, B. de Boer, L. J. A. Koster, P. W. M. Blom, Adv. Funct.

Mater. 16, 699, (2006).

[18] V. D. Mihailetchi, H. Xie, L. J. A. Koster, B. de Boer, L. M. Popescu, J. C.

Hummelen, P. W. M. Blom, Appl. Phys. Lett. 89, 012107, (2006)

7/31/2019 10 Complete


48


7/31/2019 10 Complete



Chapter 3

Charge dissociation in polymer:fullerene

bulk heterojunction solar cells with enhanced permittivity

Abstract

The dissociation efficiency of bound electron-hole pairs at the donor-acceptor

interface in bulk heterojunction solar cells is partly limited due to the low

dielectric constant of the polymer:fullerene blend. In this chapter the photocurrent

generation in blends consisting of a fullerene derivative and an oligo(oxyethylene)

substituted poly(p-phenylenevinylene) derivative with an enhanced relative permittivity of 4 is investigated. It is demonstrated that in spite of the relatively low

hole mobility of the glycol substituted PPV the increase of the spatially averaged

permittivity leads to an enhanced charge dissociation of 72% at short circuit

conditions for these polymer:fullerene blends.

REFERENCES

M. Lenes, F. B. Kooistra, J.C. Hummelen

I. Van Severen, L. Lutsen

D. Vanderzande, T. J. Cleij and P. W.M. Blom J. Appl. Phys, 2008 104, 114517

7/31/2019 10 Complete


CHAPTER 3

50


3.1 Introduction

An important difference between semiconductors used in inorganic and organic

solar cells is the much lower permittivity of the latter. As a result, strongly bound

excitons are created after absorption of light instead of free charge carriers. To

overcome this problem a donor-acceptor system is used in which the electron

transfers from the donor to the acceptor material. However, electron-hole (e-h)pairs generated in this way, are still strongly bound by Coulomb interaction and

need to be dissociated into free carriers in order to be collected at the electrodes.1,2

The occurrence of such an interfacial geminate charge pair has been

spectroscopically observed for polymer:polymer systems 3,4 and recently also for

polymer:fullerene blends.5

In addition, a strong indication for the existence of a

bound e-h pair in polymer:fullerene blends is the field- and temperaturedependence of the photocurrent at reverse bias. At sufficiently high reverse bias all

bound e-h pairs are dissociated, leading to a saturated photocurrent that is field-

and temperature independent.2

As a result, the saturated photocurrent is a direct

measure for the amount of photons absorbed in the blend.6

Typically, thedissociation efficiency in poly(2-methoxy-5-(3´,7´-dimethoxyloctyloxy)-p-

phenylene vinylene) (MDMO-PPV) and [6,6]-phenyl-C61-butyric acid methyl ester

(PCBM) bulk heterojunction (BHJ) solar cells is only 60% at short circuit

conditions, representing a major loss mechanism in these devices.2,7

In this chapter

an oligo(oxyethylene) substituted PPV derivative (Figure 3.1) with an enhancedrelative permittivity of 4 is investigated to see whether the dissociation efficiency

in PPV based solar cells can be enhanced.

Figure 3.1: Chemical structure of poly[2-methoxy-5-(triethoxymethoxy)-1,4-phenylene vinylene](PEO-PPV) and phenyl-C61-butyric acid 2-ethylhexyl ester (PCB-EH) being the donor and acceptormaterial used in this chapter, respectively.

7/31/2019 10 Complete




51

3.2 Parameters governing the charge dissociation

The field- and temperature dependent process of charge dissociation of thebound e-h pair in polymer:fullerene solar cells can be described using Onsager‟stheory of ion pair dissociation.

8This e-h pair can dissociate into free carriers by a

rate constant k D given by:

1801831

4

3)(

432 /

3

bbbbe

ak E k

kT E

R D B (3.1)

with E B being the electron-hole pair binding energy, )8 /( 223T k E qb B , E is

the field strength and ε=ε0εr the permittivity.1

Free carriers generated in this way

can recombine back to the bound state by Langevin recombination with a rate k R9

),min(k R he

q(3.2)

where it has been pointed out that the recombination strength is dominated by the

slowest carrier mobility in bulk heterojunction solar cells.10

Finally, the bound state

can decay to the ground state with a rate k F . The model predicts the probability thatfree charge carriers will be produced at a particular field ( E ), temperature (T ), and

donor-acceptor separation radius (a):

F D

D

k E k

E k E T aP

)(

)(),,( (3.3)

From the above-mentioned equations, it is clear that the charge dissociation is

governed by four relevant parameters, viz. the charge carrier mobility μ, the

permittivity of the blend ε, the initial separation distance a, and the decay rate k F .

For the MDMO-PPV:PCBM system it has been shown to be vital to take into

account the overall relative permittivity of the blend when describing the chargedissociation.

11The relative permittivity εr of MDMO-PPV is 2.1,

12whereas the εr

of PCBM amounts to 4. As a result, when going from a 1:1 to a 1:4

polymer:fullerene weight ratio the increase in photocurrent is shown to originate

not only from an increase in the hole mobility, but also from an increase of the

average permittivity, due to the loading of more PCBM. An increase in the

permittivity of the donor polymer is therefore expected to enhance the dissociationefficiency and thus the efficiency of the solar cells. In this study an

oligo(oxyethylene) substituted PPV derivative with εr = 4 is used to study the effect

of an enhanced permittivity of the blend on the photogeneration.

7/31/2019 10 Complete


CHAPTER 3

52


0,1 1 10

0,6

0,8

1

2,0 2,5 3,0 3,5 4,00,60

0,65

0,70

0,75

r=4

D i s s o

c i a t i o n E f f i c i e n c y

V0-V [V]

r=2 D

i s s o c i a t i o n

a t S C

Permittivity

Figure 3.2: Calculated dissociation efficiency versus effective applied voltage with a polymerpermittivity ranging from 2 to 4. The dissociation efficiency at short circuit conditions is indicated inthe inset.

In figure 3.2 the expected increase in dissociation efficiency is shown. Here we

start with calculating the dissociation using the Braun model and input parameters

as determined in ref 7. The permittivity of the polymer donor material is nowraised until the value of 4, taking into account the 1:4 polymer:fullerene weight

ratio in the blend. From the calculations we expect a significant increase in

dissociation efficiency at short circuit conditions to 73%.

7/31/2019 10 Complete




53

3.3 Single Carrier Devices

Figure 3.1 shows the chemical structure of poly[2-methoxy-5-(triethoxymethoxy)-1,4-phenylene vinylene] (PEO-PPV), the donor material used

in this chapter. The synthesis and characterization have been previously reported.13

Furthermore, previous studies using impedance measurements have shown the

relative permittivity of the material to be equal to 4.14

Also shown in figure 3.1 is

the chemical structure of phenyl-C61

-butyric acid 2-ethylhexyl ester (PCB-EH), the

acceptor used here. PCB-EH is used instead of PCBM to provide a better mixing of

donor and acceptor. The charge transport in pristine polymer films was investigatedby sandwiching a layer of PEO-PPV between a layer of indium tin oxide (ITO)

covered with ~70 nm of poly(3,4ethylenedioxythiophene)/poly(styrenesulfonate)(PEDOT:PSS) and a palladium (Pd, 20 nm)/gold (Au, 80 nm) top electrode. The

high work function of Pd prevents electron injection and only holes flow through

the device. Figure 3.3 shows the J-V characteristics of such a PEO-PPV hole-only

device, corrected for the built-in voltage (V bi) and the series resistance of the

electrodes. The J-V curve is fitted to a space-charge limited current (SCLC),

yielding a hole mobility of 1.8 x 10-10

m2 /Vs. The observed hole mobility is

comparable to hole mobilities reported for pristine MDMO-PPV films.15

ForMDMO-PPV, however, blending the polymer in a 1:4 weight ratio with PCBM

results in a dramatic increase in hole mobility of more than two orders of

magnitude,16 which turns out to be essential for device operation.2,7

Figure 3.3: Current density versus voltage, corrected for built-in voltage and series resistance of PEO-PPV hole only device. Data (symbols) is fitted (solid line) using a space-charge limited current.

As a next step the hole transport in a PEO-PPV:PCB-EH blend (1:4 weightratio) is investigated , also using a Pd top electrode to prevent electron injection

into the PCB-EH. In figure 3.4 the J-V characteristics of such a hole only device

are shown. The determined hole mobility of 4 x 10-11

m2 /Vs indicates that an

enhancement of the hole mobility, as seen in blends of MDMO-PPV, does not

7/31/2019 10 Complete


CHAPTER 3

54


occur for blends of PEO-PPV and PCB-EH. As was shown in chapter 2 a low hole

mobility is likely to lead to the formation of space-charges as well as a lowdissociation efficiency and is expected to limit the performance of the PEO-

PPV:PCB-EH solar cells. To separately illustrate the effect of a low charge carrier

mobility on the charge dissociation efficiency, figure 3.5 shows calculations of the

Braun model using parameters for the MDMO-PPV:PCBM system;7

a=1.25x10-9m, k f =1x105

s-1

r =3.4 and h=4x10-8

m2 /Vs and identical parameters

but now with the lower hole mobility of PEO-PPV, 4x10-11

m2 /Vs. The calculated

charge dissociation at short-circuit is reduced from 62% to only 22% by lowering

the hole mobility to such a low value. Furthermore, the formation of space-charges17

is expected to limit the fill factor.

Figure 3.4: Current density versus voltage, corrected for built-in voltage and series resistance of PEO-PPV:PCB-EH hole only device. Data (symbols) is fitted (solid line) using a space-charge limitedcurrent..

Figure 3.5: Dissociation efficiency for an MDMO-PPV:PCBM solar cell as determined in ref 7 (solid

line), together with a calculated dissociation efficiency for a similar system but with a lower hole

mobility of 4x10-11 m2 /Vs. The dissociation efficiency at short circuit conditions (as indicated by thevertical solid line) drops from 62% to 22% due to the lower hole mobility.

7/31/2019 10 Complete




55

3.4 PEO-PPV:PCB-EH Solar Cells

Polymer:fullerene bulk heterojunction solar cells were fabricated by spincoating a layer of PEO-PPV:PCB-EH in a 1:4 weight ratio from chlorobenzene on

top of PEDOT:PSS covered ITO. As a cathode, 1 nm of lithium fluoride and 100

nm of aluminium was evaporated. Figure 3.6 shows the J-V characteristics of a

PEO-PPV:PCB-EH solar cell measured under illumination of a white light halogen

lamp set at 1000W/m2. Due to the low charge carrier mobility the optimal device

thickness is only 68 nm,18

resulting in a low short circuit current of 13.8 A/m2.

Combined with a fill factor of 52% and an open circuit voltage of 0.67 V theestimated overall power efficiency is 0.5% which is considerably lower compared

to the model system MDMO-PPV-PCBM. The main reason for this lower

efficiency is the low short circuit current which is likely to be caused by an

unfavourable large domain formation of polymer and fullerene, leading to a loss of

excitons.

Figure 3.6: Current density versus voltage characteristics for a PEO-PPV:PCB-EH solar cell under

illumination of a 1000 W/m2 halogen lamp.

7/31/2019 10 Complete


CHAPTER 3

56


3.5 Device Simulations

Even though the power conversion efficiency is much lower as compared toMDMO-PPV:PCBM the devices can still be used to study the effect of the raised

permittivity on the charge dissociation. For this, the photocurrent density is plotted

as a function of effective applied voltage V 0-V A, as is shown in figure 3.7. The

previously mentioned Braun model can only be applied to describe the

photocurrent at relatively high effective voltages where the photocurrent is in the

saturated regime and is governed by the dissociation of bound electron-hole pairs.2

To fully describe the photocurrent the numerical program was used.7

Inputparameters for the numerical program were identical to the above described Braun

model i.e., the charge carrier mobilities including their field dependence, the

average permittivity εr , separation distance a and decay rate k f . Since the charge

carrier mobility has been determined using hole only diodes and the electron

mobility of the fullerene is known19

only a and k f were used as fitting parameters.

Both separation distance a and decay rate k f have a different effect on the fits of the

photocurrent. The distance a determines at which voltage the dissociation saturates

whereas the decay rate k f determines how fast the dissociation drops when the field

in the device is lower. Indicated in figure 3.7 is a fit to the experimental

photocurrent using our numerical program with input parameters εr =4, h=4x10-11

m2 /Vs, a=1.5x10

-9m, e=2.5x10

-7m

2 /Vs and k f =4x10

4s

-1. Note that the value to

which the photocurrent saturates is considerably lower compared to normalPPV:PCBM cells. This is due to the in paragraph 3.4 mentioned coursemorphology which leads to a loss of excitons. The amount of lost excitons can in

principle be calculated by comparing the saturated photocurrent with the amount of

absorbed photons. For the analysis of the dissociation efficiency these lost photons

do not play an important role.

