fullerene based organic solar cells

Fullerene based Organic Solar Cells

Lacramioara Mihaela Popescu

Zernike Institute PhD thesis series 2008-17ISSN 1570-1530

The work described in this thesis was performed in the research group Molecu-lar Electronics of the Zernike Institute for Advanced Materials at the Universityof Groningen, the Netherlands.

PRINTED BY: Drukkerij van Denderen B.V., Groningen, the Netherlands

RIJKSUNIVERSITEIT GRONINGEN

Fullerene based Organic Solar Cells

Proefschrift

ter verkrijging van het doctoraat in deWiskunde en Natuurwetenschappenaan de Rijksuniversiteit Groningen

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

vrijdag 3 oktober 2008om 16:15 uur

door

Lacramioara Mihaela Popescu

geboren op 23 april 1978te Timis, oara, Roemenie

Promotor: Prof. dr. J. C. HummelenCopromotor: Dr. H. T. Jonkman

Beoordelingscommissie: Prof. dr. O. InganasProf. dr. ir. R. A. J. JanssenProf. dr. ir. P. W. M. Blom

ISBN 978-90-367-3482-0

Familiei mele,pentru iubirea, sprijinulsi incurajarea constanta

de-a lungul aniilor

Contents

1 Introduction 1

1.1 Photovoltaics: A historical perspective . . . . . . . . . . . . . . . . 1

1.2 Organic solar cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

1.3 Characterization of organic solar cells . . . . . . . . . . . . . . . . 8

1.3.1 Determining the charge carrier mobility . . . . . . . . . . . 8

1.3.2 Characterization under illumination . . . . . . . . . . . . . 10

1.4 Motivation and outline of the thesis . . . . . . . . . . . . . . . . . 11

References and notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

2 Materials and experimental techniques 19

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

2.2 Device preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

2.2.1 Materials and solutions . . . . . . . . . . . . . . . . . . . . 20

2.2.2 Preparation of various device structures used . . . . . . . 23

2.3