In figure 3.8, using the same input parameters, a fit using the Braun model

is shown. From these figures one can see that for voltages V 0-V >3 the photocurrent

is dominated by the field dependent dissociation of bound electron hole pairs and

that bimolecular recombination and space-charge effects do not play a role.

7/31/2019 10 Complete




57

Figure 3.7: Experimental photocurrent density Jph as a function of effective applied voltage V 0−VA under 1kW/m2 illumination for a PEO-PPV:PCB-EH solar cell. Circles indicate experimental data,

solid line fit of the photocurrent.

This allows one to directly determine the dissociation efficiency by comparing

the generated photocurrent at short circuit with the saturated photocurrent at a large

reverse bias resulting in a dissociation efficiency of no less than 72%. Thus, we canconclude that, even when the charge carrier mobility is significantly lower as

compared to MDMO:PPV, the charge dissociation at short circuit is increased by

using a high permittivity polymer. Above we calculated the dissociation efficiency

for materials with a hole mobility of 4x10-11

m2

/Vs and normal permittivity to beonly 22%. The origin of the observed enhanced dissociation efficiency in the PEO-

PPV:PCB-EH blend is due to two effects, as shown in figure 3.8; First of all, the

initial separation distance of charges is enlarged from 1.25 x10-9

to 1.5 x 10-9

m and

the decay rate k f is lowered from 1x105

s-1

to 4x104

s-1

. Using these parameters we

predict the dissociation efficiency to be 45%. The second effect is the direct effectof the higher relative permittivity of PEO-PPV of 4, raising the dissociation

efficiency even more to the measured 72%. With these parameters, but now

combined with the MDMO:PPV hole mobility of h=4x10-8

m2 /Vs, the calculated

dissociation is as high as 78%. Combined with an increase in fill factor, our

numerical program predicts a power conversion efficiency of 3.5% to be possible

for a PPV-type polymer with a relative permittivity of 4.

7/31/2019 10 Complete


CHAPTER 3

58


Figure 3.8: Dissociation efficiency as a function of effective voltage calculated for various values of

permittivity of the polymer ε, initial separation distance a and decay rate k F . In all cases the holemobility is taken to be the measured value of 4x10 -11 m2 /Vs. Starting from a low dissociation of 22%at short circuit conditions (indicated by the vertical solid line) the dissociation efficiency is increasedfirst to 45% by an improved separation distance and decay rate and subsequently to the measuredvalue of 72% by increasing the relative permittivity of the polymer from 2 to 4. The measuredphotocurrent is indicated as a reference (symbols).

7/31/2019 10 Complete




59

3.6 Conclusions

To conclude, an oligo(oxyethylene) substituted PPV derivative was used tostudy the effect of an enhanced permittivity on the dissociation efficiency in

polymer:fullerene bulk heterojunction solar cells. Besides the permittivity, also the

charge carrier mobility, separation distance, and decay rate are important factors

determining the charge dissociation. Despite a low hole mobility of 4x10-11

m2 /Vs

in the blend of PEO-PPV and PCB-EH, a dissociation efficiency of 72% was

observed. It was shown that the effect of a higher relative permittivity is twofold.

Not only a direct enhancement of the charge dissociation is observed, but also theseparation distance and decay rate are improved upon increasing the relative

permittivity. Therefore, it is concluded that enhancing the relative permittivity of

the polymer can be very beneficial for the device operation of polymer:fullerene

solar cells.

7/31/2019 10 Complete


CHAPTER 3

60


REFERENCES

[1] C. L. Braun, J. Chem. Phys. 80, 4157 (1984).

[2] V. D. Mihailetchi, L. J. A. Koster, J. C. Hummelen, and P. W. M. Blom, Phys.Rev. Lett. 93, 216601 (2004).

[3] C. Yin, T. Kietzke, D. Neher and H.H. Horhold, Appl. Phys. Lett 80, 092117

(2007).

[4] A.C. Morteani, P. Sreearunothai, L.M. Herz, R.H. Friend and C. Silva, Phys. Rev.Lett. 92, 247402 (2004).

[5] D. Veldman, O. Ipek, S. C. J. Meskers, J. Sweelssen, M. M. Koetse, S. C.

Veenstra, J. M. Kroon, S. S. van Bavel, J. Loos, R. A. J. Janssen, J.Am. Chem.Soc. 130, 7721 (2008).

[6] J.D. Kotlarski, P.W.M. Blom, L.J.A. Koster, M. Lenes and L.H. Slooff, J. Appl.Phys. 103, 084502 (2008).

[7] L. J. A. Koster, E. C. P. Smits, V. D. Mihailetchi, and P. W. M. Blom, Phys. Rev.B, 72, 085205 (2005).

[8] L. Onsager, Phys. Rev. 54, 554 (1938)

[9] P. Langevin, Ann. Chim. Phys. 28, 433 (1903).

[10] L.J.A. Koster, V.D. Mihailetchi and P.W.M. Blom, Appl. Phys. Lett. 88, 052104

(2006).

[11] V.D. Mihailetchi, L.J.A. Koster, P.W.M. Blom C. Melzer, B. de Boer, J.K.J. vanDuren and R.A.J. Janssen, Adv. Func. Mat. 15, 795 (2005).

[12] H.C.F. Martens, H.B. Brom and P.W.M. Blom, Phys. Rev. B, 60, R8489 (1999).

[13] I. Van Severen, M. Breselge, S. Fourier, P. Adriaensens, J. Manca, L. Lutsen, T.J.

Cleij and D. Vanderzande, Macromol. Chem. Physic. 208, 196, (2007).

[14] M. Breselge, I. van Severen, L. Lutsen, P. Adriaensens, J. Manca, D. Vanderzandeand T. Cleij, Thin Solid Films, 511, 328 (2006).

7/31/2019 10 Complete




61

[15] P.W.M. Blom, M.J.M de Jong and M.G. van Munster Phys. Rev. B, 55, R656(1997).

[16] C. Melzer, E.J. Koop, V.D. Mihailetchi and P.W.M. Blom, Adv. Fuc. Mat. 14, 865(2004).

[17] V. D. Mihailetchi, J. Wildeman, and P. W. M. Blom, Phys. Rev. Lett. 94, 126602(2005).

[18] M. Lenes, L.J.A. Koster, V.D. Mihailetchi and P.W.M. Blom Appl. Phys. Lett. 88,

243502 (2006).

[19] 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. Func. Mat. 13, 43(2003)

7/31/2019 10 Complete


CHAPTER 3

62


7/31/2019 10 Complete


Chapter 4

Recombination-limited photocurrents in

small bandgap polymer:fullerene solar cells

Abstract

The charge transport and photogeneration in solar cells based on the low bandgap

conjugated polymer, poly[2,6-(4,4-bis-(2-ethylhexyl)-4H-cyclopenta[2,1-b;3,4-

b′]dithiophene)-alt-4,7-(2,1,3-benzothiadiazole)] (PCPDTBT) and a

methanofullerene is studied. The efficiency of the solar cells is limited by a

relatively low fill factor that contradicts with the observed good and balanced

charge transport in these blends. Intensity dependent measurements display a

recombination limited photocurrent, characterized by a square root dependence on

effective applied voltage, a linear dependence on light intensity, and a constant

saturation voltage. Numerical simulations show that the origin of the

recombination limited photocurrent stems from the short lifetime of the bound

electron-hole pairs at the donor-acceptor interface.

REFERENCES

M. Lenes, M. Morana

C. J. Brabec and P. W. M. Blom

Adv. Funct. Mat. 19 , 1106 (2009)

7/31/2019 10 Complete


CHAPTER 4

64


4.1 Introduction

As discussed in chapter 1 the large offset between donor and acceptor LUMO

of P3HT and PCBM results in a significant loss of energy. One way to overcome

this problem is by decreasing the LUMO level of the polymer creating so called

small bandgap donors. Due to the lowering of the donor bandgap the absorption is

expanded towards higher wavelengths, allowing more photons to be absorbed, evenwhen one takes into account that not all photons above the bandgap are absorbed.

Besides an expected enhanced efficiency compared to P3HT:PCBM when used in

single active layers, small bandgap donors are also desired for multijunction solar

cells or infrared photodetectors.1

One route towards small bandgap polymers is by coupling electron donor and

acceptor units together in a polymer. Most of the polymers created using this route,however, have resulted in significantly inferior performances compared to solar

cells based on P3HT. The reason for the low performance is mainly due to the poor

carrier transport in these polymers, resulting in low fill factors and quantum

efficiencies.2,3,4,5,6,7,8,9

One of the most promising devices following this approachare based on poly [2,6-(4,4-bis-(2-ethylhexyl)-4H-cyclopenta[2,1-b;3,4-

b′]dithiophene)-alt-4,7-(2,1,3-benzothiadiazole)] (PCPDTBT), reaching power

conversion efficiencies of up to 3.2% when combined with [6,6]-phenyl-C 71-

butyric acid methyl ester ([70]PCBM).10

In spite of the increased absorption the

power efficiency is still lower than the state-of-the-art P3HT:PCBM cells, of whichefficiencies have been reported of more than 5%.11

The efficiency is mainly limited

by a low fill factor (FF ) of only 40%. In chapter 2 it has been demonstrated that astrongly unbalanced charge transport leads to space-charge limited photocurrents,

characterized by a square-root dependence on applied voltage.12

This dependence

limits the fill factor to about 40%. Remarkably, measurements performed on

PCPDTBT-based field effect transistors resulted in hole mobilities of the polymer

as high as 2×10 – 6

m2V

– 1s – 1

. Even though field-effect mobilities are quantitatively

difficult to relate to charge carrier mobilities in actual solar cells, due to the muchlower charge carrier densities in the latter devices,

13the high field-effect mobilities

clearly indicate that the quality of the hole transport in PCPDTBT must be very

good.14

Combined with the electron transport properties of intrinsic PCBM films,

which already have been investigated in great detail,15

a balanced transport istherefore expected. Consequently, the origin of the reduced fill factors and external

quantum efficiencies in these blends is not clear. In this chapter the charge

transport and photogeneration of PCPDTBT:PCBM solar cells is studied to gain

more insight into the loss mechanisms in this type of devices.

7/31/2019 10 Complete




65

4.2 Charge transport in pristine PCPDTBT films

As mentioned above, even though field-effect mobilities give a valuable insightinto the quality of the charge carrier transport, ideally one would like to measure

the charge carrier mobility in a device geometry similar to the actual solar cell.

Here the charge transport is studied in a vertical device geometry, similar to solar

cells. By choosing suitable top and bottom contacts one can either inject both

charge carriers, or choose to block one carrier and measure either the hole or

electron current. The transport through these single carrier devices is modeled with

a space-charge limited current (SCLC). Figure 4.1 shows the measured J-V characteristics of a hole only device of PCPDTBT. From the J-V measurements the

zero-field mobility is determined to be 5.5x10-8

m2 /Vs. This mobility is about a

factor of 30 lower than the earlier reported field-effect mobility, due to the density

dependence of the mobility. However, the observed hole mobility of 5.5x10-8

m2 /Vs for PCPDTBT is about a factor of 2-3 larger than the mobility obtained from

diodes based on pristine regio-regular P3HT.16

As a result in its pristine form

PCPDTBT is at least as good a hole transporter as regio-regular P3HT.

Since the electron transport in polymer:fullerene blends occurs through the

fullerene phase, electron transport through the polymer is of no importance for the

device operation of organic solar cells. Nevertheless, also the electron transport

through pristine PCPDTBT films is studied as shown in Figure 4.1. As reported

previously,10 the polymer also shows signs of electron transport. In fact, theobserved electron mobility of 4x10

-9m

2 /Vs is only one order of magnitude lower

than the hole mobility. Interestingly, the electron transport in the pristine material

exhibits normal SCLC behavior where polymers usually show a much stronger

voltage dependence due to charge trapping.17

0,0 0,4 0,8 1,2 1,6 2,0 2,41E-3

0,01

0,1

1

10

100

1000

Hole Only Device

Electron Only Device

J [ A / m 2 ]

V-Vres

-Vbi

Figure 4.1: J-V characteristics, corrected for built-in voltage and series resistance, of a PCPDTBT

hole and electron only device. Data are fitted (solid line) with a space-charge limited current using afield dependent mobility resulting in a hole mobility of 5.5 x 10 -8 m2 /Vs and electron mobility of 4x10-9 m2 /Vs, respectively.

7/31/2019 10 Complete


CHAPTER 4

66


4.3 Charge transport in PCPDTBT:PCBM blends

Blending a polymer with a fullerene often significantly alters the charge carriertransport in both polymer and fullerene compared to the pristine case. For instance,

in the case of MDMO-PPV, a 200 fold increase in hole mobility is observed when

blending the polymer with PCBM.18

On the other side, blending P3HT with PCBM

results in a reduced hole transport, only to be recovered by thermal or solvent

annealing.19,20

Therefore, single carrier measurements on the actual blend used in

the solar cell are needed to relate the charge carrier transport to the solar cell

performance. Blends of PCPDTBT and PCBM were prepared in a 1 to 4 weightratio, which was reported to be optimal.

10The charge transport is determined in

single carrier devices as described above for pristine polymer films. Figure 4.2

shows the J-V characteristic of a hole and electron only device of a

PCPDTBT:PCBM blend with a weight ratio of (1:4). The determined hole mobility

of PCPDTBT in the blend of 3x10-8

m2 /Vs almost equals the hole mobility in

pristine films. This indicates that the hole transport in the polymer is not altered by

blending it with PCBM. Furthermore, the determined hole mobility is equal to hole

mobilities reported in MDMO-PPV:PCBM (1:4) blends18

and P3HT:PCBM (1:1)

blends after annealing.16

The determined electron mobility of 7x10-8

m2 /Vs is

slightly lower than values reported for MDMO-PPV:PCBM and P3HT:PCBM

blends, that typically amount to 1.0x10-7

-2.0x10-7

m2 /Vs.

18,19,20Similar electron and

hole mobility values were found by ambipolar transport studies on OFETs at highfullerene loadings.

21

0,0 0,7 1,4

0,1

1

10

100

1000

10000

Hole Only DeviceElectron Only Device

J

[ A / m 2 ]

V-Vbi-V

rs

Figure 4.2: J-V characteristics, corrected for built-in voltage and series resistance, of a

PCPDTBT:PCBM hole- and electron-only device with a weight ratio of (1:4). Data are fitted with a

space-charge limited current using a field-dependent mobility, resulting in a hole mobility in the

7/31/2019 10 Complete




67

blend of 3x10-8 m2 /Vs and an electron mobility of 7x10-8 m2 /Vs.

The single carrier measurements presented here demonstrate that in the blendsthe hole and electron mobilities are balanced and closely match the mobilities

reported for MDMO-PPV and P3HT based blends. It is therefore highly unlikely

that the relatively low quantum efficiencies and fill factors are a consequence of

unbalanced transport or too low charge carrier mobilities and more investigation of the solar cells is needed.

7/31/2019 10 Complete


CHAPTER 4

68


4.4 PCPDTBT:PCBM Solar Cells

The inset of Figure 4.3 shows the current vs. voltage ( J-V ) curve at roomtemperature of a typical PCPDTBT:PCBM solar cell made in this study. The

external quantum efficiency (see figure 4.4) has been determined at ECN in Petten

to estimate the correct short circuit current under AM 1.5 illumination and thus the

mismatch factor of our measurements. Efficiencies of 2.2% are obtained which is

somewhat lower than the 2.7% reported previously for PCPDTBT:[60]PCBM.10

As

reported previously the power conversion efficiency is limited by a low external

quantum efficiency (<35%) and fill factor (40%).

-0.2 0.0 0.2 0.4 0.6

-60

-40

-20

0

0.1 1 10

10

100

J L

[ A / m 2 ]

V [V]

J p h

[ A / m 2 ]

V0-V

Figure 4.3: Photocurrent of a PCPDTBT:PCBM solar cell versus effective applied voltage. The black line indicates a square root dependence. Inset: J-V characteristics of a PCPDTBT:PCBM solar cell.

7/31/2019 10 Complete




69

400 500 600 700 800 9000

5

10

15

20

25

30

35

40

45

E Q E [ % ]

Wavelength [nm]

Figure 4.4: External Quantum Efficiency of a PCPDTBT:PCBM solar cell.

For studying the device physics it is very useful to plot the photocurrent of asolar cell as a function of effective applied voltage. The photocurrent density is

defined as J ph=J L−J D, where J L and J D are the current density under illumination

and in dark, respectively, and the effective applied voltage as V eff =V 0−V A. Here V 0

is the compensation voltage defined as J ph(V0)=0 and V A is the applied bias. The

photocurrent versus effective applied voltage of a PCPDTBT:PCBM solar cell isalso shown in figure 4.3. It is clear that at large reverse bias the photocurrentsaturates, at which point all generated electron-hole pairs are dissociated and

collected at the electrodes, which indicates that the mean electron and hole drift

lengths we(h) = e(h)τ e(h) E are equal to, or larger than the sample thickness L and no

recombination occurs.22

The photocurrent shows a sharp decrease at lower effective

applied voltages, resulting in a rather low short circuit current and low fill factor.

Furthermore, a square root dependence of the photocurrent as a function of

effective voltage is observed, as is indicated by the black line. The origin of such a

square root dependence of the photocurrent has been explained by Goodman and

Rose in 1971.22

If the mean electron or hole (or both) drift length becomes smaller

than L, recombination of charge carriers becomes considerable. If there is also a

difference between hole and electron drift length, a non uniform electric field will

occur across the devices, which will give rise to a square-root dependentphotocurrent:

V qG J eheh ph )()( (4.1)

with G the generation rate of free charge carriers. Here a low mobility or shortlifetime of the free carriers, due to recombination or trapping, limits the

7/31/2019 10 Complete


CHAPTER 4

70


photocurrent. Additionally, at high light intensities the build up of space-charges

(which is the origin of the non-uniform electric field) reaches a fundamental limit.In this limit the maximum electrostatically allowed photocurrent is limited by the

mobility of the slowest charge carrier and is given by

V qG J hr ph

25.0

0

75.0

8

9)( (4.2)

which again has a square-root dependence on voltage. The latter has been

experimentally demonstrated in a system where the charge carrier mobilities are

heavily unbalanced.12 The way to distinguish between these two physically distinct

cases is by light-intensity dependent measurements. Where in the first(recombination limited) case the photocurrent scales linearly with light intensity,

in the second (the space-charge limited) case it scales with a ¾ power law

dependence. Furthermore, the point at which the square root regime forms a

transition into the saturation regime, the saturation voltage V sat , is either

independent on light intensity (recombination-limited) or scales with a one half

power on light intensity (space-charge limited case). From figure 4.5 it is clear that

with decreasing light-intensity V sat is not changing, as expected for arecombination-limited photocurrent. Furthermore, in Fig. 4.6 it is shown that in the

square-root regime the photocurrent is linearly scaling with light-intensity. As a

result the photocurrent observed for PCPDTBT:PCBM devices clearly shows thefingerprints of a recombination-limited photocurrent. Since the mobilities of the

charge carriers in the device are known we can estimate the lifetime using equation

(2), resulting in a lifetime of ~ 10-7

s. This value, estimated under the assumption

that the dominant limitation comes from the hole transport, may slightly change if

electron transport is considered as well.

7/31/2019 10 Complete




71

0.1 1 10

10

100

J p h

[ A / m 2 ]

V0-V [V]

Vsat

Figure 4.5: Photocurrent of a PCPDTBT:PCBM solar cell versus effective applied voltage at differentintensities varying over more then 1 order of magnitude. Solid lines indicate square root andsaturation regimes as a guide for the eye where Vsat indicates the saturation voltage.

1000

10

100

Veff=0.3 S=1.01V

eff=0.7 S=1.02

Veff

=4.0 S=1.04

J p h

[ A / m 2 ]

Intensity [W/m2

]

Figure 4.6: Intensity dependence of the photocurrent at different effective voltages. The slope (S)

determined from the linear fit (solid lines) to the experimental data is indicated in the figure.

7/31/2019 10 Complete


CHAPTER 4

72


4.5 Device Simulations and Discussion

In a polymer:fullerene solar cell the photogenerated excitons dissociate at thedonor-acceptor interface via an ultrafast electron transfer from the donor to the

acceptor. However, the ultrafast electron transfer to the acceptor does not directly

result in free carriers, but in a bound electron-hole pair (due to the Coulomb

attraction between the carriers). This pair also needs to be dissociated, assisted by

temperature and by the internal electric field, before it decays to the ground state.20

As proposed by Braun, this bound pair is metastable, enabling multiple

dissociations and being revived by the recombination of free charge carriers viaLangevin recombination.

23Finally, the free carriers are transported to the

electrodes, a process governed by charge carrier mobility. In the above mentioned

Goodman and Rose model a direct generation of free carriers (from now on calledGGR) is assumed. In a polymer:fullerene solar cell, however, the amount of

generated free carriers will not only depend on the amount of generated bound

electron-hole pairs (G B), but also on their dissociation probability (P). In that case

the generation rate of bound pairs G B is proportional to the incident light intensity

and is taken as a measure for the amount of absorbed photons (assuming that all

generated excitons dissociate at the donor-acceptor interface). As a result, when

equation 4.1 is applied to an organic solar cell the calculated lifetime can only be

considered as an effective lifetime (τ eff ). This can be seen more clearly when one

considers the device at open circuit voltage: Since no charges are extracted ( J ph=0)there is an equilibrium between the generation and recombination of free charge

carriers in the device, given by:

eff

GR

nG (4.3)

with n=p being the free electron/hole density and GGR the recombination rate of

free carriers. Thus, if τ eff is small, indicating lots of recombination, also the freecarrier density will be small for a given generation rate of free carriers GGR. When

the formation and dissociation of bound electron-hole pairs as an intermediate step

is taken into account the amount of free carriers that are generated will be given by

PG B and hence one can state that

nPG B

(4.4)

where τ is now the true lifetime of free charge carriers as given by Langevin

recombination.

One can also say that

7/31/2019 10 Complete




73

eff

B

n

P

nG (4.5)

Thus, when the bound-pair generation rate G B is taken as a measure for the amount

of generated charges, as has been done in device modeling, the effective lifetime

τeff can be small either due to a small life time τ of the free carriers or due to a low

dissociation probability P of the bound pairs.

0.1 1 10

10

100

0.1 1 10

10

100

Data

Simulations

J p h

[ A / m 2 ]

V0-V [V]

Dissociation Probability

DissociationProbability[%]

Figure 4.7: Simulation of the photocurrent at different intensities. Symbols represent measurement,solid line fit to the data, dotted line calculated dissociation efficiency

In order to disentangle the effects of P and τ on τ eff we performed device

simulations using a numerical program which solves Poisson‟s equation and thecontinuity equations, including diffusion, space-charge effects and chargedissociation of bound electron-hole pairs.

24Relevant parameters for the simulation

program are the charge carrier mobilities, including their field and/or densitydependence, dielectric constant , separation distance a and the decay rate of the

bound electron-hole pairs k f . Since the charge carrier mobilities are measured and

the dielectric constant is known only a and k f are used as fitting parameters. Figure

4.7 shows the fit of the simulation program using a = 2.1x10-9

m and k f = 1.7x107

s-

1. Using the same fit parameters we can fit all measured light intensities. When the

calculated dissociation probability is compared with the measured and simulatedphotocurrent (as is indicated in figure 4.7) it is clear that the strong field

dependence of the photocurrent for effective voltages > 0.4 V originates from the

field dependent dissociation of the bound electron-hole pairs. What is striking in

7/31/2019 10 Complete


CHAPTER 4

74


the device simulations is the high value of k f needed to fit the data. As an

indication, for MDMO-PPV:PCBM and P3HT:PCBM cells a value of ~ 104

isfound. This indicates that the solar cells are limited by a high decay-rate, and thus

short lifetime, of the bound electron-hole pair. Recently Hwang et. al. have shown

experimental evidence of such an intermediate charge transfer state with a short

lifetime using photoinduced absorption spectra.25

To indicate the strong effect thisdecay rate has on the performance of the solar cells, simulations with a decreasing

k f have been performed up to the point at which the decay rate is equal to P3HT

and MDMO-PPV values (see figure 4.8). Upon lowering of the decay rate the

typical square root behavior disappears and the photocurrent becomes significantly

less field dependent, manifesting itself in a greatly increased short circuit current

and fill factor. The simulations indicate that when k f for the PCPDTBT:PCBMdevices would be as low as for the P3HT:PCBM cells an efficiency of ~ 7% can be

achieved. This demonstrates the potential of these low band gap polymer:fullerene

devices, when the increased recombination of the bound pairs can be prevented.

-2.0 -1.5 -1.0 -0.5 0.0 0.5

-120

-100

-80

-60

-40

-20

0

0.1 1 10

10

100

DataFit

J L

[ A / m 2 ]

Voltage [V]

P h o t o c u r r e

n t [ A / m 2 ]

V0-V

decreasing kf

Figure 4.8: Simulation of photocurrent for different values of the decay rate k f starting from thedetermined decay rate of 1.7x107 s-1 and decreasing one order at a time until the value of 1.7x10 4 s-1

typical for normal polymer:fullerene systems. Inset: Simulation of the current under illumination for

these values of k f .

Above, we have shown that PCPDTBT:PCBM solar cells show signs of a

recombination limited photocurrent as predicted by Goodman and Rose. Using our

numerical simulation program we are able to show that the low dissociation

probability is the cause of the low effective lifetime of the free carriers τ eff . We

show the lifetime of the bound electron-hole pair to be significantly shorter

7/31/2019 10 Complete




75

compared to other polymer:fullerene systems. Moreover, the effective lifetime

predicted using equation (4.2) ~ 10-7

s, matches the lifetime of the bound pair 1/ k f .(k f .= 1.7x10

7s

-1) predicted by the simulation model.

Therefore, we can conclude that the decrease of the photocurrent at low

effective voltages, and hence low fill factor of the device, is due to a short lifetime

of the bound electron-hole pairs. In earlier work an almost complete intermixing of the PCPDTBT polymer with PCBM at the molecular level was reported.

26When

donor and acceptor are too closely intermixed carriers can end up being trapped in

dead ends and will not dissociate fully into free carriers leading to a large decay

rate and hence small effective lifetime. This hypothesis seems to be confirmed by

recent results on PCPDTBT:PCBM solar cells by Peet et al.27

Here they show that

the addition of alkanedithiols to the solution results in a dramatic increase in deviceperformance. It is shown that adding alkanedithiol results in larger phase separation

of donor and acceptor which in turn results in a much higher fill factor and external

quantum efficiency. Apparently the larger phase separation results in an increase of

the effective lifetime of the charge carriers, such that the device is no longer

recombination limited, as predicted by the simulations.

7/31/2019 10 Complete


CHAPTER 4

76


4.6 Conclusions

The charge transport and photogeneration in PCPDTBT:PCBM solar cells isstudied to gain insight into the loss mechanisms in these devices. The hole

transport in the polymer phase has been shown to be unaffected upon blending with

fullerenes, with a mobility of 5.5 x 10-8

m2 /Vs. The electron mobility of PCBM in

the blend has been determined to be 7x10-8

m2 /Vs, which is slightly lower than the

pristine value for PCBM. Thus the electron and hole transport are almost balanced

and the mobilities are sufficiently high to reach high fill factors and efficiencies.

Nevertheless, the fill factor of PCPDTBT:PCBM solar cells is relatively low,originating from a square root regime in the photocurrent as a function of effective

voltage. Where in chapter 2 this square root dependence was shown to be due to an

unbalance in charge carrier mobilities this is not the case here. The photocurrent is

shown to be recombination limited, characterized by a square-root dependence on

effective applied voltage, a linear dependence on light intensity and a constant

saturation voltage. Simulations of the photocurrent show that the solar cells are

limited by a short lifetime of bound electron-hole pairs. It is suggested that this

short lifetime is due to an unfavorable morphology where donor and acceptor are

too intimately mixed.

7/31/2019 10 Complete




77

REFERENCES

[1] A. Hadipour, B. de Boer and P. W. M. Blom, , Adv. Funct. Mater., 18, 169, (2008).

[2] S. E. Shaheen, D. Vangeneugden, R. Kiebooms, D. Vanderzande, T. Fromherz, F.

Padinger, C. J. Brabec, N. S. Sariciftci, Synth. Met. 2001, 121, 1583.

[3] C. Winder, G. Matt, J. C. Hummelen, R. A. J. Janssen, N. S. Sariciftci, C. J.

Brabec, Thin Solid Films 2002, 403 – 404, 373.

[4] A. Dhanabalan, J. K. J. van Duren, P. A. van Hal, J. L. J. van Dongen, R. A. J.

Janssen, Adv. Funct. Mater. 2001, 11, 255.

[5] A. P. Smith, R. R. Smith, B. E. Taylor, M. F. Durstock, Chem. Mater. 2004, 16 ,

4687.

[6] X. Wang, E. Perzon, F. Oswald, F. Langa, S. Admassie, M. R. Andersson, O.Inganäs, Adv. Funct. Mater. 2005, 15, 1665.

[7] X. Wang, E. Perzon, J. L. Delgado, P. de la Cruz, F. Zhang, F. Langa, M.

Andersson, O. Inganäs, Appl. Phys. Lett. 2004, 85, 5081.

[8] F. Zhang, E. Perzon, X. Wang, W. Mammo, M. R. Andersson, O. Inganäs, Adv.Funct. Mater. 2005, 15, 745.

[9] L. M. Campos, A. Tontcheva, S. Günes, G. Sonmez, H. Neugebauer, N. S.

Sariciftci, F. Wudl, Chem. Mater. 2005, 17 , 4031.

[10] D. Mühlbacher, M. Scharber, M. Morana, Z. Zhu, D. Waller, R. Gaudiana, C.

Brabec, Adv. Mater. 2006, 18, 2884.

[11] M. D. Irwin, D. B. Buchholz, A. W. Hains, R. P. H. Chang, T. J. Marks, PNAS2008, 8, 2783

[12] V. D. Mihailetchi, J. Wildeman, P.W. M. Blom, Phys. Rev. Lett. 2005, 94, 126602

[13] C. Tanase, P. W. M. Blom, and D. M. de Leeuw , Phys. Rev. B 2004 70, 193202.

[14] M. Morana, P. Koers, C. Waldauf, M. Koppe, D. Muehlbacher, P. Denk, M.Scharber, D. Waller, C. Brabec, Adv. Funct. Mater. 2007, 17 , 3274.

[15] 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.

7/31/2019 10 Complete


CHAPTER 4

78



Mater 2006 , 16 , 699.

[17] M. M. Mandoc, B. de Boer, P. W. M. Blom, Physical Review B, 2007, 75, 193202

[18] V. D. Mihailetchi, L. J. A. Koster, P. W. M. Blom, C. Melzer, B. de Boer, J. K. J.van Duren, R. A. J. Janssen, Adv. Funct. Mater. 2005, 15, 795.

[19] V.D. Mihailetchi, H. Xie, L.J.A. Koster, B. de Boer, L.M. Popescu, J.C.Hummelen, P.W.M. Blom, , Appl. Phys. Lett. 2006, 89, 012107

[20] V. D. Mihailetchi, L. J. Koster, J. C. Hummelen, P. W. Blom, Phys. Rev. Lett.2004, 93, 216 601.

[21] M. Morana, M. Wegscheider, A. Bonanni, N. Kopidakis, S. Shaheen, M. Scharber,

Z. Zhu, D. Waller, R. Gaudiana and C. J. Brabec, Adv. Funct. Mat , 2008, 18, 1757

[22] A. M. Goodman, A. Rose, J. Appl. Phys. 1971, 42, 2823.

[23] C. L. Braun, J. Chem. Phys. 1984, 80, 4157.

[24] L. J. A. Koster, E. C. P. Smits, V. D. Mihailetchi, P. W. M. Blom, Phys. Rev. B

2005, 72, 085 205.

[25] I.W. Hwang, C. Soci, D. Moses, Z. Zhu, D. Waller, R. Gaudiana, C. J. Brabec, A.J.

Heeger Adv. Mater. 2007, 19, 2307.

[26] Z Zhu, D Waller, R. Gaudiana, M. Morana, D. Muhlbacher, M.Scharber,C.J.Brabec, Macromolecules 2007 , 40, 1981.

[27] J. Peet, J. Y. Kim,, N. E. Coates, W. L. Ma, D. Moses, A. J. Heeger, G. C. Bazan,

Nature Materials 2007, 6, 497.

7/31/2019 10 Complete



Chapter 5

Higher adduct fullerenes for enhanced open

circuit voltage and efficiency in polymer solar cells Abstract

The bisadduct analog of PCBM, bisPCBM, is investigated which has a significant

lower electron affinity as compared to the standard acceptor PCBM. By this raise

of the LUMO level the energy loss in the electron transfer from donor to acceptor

material is reduced, manifesting itself as an increase of the open circuit voltage of

polymer:fullerene bulk heterojunction solar cells. Maintaining high currents and

fill factors an externally verified power conversion efficiency of 4.5% is achieved

for a P3HT:bisPCBM solar cell, 20% higher as compared to the efficiencies of

P3HT:PCBM cells, clearly showing bisPCBM to be the superior acceptor

compared to standard PCBM. Next to bisPCBM, other higher adduct fullerenes

are investigated, including C 70 and thienyl based materials. It is shown that theoccurrence of a multitude of different isomers results in a decrease in charge

carrier transport in single carrier devices for some of the materials. Surprisingly,

the solar cell characteristics are very similar for all materials. This apparent

discrepancy is explained by a significant amount of shallow trapping occurring in

the fullerene phase which does not hamper the solar cell performance due the

filling of these shallow traps during illumination. Furthermore, the trisadduct

analogue of [60]PCBM is investigated which, despite an even further increase in

open circuit voltage, results in a significantly reduced device performance due to a

strong deterioration of the electron mobility in the fullerene phase.

REFERENCES

M. Lenes, G. A. H. Wetzelaer, F. B. Kooistra, S. C. Veenstra,J. C. Hummelen, P. W. M. Blom

Adv. Mater . 2008, 20, 2116

M. Lenes, S. W. Shelton, A. B. Sieval, D. F. KronholmJ. C. Hummelen, P. W. M. Blom

Adv. Funct. Mat. Published Online

7/31/2019 10 Complete


CHAPTER 5

80


5.1 Introduction

In the first chapter of this thesis the prototypical polymer:fullerene system

P3HT:PCBM was introduced. With a record efficiency of 5.4%1

this material

system is approaching efficiencies warranting large scale commercialization.

Already significant research effort is put into developing large area technologies

employing these materials.2

In order for this research effort to be successful it iscrucial that materials are available on a relatively large scale. As discussed in the

introductory chapter, however, solar cells based on P3HT and PCBM are nearing

their maximum performance. Three strategies to improve the performance beyond

that of P3HT:PCBM solar cells are given. Firstly, small bandgap polymers such as

the ones used in the previous chapter can be used to enhance the light absorption of

the solar cell. Secondly, polymers with lower HOMO and LUMO levels can beused in order to increase the open circuit voltage. Lastly, acceptors with higher

LUMO levels can be employed to raise the open circuit voltage. In this chapter the

third strategy is employed. Thus far, fullerenes have always been the acceptor of

choice when making polymer solar cells. Even though polymer n-type materialshave a large potential due to the additional absorption in the acceptor, so far

efficiencies have been moderate due to problems with charge trapping,

dissociation, and phase separation.3,4

Hybrid solar cells combine polymers with

inorganic nanoparticles and are also considered to have great potential, but so far

stay behind in performance.

5

Therefore, fullerenes with higher LUMO levels arehighly desired. Changing the substituent of PCBM has shown to result in a slightly

higher Voc however, the amount of enhancement using this method is limited.6

Other fullerenes as reported thus far have not resulted in a significant improvement

compared to PCBM.7,8

In this chapter higher adduct fullerenes are investigated as a

candidate for acceptors in polymer solar cells.

7/31/2019 10 Complete




81

5.2 The bisadduct analogue of PCBM

First we introduce bisPCBM, which is the bisadduct analogue of [60]PCBM, asa new fullerene based n-type semiconductor material. BisPCBM is normally

obtained as a side product in the preparation of PCBM.9

The material used is

obtained by standard chromatographic separation from the other reaction products.

The material consists of a number of regio-isomers. The general structure of these

isomers (with the second addend at various positions on the fullerene cage) is

depicted in Fig. 5.1. BisPCBM has a substantially higher LUMO than PCBM,10

as

can be seen by cyclo voltametric (CV) comparison of bisPCBM and PCBM (Fig.5.1). An increase of the LUMO level of ~ 100 meV was found, raising the LUMO

to 3.7 eV below the vacuum level. Here the pure isomeric mixture of bisadducts

(free of monoadduct and higher adducts) was used. The bisadduct isomer mixture

is made up of a minimum of 17 isomers, as indicated by LC-MS traces. The 1H-

NMR data further indicate that the bisadducts consist of very complex mixture of

isomers, showing at least 17 methoxy resonance signals. First, layers of pristine

bisPCBM were investigated to see whether the additional functionalization of the

fullerene, and the fact that the material is made up out of a mixture of isomers,

have any negative side effects on the charge transport properties.

-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0-4

-3

-2

-1

0

1

2

3

4

5

6

O

O

O

O

PCBM

bis PCBM

C u r r e n t [ A ]

Voltage [V]

Figure 5.1: Cyclic Voltametry measurement performed on PCBM (solid line) and bisPCBM (dashedline). Experimental conditions: V vs Fc/Fc+, Bu 4NPF6 (0.1 M) as the supporting electrolyte,ODCB/acetonitrile (4/1) as the solvent, 10 mV/s scan rate. The inset shows the generalized chemical

structure of the bisPCBM regio-isomers (i.e. the bottom addend is attached in a cyclopropane mannerat various [6,6] positions, relative to the top one).

7/31/2019 10 Complete


CHAPTER 5

82


The electron transport through the fullerene was measured by sandwiching a

layer of bisPCBM between a layer of indium tin oxide (ITO) covered with ~70 nmof poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS) and a

samarium(5 nm)/aluminium(100 nm) top electrode. Since the work function of

PEDOT:PSS (5.2 eV) is significantly lower than the HOMO of bisPCBM (6.1 eV),

hole injection into the fullerene can be neglected and only electrons flow atforward bias. Figure 5.2 shows the J-V characteristic of a bisPCBM electron only

device with a thickness of 182 nm, with the applied voltage corrected for the built-

in voltage and series resistance of the contact. The transport through these single

carrier devices is space-charge limited, resulting in a low-field electron mobility of

7 x 10-8

m2 /Vs. Even though the measured electron mobility for bisPCBM is lower

compared to values reported for normal PCBM (2 x 10-7

m2

/Vs), measured underthe same conditions,

11the observed electron mobility is still expected to result in a

balanced charge transport when combined with P3HT.

0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1

100

1000

Fit

Data

J [ A

/ m 2 ]

V-Vbi-V

rs

Figure 5.2: Current density versus voltage, corrected for built in voltage and series resistance of abisPCBM electron only device. Data (symbols) is fitted (solid line) using a space-charge limited

current with a field dependent mobility.

5.2.1 P3HT:bisPCBM solar cells

Next, bisPCBM was used as an acceptor in a polymer:fullerene solar cellsusing the solvent annealing technique.

12P3HT and bisPCBM were dissolved in

1,2-dichlorobenzene (oDCB) by stirring the mixture for 2 days. The blend was spin

cast on top of ITO covered with PEDOT:PSS and left to dry in a closed petri dish

for 48 hours. After the solvent annealing a short (5 minute) thermal annealing step

7/31/2019 10 Complete




83

was done at 110o

C. To finish the devices a samarium(5 nm)/aluminium(100 nm)

top contact was evaporated. Since bisPCBM has a lower electron mobility andhigher molecular weight compared to normal PCBM, also a different optimal

composition of the blend was anticipated. Indeed, optimization showed a polymer

fullerene weight ratio of (1:1.2) to give the highest efficiencies. The optimal active

layer thickness for P3HT:bisPCBM was found to be ~ 250-300 nm. Afterfabrication the samples were evaluated and the best cells were shipped inside a

nitrogen filled container to the Energy research Centre of the Netherlands (ECN),

to accurately determine the device performance. As a reference, P3HT cells with

normal PCBM in a 1:1 weight ratio were made with the same fabrication

procedure. The optimal thickness of these cells was somewhat higher than for

bisPCBM, around 350 nm.

400 500 600 700 800 900 10000

10

20

30

40

50

60

70

80

bis PCBM

PCBM

E . Q . E .

[ % ]

Wavelength [nm]

Figure 5.3: External quantum efficiency of a P3HT:PCBM and P3HT:bisPCBM solar cell.

Figure 5.3 shows the external quantum efficiency (EQE) determined atECN for P3HT:bisPCBM and P3HT:PCBM solar cells. Even though similar in

shape normal PCBM devices result in slightly higher external quantum

efficiencies, probably due to a thicker active layer. From the EQE measurements

the short circuit current density under AM 1.5 conditions was estimated to be 96

A/m2

for P3HT:bisPCBM versus 104 A/m2

for P3HT:PCBM. Figure 5.4 shows the

J-V characteristics of the cells measured using a halogen lamp with a light output

equivalent to an AM1.5 light source with an intensity of 1.16 kW/m2. The open

circuit voltage of the P3HT:bisPCBM cell amounted to 0.73 V, which is 0.15 V

higher than the cell with P3HT:PCBM. As predicted by the EQE measurements the

short circuit current is only slightly lower for P3HT:bisPCBM. Due to the

enhanced V oc, bisPCBM is clearly the superior acceptor in combination with P3HT.

In order to accurately quote efficiencies, calibrated measurements are needed. Our

7/31/2019 10 Complete


CHAPTER 5

84


best cell was measured under a 1000 W/m2, simulated AM1.5 illumination from a

WXS-300S-50 solar simulator (WACOM Electric Co). These externally verifiedmeasurements resulted in an open circuit voltage of 0.724 V, fill factor of 68% and

a short circuit current of 91.4 A/m2. The resulting power conversion efficiency

amounts to 4.5% for the P3HT:bisPCBM solar cell with an active area of 0.16 cm2.

Devices with larger active areas of 1 cm2

showed a small decrease in fill factor to62%, resulting in efficiencies of 4.1%. The discrepancy between the calculated

short circuit current from the EQE measurements and the AM 1.5 current is

probably due to the absence of a bias illumination during the EQE measurement.

The efficiency of 4.5% is about a factor 1.2 larger as compared to the efficiencies

of our best P3HT:PCBM cells of 3.8%. This improvement is entirely due to the

increase of Voc.

-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8150

100

50

0

50

100

150

bis PCBM

PCBM

C u r r e n t

D e n s i t y [ A / m 2 ]

Voltage [V]

Figure 5.4: Current density versus voltage of P3HT:PCBM and P3HT:bisPCBM solar cells underillumination of a halogen lamp with an intensity equivalent to 1.16 sun.

7/31/2019 10 Complete




85

5.3 Other higher adduct fullerenes

In the previous paragraph, the bisadduct analog of PCBM was introduced as anew acceptor for use in polymer solar cells. Next, other higher adducts analogues

were investigated in order to see whether the concept can be extended to other

fullerenes. Furthermore, the charge transport of the various higher adducts was

studied in more detail to explain the peculiar feature that the bisadduct mixture of

isomers can be used to produce high performance solar cell. It is noted that in this

follow-up study the device performance is somewhat lower compared to the one

described above. Furthermore, a larger difference in short circuit current betweenmono and bisadducts is observed. The reason for this behaviour is the different

fabrication technique used here, based on chloroform as solvent and thermal

annealing. This fabrication technique is chosen above spin coating from

orthodichlorobenzene and solvent annealing, due to the much larger spread in

device performance of the latter, making a comparison of single carrier devices and

solar cells difficult. Furthermore, other than accounting for the difference in weight

ratio, all cells were fabricated using identical procedures and no optimisation was

done for the individual materials.

Figure 5.5: Materials used in this chapter. From top left to bottom right, regioregular poly[3-hexylthiophene] (P3HT), [60]PCBM, and highly generalized structures for the isomeric mixtures of

the bisadducts bis[60]PCBM, bis[70]PCBM, bis[60]ThCBM, bis[70]ThCBM, and the trisadducttris[60]PCBM.

S

C6H13

n

O

O

O

O

O

O

O

OS

O

O

S

O

OS

O

O

S

O

O

O

O

O

O

O

O

OO

7/31/2019 10 Complete


CHAPTER 5

86


Due to its more asymmetric shape the C70 based [70]PCBM has a higher

absorption coefficient compared to [60]PCBM which has shown to be useful forcomplementing the absorption of small bandgap polymers.

13The thienyl based

[6,6]-thienyl-C61-butyric acid methyl ester (ThCBM) has been developed to

provide a better conformity between polymer and fullerene in P3HT based

devices.14

The generalized chemical structures of these materials are shown infigure 5.5. All fullerenes were synthesised according to a procedure reported in our

previous work. Next to the standard p-type polymer P3HT and n-type molecule

[60]PCBM, the bisadduct analogues of [60]PCBM, [70]PCBM, [60]ThCBM,

[70]ThCBM and the trisadduct analogue of [60]PCBM have been investigated. In

order to study the effect of the additional fuctionalisation and the fact that a

mixture of isomers is used, electron transport measurements have been performedon blends of P3HT and fullerenes.

In figure 5.6 the J-V characteristics of electron-only devices (using an AlOx

electrode) of all P3HT:fullerene blends at room temperature are shown. The device

currents of the bisadducts are all lower compared to P3HT-PCBM blends where the

biggest difference occurs for the blend based on [70]ThCBM. For the trisadduct the

electron current is even further decreased by 3 orders of magnitude. A possible

explanation for this drop in device current can be an increase in disorder in thefullerene phase, due to the presence of a multitude of isomers of the fullerenes. In

order to determine the amount of disorder in the materials the temperature

dependence of the zero-field mobility as determined from the various electron only

devices is studied. According to the Gaussian disorder model this temperature

dependence is governed by the width of a Gaussian density of states σ following15

7/31/2019 10 Complete




87

0.0 0.5 1.0 1.5 2.0 2.5 3.0

1E-4

1E-3

0.01

0.1

1

10

100

1000

10000

100000

[60]PCBM trisPCBM

bis[60]PCBM bis[60]ThCBM

bis[70]PCBM bis[70]ThCBM

J

[ A / m 2 ]

V [V]

Figure 5.6: J-V characteristics of P3HT:methanofullerene blend electron single carrier devices.

eaE

T k T k B B

278.05

3exp

2 / 32

(5.3)

where μ∞ is the mobility as the temperature goes to infinity, a is the intersite

spacing and k B is Bolzmann‟s constant. Figure 5.7 shows the temperature

dependence of the zero-field mobility as determined from the electron onlydevices. According to Equation 5.3 the amount of disorder σ can be calculated

from the slope of the (log) mobility versus 1/T 2. For PCBM a σ of 68 meV is

determined, which agrees with the previously reported value.11

For the bisadducts

the magnitude of the disorder is significantly larger as given in the inset of figure5.7. Note that for tris[60]PCBM σ could not be determined. At low temperatures

the electron current decreased below the leakage current of the devices due to local

shorts, so that the electron mobility could not be measured in this material.

Looking at the room temperature zero-field mobilities, as compared to PCBM,

7/31/2019 10 Complete


CHAPTER 5

88


0.000010 0.000012 0.000014 0.000016 0.000018 0.000020 0.000022 0.000024

1E-15

1E-14

1E-13

1E-12

1E-11

1E-10

1E-9

1E-8

1E-7

0

[ m 2 / V s ]

1/T2

PCBM 68

bis[60]PCBM 94

bis[70]PCBM 100

bis[60]ThCBM 88

bis[70]ThCBM 129

trisPCBM

[meV]

Figure 5.7: Temperature dependence of the zero field mobility of P3HT:fullerene single carrierdevices. The Gaussian disorder model is used to determine the disorder parameter σ for the variousmethanofullerenes in the blend.

a decrease in mobility of up to 2 orders of magnitude is seen for the bisadducts, and

an even higher decrease for the trisadduct. It is expected that the device

performance of solar cells based on these materials will suffer significantly from

the much lower mobility due to space-charge formation, additional recombinationlosses and a lower dissociation probability of the bound electron-hole pairs.

16,17,18

Using the numerical program, we have performed simulations in order to analyze

the effect such a lowering of the mobility has on the device performance. In figure

5.8 the J-V characteristics of a P3HT:PCBM reference device is shown. Using theelectron mobility for PCBM determined above and typical simulation parameters

as previously reported,19 the J-V characteristics are described adequately. After

accounting for the increase in V oc of the bisadducts and trisadducts the J-V

characteristic is then calculated using the mobilities determined using the electron-

only devices. As expected, due tothe lower electron mobilities the calculated short

circuit current and fill factor decrease dramatically resulting in a predicted drop inefficiency of up to 60% for bis[70]ThCBM.

7/31/2019 10 Complete




89

-0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

-120

-100

-80

-60

-40

-20

0

20

Data

PCBM

bis[60]ThCBM

bis[60]PCBM

bis[70]PCBM

bis[70]ThCBM

trsiPCBM

J L [ A / m 2 ]

V [V]

Figure 5.8: Predictions of solar cell characteristics for all fullerenes. First the standard P3HT:PCBM J-V curve (symbols) is fitted using the numerical program (solid line). Next the J-V curve for theother fullerenes is calculated taking into account the lower mobility as determined in figure 3 and theincrease in Voc of the bis and tris adducts.

5.3.1 Solar cells based on higher adduct fullerenes

Next, bulk heterojunction solar cells were fabricated using the methanofullerenesintroduced above. Figure 5.9 shows the J-V characteristics of all solar cells at room

temperature under simulated AM1.5 illumination. As discussed in the first part of

this chapter, the raised LUMO level of the bisadducts result in a significantenhancement of the open circuit voltage of the devices. What is very surprising

however, is that all bisadducts show an almost identical device performance in

contradiction to the predicted performance shown in Fig. 5.8. Apparently, the

lower electron currents as seen in the electron-only devices are not at all reflectedin the performance of the solar cells. For the trisadduct however, despite the even

further enhanced open circuit voltage(which is among the highest reported for a

P3HT based device)20

the power conversion efficiency drops dramatically as

predicted from the deteriorated electron transport. Another remarkable feature, as

can be seen in Figure 5.10, is that a difference in device performance of the solar

7/31/2019 10 Complete


CHAPTER 5

90


cells between monoadducts and bisadducts reappears when cooling the samples

below room temperature. Furthermore, the solar cells which show a strongertemperature dependence of their performance are those which gave low device

currents in the electron-only devices. These observations strongly suggest that the

transport in the fullerene phase is hampered by a large amount of shallow trapping.

When shallow traps are present the J-V characteristics at low voltages are alsodescribed by a quadratic dependence on voltage, given by

-0.2 -0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

-100

-80

-60

-40

-20

0

PCBM

bis[60]PCBM

bis[70]PCBM

bis[60]ThCBM

bis[70]ThCBM

trisPCBM

J L

[ A / m 2 ]

V [V]

Figure 5.9: J-V characteristics of P3HT:fullerene solar cells under illumination of a simulated AM1.5irradiation with an equivalent of 1.4kW/m2.

(5.1)

with

(5.2)

and N c the effective density of states, N t the amount of traps and E t the trap-depth. In this case θμ represents an effective mobility, that contains the ratio

of free and trapped charges. The relatively low electron-only currents for a

3

2

int0

8

9

L

V J r

T k

E

N

N

B

t

t

c exp

7/31/2019 10 Complete




91

number of bisadducts are in that case due to the fact that many electrons are

immobile because they are trapped in shallow traps. For the solar cells,

during illumination a number of trap states will be filled, leading to an

enhanced transport and the device operation then approaches the one of the

trap-free PCBM device. Such an illumination dependent transport has

recently been observed in n-type polymers..21

Further evidence for chargetrapping in solar cells can be obtained from the intensity dependence of the

open circuit voltage of the devices.22

For trap-free polymer:fullerene solarcells, when plotting the Voc versus the natural logarithm of the light

intensity, the slope of the Voc follows S=(k BT/q), where k B is the Bolzmann

constant, T is the temperature and q is the elementary charge. In the case of recombination with trapped charges, however, the intensity dependence of

the V oc is enhanced. Figure 5.11 shows the Voc dependence on light intensityfor our devices. Again, the fullerenes which exhibit lower electron currents

in the electron only devices and a stronger temperature dependence in the

solar cells, show a larger dependence of the Voc versus light intensity.

7/31/2019 10 Complete


CHAPTER 5

92


300 280 260 240 220 2000

4

8

12

16

20

24

28

32

36

40

300 280 260 240 220 2000.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

300 280 260 240 220 2000

10

20

30

40

50

60

70

80

90

300 280 260 240 220 2000.3

0.4

0.5

0.6

PCBM

bis[60]PCBM

bis[70]PCBM

bis[60]ThCBM

bis[70]ThCBM

M

a x i m u m P o w e r P i n t [ W / m 2 ]

Temperature [K]

PCBM

bis[60]PCBM

bis[70]PCBM

bis[60]ThCBM

bis[70]ThCBM

O p e n C i r c u i t V o l t a g e [ V ]

Temperature [K]

PCBM

bis[60]PCBM

bis[70]PCBM

bis[60]ThCBM

bis[70]ThCBM

PCBM

bis[60]PCBM

bis[70]PCBM

bis[60]ThCBM

bis[70]ThCBM

S h o r t C i r c u i t C u r r e n t [ A / m 2 ]

Temperature [K]

F i l l F a c t o r

Temperature [K]

Figure 5.10: Solar cell parameters; maximum power point (MPP), short circuit current density (J sc),

open circuit voltage (Voc), and fill factor (FF) of P3HT:fullerene solar cells as a function of

temperature.

7/31/2019 10 Complete




93

7,38906 20,08554 54,59815 148,41316 403,42879 1096,63316 2980,95799

0,45

0,50

0,55

0,60

0,65

0,70

0,75

0,80

0,85

0,90

0,95

PCBM

bis[60]PCBM

bis[70]PCBM

bis[60]ThCBM

bis[70]ThCBM

tris[60]PCBM

1.28

1.30

1.30

1.40

1.55

1.57

V o c

[ V ]

Intensity [W/m2]

slope [kT/q]

Figure 5.11: Open circuit voltage versus the natural logarithm of the intensity of P3HT:fullerene solarcells. The slope of the Voc vs. intensity in units of [kT/q] is given in the legend.

5.3.2 Device simulations using charge trapping

In order to quantitatively describe the electron only devices and solar cells the

effect of charge trapping on the simulations is incorporated. Figure 5.12 shows the J-V characteristics of a P3HT:bis[70]PCBM electron-only device on a double log

scale. When modelling these electron currents, it is assumed that the mobility of

bis[70]PCBM is identical to reference [60]PCBM. We note that field effect

transport studies have shown the mobility of [60]PCBM and [70]PCBM to be equal

within experimental error.23

Next, we introduce shallow traps that are exponentially

distributed in energy. We observe that a relatively narrow distribution in energy, asexpected for (random) disorder, gives better results than only a discrete trap level.

The width of the distribution is governed by a trap temperature T trap=340K.22

For

such an exponential trap distribution the effective number of traps has been shown

to vary with temperature with exp{-[1/2(σ 2

/kT]/kT trap}. Using this temperaturedependence of the effective number of traps (assuming σ to be 68meV asdetermined from the PCBM devices) we can describe the whole temperature range.

For the simulations of the photocurrent of the P3HT:PCBM andP3HT:bis[70]PCBM solar cells in figure 5.12, we again start with our description

of the P3HT:PCBM reference device. Using the same set of parameters, we

subsequently add the enhanced open circuit voltage, and the trap distribution as

determined from the electron-only device of the P3HT:bis[70]PCBM blend. As canbe seen in figure 5.13, the incorporated trap distribution indeed does not lower the

device performance and we can describe the J-V characteristics adequately. As a

7/31/2019 10 Complete


CHAPTER 5

94


result the occurrence of shallow traps simultaneously explains the reduced electron

currents and the relatively good solar cell performance, due to filling of these trapsunder illumination. The nature of the shallow trapping is likely to be one or more

specific bisadduct isomers with lower lying LUMO‟s. Although we expect that

certain single isomers of bisadducts can show improved performance, the fact that

the mixture of isomers can be used as such, and that it still results in a properdevice operation is a great benefit for commercialisation of polymer:fullerene solar

cells.

0.1 1

0.01

0.1

1

10

100

1000

T [K]

295

270

250

230

210

J [ A / m 2 ]

V-Vbi-V

rs

Figure 5.12: J-V characteristics, corrected for built-in voltage and series resistance of a

P3HT:bis[70]PCBM electron single carrier at different temperatures. The J-V curves are fitted usingPCBM mobilities and an exponential trap distribution.

7/31/2019 10 Complete




95

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

-120

-100

-80

-60

-40

-20

0

20

PCBM Databis[70]PCBM Data

PCBM Fit

bis[70]PCBM Fit

J L

[ A / m 2 ]

V [V]

Figure 5.13: J-V characteristics of a P3HT:PCBM and P3HT:bis[70]PCBM solar cell. The J-V curve

(symbols) are fitted with our numerical program where for the P3HT:bis[70]PCBM cell anexponential trap distribution is introduced.

7/31/2019 10 Complete


CHAPTER 5

96


5.4 Conclusions

A novel type of fullerene, bisPCBM, with a higher LUMO level compared to thatof PCBM, is used in order to minimize the energy loss in the electron transfer from

the donor to the acceptor material in bulk heterojunction solar cells. The additional

functionalization of the fullerene cage in bisPCBM was shown to have little

negative influence on the charge-carrier properties of the fullerene. As predicted,

the higher LUMO resulted in a significantly enhanced open-circuit voltage when

used in combination with P3HT, while maintaining a high short-circuit current and

fill factor. An externally verified power-conversion efficiency of 4.5% wasreported for a P3HT:bisPCBM solar cell. We showed that the bisadduct isomer

mixture, free of monoadduct and higher adducts, can be used without further

separation of the individual isomers, resulting in an enhanced cell performance

compared to that of PCBM. Furthermore, several other higher adduct fullerenes are

investigated in combination with P3HT. At first sight the higher adduct fullerenes

show signs of an enhanced disorder, reflected by a reduced current in electron-only

devices. Such an enhanced disorder however does not comply with the temperature

and intensity dependence of the solar cells. Instead, a substantial amount of shallow

trapping is likely to be the cause of the reduced currents in the electron-only

devices. Under illumination these trap states are filled and normal solar cell

operation is observed. An exponential trap distribution has been shown to

adequately describe both electron only and solar cell. The nature of the shallowtrapping is likely to be specific bisadduct isomers with lower lying LUMO‟s. Thefact that the mixture of isomers can be used as such, and still results in a proper

device operation is a great benefit for commercialisation of polymer:fullerene solar

cells. The trisadduct analogue of PCBM however, despite leading to a high open

circuit voltage of 813 mV, results in a significantly reduced device performancedue to a deterioration of the charge transport in the fullerene.

7/31/2019 10 Complete




97

[1] M. D. Irwin, D. B. Buchholz, A. W. Hains, R. P. H. Chang, T. J. Marks, PNAS,2008, 8, 2783.

[2] C. J. Brabec, J.D Durrant MRS Bul. 2008, 33, 607.

[3] J. J. M. Halls, C. A. Walsh, N. C. Greenham, E. A. Marseglia, R. H. Friend, S. C.Moratti, A. B. Holmes, Nature 1995, 376 , 498.

[4] C. R. McNeill, A. Abrusci, J. Zaumseil, R. Wilson, M. J. McKiernan, J. H.

Burroughes, J. J. M. Halls, N. C. Greenham, R. H. Friend, Appl. Phys. Lett . 2007,

90, 193506.

[5] D. J. D. Moet, L. J. A. Koster, B. de Boer, P. W. M. Blom, Chemistry of Materials,2007, 19, 5856.

[6] F. B. Kooistra, J. Knol, F. Kastenberg, L. M. Popescu, W. J. H. Verhees, J. M.Kroon, J. C. Hummelen, Org. Lett . 2007, 9, 551.

[7] I. Riedel, E. von Hauff, J. Parisi, N. Martín, F. Giacalone, V. Dyakonov, Adv.

Funct. Mat. 2006, 15, 1979.

[8] S. Backer, K. Sivula, D. F. Kavulak, J. M. J. Fréchet, Chem.Mater . 2007, 19, 2927.

[9] J. C. Hummelen, B. W. Knight, F. Lepec, F. Wudl, J. Yao, C. L. Wilkins, J. Org.Chem. 1995, 60, 532.

[10] F. Diederich, R. Kessinger, Acc. Chem. Res. 1999, 32, 537.

[11] 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.

[12] G. Li, V. Shrotriya, J. Huang, Y. Yao, T. Moriarty, K. Emery, 2005, Y. Yang, Nat.

Mater. 2005, 4, 864.

[13] D. Mühlbacher, M. Scharber, M. Morana, Z. Zhu, D. Waller, R. Gaudiana, C.

Brabec , Adv. Mater . 2006, 18, 2884.

[14] L. M. Popescu, P. van‟t Hof, A. B. Sieval, H. T. Jonkman, J. C. Hummelen, Appl.

Phys. Lett. 2006, 89, 213507.

[15] H. Bassler, Phys. Stat. Sol. B. 1993, 175, 15.

[16] V. D. Mihailetchi, J. Wildeman, P.W. M. Blom, Phys. Rev. Lett . 2005, 94, 126602.

7/31/2019 10 Complete


CHAPTER 5

98


[17] V. D. Mihailetchi, L. J. Koster, J. C. Hummelen, P. W. Blom, Phys. Rev. Lett .2004, 93, 216 601.

[18] M. Lenes, L. J. A. Koster, V. D. Mihailetchi, P. W. M. Blom, Appl. Phys. Lett .2006, 88, 243502.


Mater 2006, 16 , 699.

[20] R. B. Ross, C. M. Cardona, D. M. Guldi, S. G. Sankaranarayanan, M. O. Reese, N.

Kopidakis, J. Peet, B. Walker, G. C. Bazan, E. Van Keuren, B. C. Holloway, M.

Drees, Nature Materials 2009, 8, 208.

[21] C. R. McNeill, N. C. Greenham, Appl. Phys. Lett . 2008, 93, 203310.

[22] M. M. Mandoc, F. B. Kooistra, J. C. Hummelen, B. de Boer, P. W. M. Blom, Appl.

Phys. Lett . 2007 , 91, 263505.

[23] P. H. Wobkenberg, D. D. C. Bradley, D. Kronholm, J. C. Hummelen, D. M. deLeeuw, M. Colle, T. D. Anthopoulos, Synth. Met . 2008, 11, 468.

7/31/2019 10 Complete


99


PUBLICATIONS

M. Lenes, S. W. Shelton, A. B. Sieval, D. F. Kronholm, J. C. Hummelen and P.

W. M. Blom ”Electron trapping in higher adduct fullerene- based solar cells” Adv. Funct. Mat. Published online (DOI: 10.1002/adfm.200900459)

D. J. D. Moet, M. Lenes, J. D. Kotlarski, S. C. Veenstra, J. Sweelssen, M. M.

Koetse, B. de Boer, P. W. M. Blom “Impact of molecular weight on charge

carrier dissociation in solar cells from a polyfluorene derivative” Org. Electronics, 2009, 10, 1275

M. Lenes, M. Morana, C. J. Brabec, P. W. M. Blom “Recombination-limited

photocurrents in low bandgap polymer:fullerene solar cells” Adv. Funct. Mat.

2009, 19, 1106

M. Lenes, F. B. Kooistra, J. C. Hummelen, I. van Severen, L. Lutsen, D.

Vanderzande, T. J. Cleij, P. W. M. Blom “Char ge dissociation inpolymer:fullerene bulk heterojunction solar cells with enhanced permittivity” J.

Appl. Phys. 2009, 104, 114517.

R. Kroon, M. Lenes, J. C. Hummelen, P. W. M. Blom, B. de Boer “Small bandgap polymers for organic solar cells” Polymer Reviews, 2008, 48, 531

M. Lenes, G. A. H. Wetzelaer, F. B. Kooistra, S. C. Veenstra, J. C. Hummelen,

P. W. M. Blom “Fullerene Bisadducts for Enhanced Open-Circuit Voltages and

Efficiencies in Polymer Solar Cells” Adv. Mat. 2008, 20, 2116

J. D. Kotlarski, P. W. M. Blom, L. J. A. Koster, M. Lenes, L. H. Slooff

“Combined optical and electrical modelling of polymer : fullerene bulk heterojunction solar cells” J. Appl. Phys. 2008, 103, 084502.

M. Lenes, L. J. A. Koster, V. D. Mihailetchi, P. W. M. Blom “Thicknessdependence of the efficiency of polymer : fullerene bulk heterojunction solar

cells” Appl. Phys. Lett. 2006, 88, 243502.

S. Steudel, S. De Vusser, K. Myny, M. Lenes, J. Genoe, P. Heremans

“Comparison of organic diode structures regarding high-frequency rectificationbehavior in radio-frequency identification tags” J. Appl. Phys. 2006, 99, 114519.

7/31/2019 10 Complete


PUBLICATIONS

100


L. J. A. Koster, V. D. Mihailetchi, M. Lenes, P. W. M. Blom “PerformanceImprovement of Polymer: Fullerene Solar Cells Due to Balanced Charge

Transport”, Organic Photovoltaics. Materials, Device Physics, and

Manufacturing Technologies. Edited by C.J.Brabec, U. Scherf, and V.

Dyakonov, WILEY-VCH, Weinheim, ISBN: 978-3-527-31675-5.

M. Lenes, V. D. Mihailetchi, L. J. A. Koster, and P. W. M. Blom, “Space-

charge formation in thick MDMO-PPV:PCBM solar cells”, Proceedings of SPIE

6192, 120 (2006).

D. Jarzab; F. Cordella, M. Lenes, F. B. Kooistra, P. W. M. Blom, J. C.Hummelen, M. A. Loi “Charge Transfer Dynamics in Polymer-Fullerene Blends

for Efficient Solar Cells" Submitted for Publication.

M. Kuik, H. T. Nicolai, M. Lenes, M. Lu, P. W. M. Blom “Introducing trap-

assisted recombination in polymer light emitting diodes” Manuscript in

preparation.

H. Diliën, A. Palmaerts, M. Lenes, B. de Boer, P. W. M. Blom, T. J. Cleij, L.

Lutsen, D. Vanderzande “Soluble poly(thienylene vinylene) derivatives for photovoltaic applications” Manuscript in preparation.

7/31/2019 10 Complete


101


SUMMARY

The main goal in organic photovoltaics is the development of a large-area, flexible,

and most importantly, a low-cost energy source. The materials used in this thesis,conjugated polymers and fullerene derivatives, can be made soluble enabling low

temperature processing techniques such as spin-coating, doctor blading and ideally

roll-to-roll fabrication (think of solar cells being printed at high speed, similar to

newspapers). On the downside, the inherently disordered nature of the used

materials and processing conditions leads to devices with inferior charge carrier

mobilities compared to their inorganic counterparts such as silicon. This in turn

places the research field of organic electronics in the area of low frequency low

performance devices. The balance between production cost, device lifetime and

efficiency will in the end determine the viability of organic solar cells.

In this thesis that latter part of this balance between cost, lifetime and

efficiency is investigated. In the first introductory chapter the current

understanding of the working principles of polymer fullerenes bulkheterojunction

(BHJ) solar cells is discussed. Using this information, the main loss mechanisms in

this type of devices are identified. It is shown that the intrinsic low mobility of the

polymer and fullerene does in fact not significantly limit the device performance,

as long as they are well balanced. This is reflected by the high (~90%) internal

quantum efficiencies achieved in, for instance, devices made from a polythiophene

(P3HT) and a methanofullerene (PCBM). Nevertheless these devices only achieve

power conversion efficiencies of typically 4%. The question then arises whichprocesses are responsible for the efficiency losses. It is shown that a significant

amount of energy is lost due to misalignment of the energy levels of the used

materials. In polymer solar cells a donor-acceptor (D-A) system is used in order to

separate excitons into free carriers. Unfortunately, during the transfer of the

electron from the lowest unoccupied molecular orbital (LUMO) of the donor to theLUMO of the acceptor, energy is inevitably lost. This loss in energy is manifested

in the low open circuit voltage of a D-A BHJ solar cell compared to the bandgap of

the absorber. Three strategies are suggested in this thesis to reduce this offset,

either resulting in an expected increase in the amount of absorbed light or an

increase of the output voltage of the solar cell.

The second and third chapter of this thesis focuses on poly(p-phenylene

vinylene) (PPV) type polymers, a class of materials heavily used and studied inpolymer electronics. The switch to chlorobenzene as a solvent for spincoating

MDMO-PPV:PCBM layers led to the first reasonable performing device, achieving

an efficiency of 2.5%. Devices based on MDMO-PPV:PCBM blends typically

have an active layer thickness of 100 nm at which still a significant portion of light

is not absorbed. In chapter 2 of this thesis the origin of the use of such a relatively

thin active layer is investigated. It is shown that the decrease in fill factor which,from a device point of view, is the origin for the decreasing efficiency upon

increasing the active layer thickness, is due to a combination of space-charge

effects, a decreasing dissociation efficiency and charge recombination.

7/31/2019 10 Complete


SUMMARY

102


Besides the thin active layers used in PPV solar cells, another loss

mechanism can be identified. The dissociation of a bound electron hole pair at thedonor acceptor interface has been shown to be a significant limiting factor for PPV

based devices. In chapter three a new glycol substituted PPV is investigated, which

has a higher permittivity compared to normal PPV‟s. The aim here is to increasethe above mentioned low dissociation efficiency of PPV based devices, which isstrongly dependent on the average permittivity of the active layer blend. Due to a

significant lower hole mobility of the polymer and morphology problems, devices

based on this new polymer did not show the expected improved power conversion

efficiencies compared to the model MDMO-PPV system. Nevertheless, an increase

in dissociation efficiency from 60 to 72% was observed for the enhanced

permittivity polymer, indicating the importance of the average permittivity inpolymer:fullerene devices.

As mentioned above, optimizing the LUMO level offset between donor and

acceptor is one of the most forward ways of increasing device performance. By

lowering the LUMO of the donor in theory more light can be absorbed resulting in

an increased device current. One of the most promising materials following this

route is PCPDTBT. In chapter 4 the charge transport and photogeneration of this

material blended with PCBM is investigated. Despite an almost balanced transportthe photocurrent shows a square root dependence on effective voltage. It is shown

that this square root dependence does not stem from an unbalance in mobilities as

is seen in chapter 2, but from an enhanced recombination of the bound electron

hole pair. This enhanced recombination is likely due to a too close intermixing of

polymer and fullerene.In chapter 5 another route towards efficiency enhancement is investigated.

Instead of lowering the LUMO of the donor now the LUMO of the acceptor israised allowing for a very direct enhancement of the efficiency due to a larger open

circuit voltage. In order to achieve this raising of the LUMO level, the bisadduct

analog of PCBM was used. The additional functionalisation of the fullerene cage

leads to a saturation of the double bonds, raising the LUMO level of the molecule

significantly. It is shown that, despite the additional functionalisation that increases

the disorder (due to having a multitude of isomers of the molecule), replacing

PCBM with bisPCBM results in only a very slightly decreased photogeneration

and transport properties. Combined with a significantly enhanced open circuit

voltage a power conversion efficiency of 4.5% was achieved, a relative increase of

20% compared to PCBM and among the highest reported for P3HT based

polymer:fullerene solar cells. In the second part of chapter 5, bis- and trisadduct

analogs of other fullerenes are investigated. It is shown that the existence of

multiple isomers leads to shallow trapping for single carriers devices, which do not

affect the device operation of the solar cells itself.

In conclusion, existing device models were used to identify limiting factors

for the power conversion efficiency of polymer solar cells. Several strategies are

employed in order to reduce these limits. In some cases the proposedimprovements worked as expected, only to be countered by unexpected side effects

7/31/2019 10 Complete




103

(chapter 3). In other cases new physical phenomena were observed (a lifetime

limited photocurrent, chapter 4). A breakthrough in performance is described inchapter 5, where the proposed raising of the LUMO level of the acceptor resulted

in an increased open circuit voltage and no negative side effects and hence a

significant improved efficiency, exactly as predicted.

With the results presented in this thesis, another step is made towards theunderstanding of the device physics of polymer solar cells, and higher efficiencies

are achieved. Maybe it is time to focus on the other side of the balance, to actually

make the devices low-cost and long living.

7/31/2019 10 Complete


SAMENVATTING

104


SAMENVATTING

Het voornaamste doel van de organische fotovoltaïsche technologie is het

ontwikkelen van een grote schaal, flexibele en vooral goedkope energiebron. De

materialen beschreven in dit proefschrift, geconjugeerde polymeren en fullereen

derivaten, kunnen oplosbaar gemaakt worden en zijn hierdoor geschikt voor lage

temperatuur processen zoals spin coaten, doctor blading en idialiter roll-to-roll

fabricatie (denk hierbij aan zonnecellen die als kranten op hoge snelheid gedrukt

worden). Helaas geldt ook voor deze materialen dat ze inherent wanordelijk zijn,

mede door deze fabricage technieken, wat leidt tot lage ladingsdrager mobiliteiten

vergeleken met hun inorganische tegenhangers zoals silicium. Dit plaatst het

vakgebied van de organische halfgeleiders in een gebied van lage frequentie, lage

performance toepassingen. Uiteindelijk zal de balans tussen kosten, levensduur en

efficiëntie de waarde van de organische zonnecel moeten bepalen.

Deze thesis zal het laatste onderdeel van deze balans, de efficiëntie,behandelen. In het inleidende eerste hoofdstuk wordt de huidige stand van kennis

met betrekking tot de principes van polymere fullerene bulk heterojunctie (BHJ)

zonnecellen behandeld. Met deze informatie worden de voornaamste

verliesprocessen geïdentificeerd. Er wordt aangetoond dat de lage ladingsdrager

mobiliteiten in feite de efficiëntie niet significant beïnvloed, mits ze gebalanceerd

zijn. Dit komt naar voren in de hoge interne efficiëntie gehaald in bijvoorbeeld

zonnecellen gemaakt van een mix van polythiophene (P3HT) en een

methanofullerene (PCBM). Desalniettemin behalen dit soort devices slechts eenefficiëntie van rond de 4%. De vraag is dan welke processen er dan

verantwoordelijk zijn voor het energieverlies. Er wordt getoond dat een significant

deel van de energie verloren gaat door een onvoordelige afstand tussen de

energieniveau‟s van de gebruikte materialen. In polymere zonnecellen wordt

gebruik gemaakt van een zogeheten donor-acceptor systeem om excitonen in vrijeladingsdragers te om te zetten. Helaas gaat er tijdens de overdracht van een

electron van de laagste ongevulde moleculaire orbitaal (LUMO) van de donor naar

die van de acceptor energie verloren. Dit verlies in energie manifesteert zich in de

lage openklemspanning in vergelijking met de bandgap van het absorberendmateriaal. Drie strategiën om dit verlies van energie te verminderen worden

voorgesteld, waarvan ofwel wordt verwacht dat de hoeveelheid geabsorbeerd licht

wordt verhoogd, ofwel een verhoging van het geleverde voltage wordt verhoogd.Het tweede en derde hoofdstuk van deze thesis richten zich op poly(p-

phenylene vinylene) (PPV) type polymeren, een veel gebruikte en bestudeerde

klasse materialen in de polymere electronica. Door gebruik te maken van

chlorobenzeen als oplosmiddel tijdens het spincoaten van MDMO-PPV:PCBMlagen zijn in het verleden de eerste relatief efficiënte devices gemaakt met een

efficiëntie van 2.5%. Zonnecellen gebaseerd op mengsels van deze twee materialen

hebben typisch een actieve laag van 100nm dik, welke nog een significant deel van

het zonlicht doorlaat. In hoofdstuk twee wordt de reden voor het gebruik van een

7/31/2019 10 Complete




105

dergelijk dunne actieve laag bestudeerd. De daling van de vulfactor, wat vanuit een

device oogpunt de oorzaak van de daling in efficiëntie is, blijkt veroorzaakt dooreen combinatie van ruimtelading effecten, een verminderde dissociatie efficiëntie

en ladings recombinatie.

Buiten de dunne actieve laag in PPV zonnecellen kan nog een ander

verliesmechanisme aangeduid worden. Aangetoond is dat de dissociatie vangebonden electron-gat paren aan het donor acceptor interface een belangrijk

verliespad is in PPV gebaseerde zonnecellen. In hoofdstuk drie wordt een nieuw

glycol gesubstitueerde PPV gebruikt met een hogere diëlectrische constante in

vergelijking met normale PPV‟s. Het doel is nu de bovengenoemde dissociatieefficiëntie te verhogen, welke sterk afhankelijk is van de gemiddelde dielectrische

constante van de actieve laag. Door een significant lagere gatenmobiliteit enmorfologie-problemen gaven de zonncellen gebaseerd op dit polymeer niet de

verwachte winst in efficiëntie. Desalniettemin werd een verhoging van de

dissociatie efficiëntie van 60 naar 72% geobserveerd voor het nieuwe polymeer,

wat het belang aangeeft van de gemiddelde dielectrische constante op de werking

van een polymere zonnecel.

Zoals hierboven beschreven, is het optimaliseren van de LUMO energie

niveau‟s van donor en acceptor één van de meest directe manieren om de deviceefficiëntie te verhogen. Door het LUMO energieniveau te verlagen kan in theorie

meer licht door het polymeer worden geabsorbeerd, wat leidt tot een hogere device

stroom. Eén van de meest veelbelovende materialen in die categorie is PCPDTBT.

In hoofdstuk vier wordt het ladingstransport en photogeneratie van dit materiaal

gemixed met PCBM onderzocht. Ondanks een gebalanceerd transport vertoont dephotostroom een wortelafhankelijkheid ten opzichte van het effectieve voltage. Er

wordt aangetoond dat deze wortelafhankelijkheid niet komt door eenongebalanceerd transport, zoals beschreven in hoofdstuk twee, maar door een

verhoogde recombinatie van het electron gat paar. Deze verhoogde recombinatie

wordt waarschijnlijk veroorzaakt door een te fijne mix van polymeer en fullereen.

In hoofdstuk vijf wordt een andere strategie om de device efficiëntie van

een polymere zonnecel te verhogen onderzocht. In plaats van het verlagen van de

LUMO van de donor wordt nu de LUMO van de acceptor verhoogd, wat tot een

zeer directe verhoging van de efficiëntie leidt door middel van een hogere

openklemspanning. Het verhogen van de LUMO van het fullereen wordt

bewerkstelligd door gebruik te maken van de bisadduct analoog van PCBM. De

additionele functionalisering van het fullereen leidt tot een saturatie van het aantal

dubbele bindingen wat het LUMO energieniveau van het molecuul aanzienlijk

verhoogt. Er wordt aangetoond dat, ondanks de extra functionalisatie, welke de

wanorde in het systeem verhoogd (doordat er nu meerdere isomeren van het

molecuul bestaan) het vervangen van PCBM met bisPCBM slechts in zeer kleine

mate de kortsluitstroom en transport van de zonnecel vermindert. Gecombineerd

met een significant verhoogde openklemspanning werd een efficiëntie van 4,5%

behaald, relatief 20% hoger in vergelijking met PCBM, en één van de hoogstgerapporteerde efficiënties voor een P3HT gebaseerde zonnecel. In het tweede deel

7/31/2019 10 Complete


SAMENVATTING

106


van hoofdtuk vijf worden bis,- en trisadduct analogen van andere fullerenen

onderzocht. Er wordt aangetoond dat het bestaan van meerdere isomeren leidt totondiepe vangstcentra voor het electronentransport die echter de zonnecel niet

beperken tijdens zijn werking.

Tot besluit; bestaande device modellen zijn gebruikt om limiterende

factoren voor de efficiëntie van polymere zonncellen de identificeren.Verschillende strategien worden behandeld om deze limiten op te heffen. In

sommige gevallen werkten de voorgestelde veranderingen als verwacht echter

werden verbeteringen gecompenseerd door onverwachte bijverschijnselen van de

veranderingen (hoofdstuk drie). In andere gevallen werden nieuwe fysische

fenomenen geobserveerd zoals een de levensduur begrensde fotostroom in

hoofdstuk vier. Een doorbraak in performance is beschreven in hoofdstuk vijf,waar de voorgestelde verhoging van het LUMO energie niveau van de acceptor

resulteerde in een verhoogde openklemspanning zonder daarbij in te boeten op

andere vlakken. De behaalde efficiëntie steeg daardoor significant, precies als

voorspeld.

Met de resultaten in deze thesis is een volgende stap gemaakt richting het

begrijpen van de device fysica van polymere zonnecellen en werden hogere

efficiënties behaald. Misschien wordt het wel tijd om naar de andere kant van depolymere zonnecel te kijken, het daadwerkelijk goedkoop maken en de levensduur

van de devices.

7/31/2019 10 Complete



107


ACKNOWLEDGEMENTS

The work described in this thesis, and the actual appearing of this thesis, would not

have been possible without the aid and support of many people. First and foremost,

I would like to thank Paul Blom for his excellent guidance over the last 4 years.

You have shown me how science should be done, always having a clear vision of

the bigger picture, starting from the first experiments right up to the presenting of

results in journals and on conferences. Groningen is going to have a hard time

finding a proper replacement.

The work in this thesis is a direct continuation of the excellent work done byValentin Mihailetchi and Jan Anton Koster. Their results, guidance in

experimenting and modeling gave the beginning of my Ph D a jumpstart like noother and I will be the last to forget how lucky I was following up their projects.

The name of Kees Hummelen has been a continuous appearance during mytime in Groningen. As a supplier of state of the art materials, co-author, a member

of the reading committee, and company during conferences en meetings. Yet I will

mostly remember you as an example of an extremely passionate scientist and your

role as ambassador for our funny plastic solar cells. Besides Kees I would like to

thank all the other bucky‟s; Floris Kooistra, Alex Sieval and David Kronholm, who

made sure I never had any shortage of acceptor materials.Besides ample acceptor materials, I also did not lack donor materials. Without

the low bandgap polymer supplied by Konarka I would not have been able to attain

the results in chapter 4. I thank Christoph Brabec and Mauro Morana, not only for

providing the polymer, but also for the thorough discussions about the results and

comments on the AFM paper. Dirk Vanderzande and Thomas Cleij I thank for their

collaboration on the PEO-PPV project as well as the other experiments through the

exchange with Arne. Arne, we had a great time in Groningen. It was sometimeshard to keep up with your desires to grab a pintje and your eierballen have reached

a legendary status. Yet it was also very nice working so close with a chemist. I

hope company life will not wear you down too much.

The work presented here was part of project#524 of the Dutch PolymerInstitute. I thank John van Haare as project leader, Sjoerd Veenstra and Jan Kroon

from ECN, and all other people from within the cluster for their cooperation.

Where Paul can be seen as the front engine of the train called MEPOS, Minte

Mulder and Jan Harkema are the engines in the back making sure the whole bunchstays together. Any you manage to run a cleanroom in the meantime. Great job!

Renate, the same thing holds for you. Thank you for getting all the paperwork for

the thesis defense in order.

Paul Heremans and Laurens Siebbeles are acknowledged for their effort and

time to read my thesis and making corrections and suggestion. Paul Heremans I

also would like to thank for introducing me to the field of organic electronics

during my industrial internship some years ago.

I was lucky to be assigned to two valuable students. The record breaking solar

cells in this thesis were made by Gertjan Wetzelaer, and I have to constantly

7/31/2019 10 Complete


ACKNOWLEDGMENTS

remind myself he was only doing a “korte stage” at that time. I enjoyed your visitto Valencia and our Spanish lessons with “Helga”. Steve, you crack me up. I havenever heard of a person ending up in the wrong place, train, plane so often and still

remain absolutely calm. I hope you had a good time in Groningen, we definitely

were glad to have you over. Also from a scientific point of view, your experiments

resulted in some nice research.Herman and Kriszty, as roomies it was only natural for you to be my

paranimfen, you have been close to me all the time during my Ph D, we might as

well finish it that way. Kuik, you have come to be such a close friend to me during

these years that I asked you to be my best man last year. I think that pretty much

says enough. Hylke, without your and Herman‟s work the last 48 hours, I couldn‟t

even think of finishing the thesis in time. I hope things work out for you in theStates, anyways we will be working together in our own company in a few years.

Did you think of a great idea yet?

It has been said many times by Hylke: “MEPOS is the best group”. He wasright. Besides the people already mentioned here I thank Afshin, Auke, Johan,

Rene, Dorota, Fatemeh, Milo, Yuan, Irina, Francesco, Magda, Edsger, Date, Mark

Jan, Eek, Alex, Paul, Ilias, Jia, Andre, Fabrizio, Ronald, Dennis, Christina, Teunis,

Maria Antonietta for the great time in Groningen.Caroline, by correcting my Dutch summary you have officially helped

finishing this thesis. If I were to write a thesis covering the rest of my life you

would be the only co-author.

There is still one person missing in these acknowledgements. Even though not

officially involved in my project, Bert de Boer has played an important role duringmy Ph. D. He was always willing to discuss the more political issues of science;

career choices, how to read between the lines of a referee comment, who to put onyour paper as co-author, how to collaborate with partners, etc. and was a huge

driving force in the group. His passing away earlier this year is an immense loss,

both inside and outside the scientific community.

Martijn Lenes

10 Complete

Documents

Transcript of 10 Complete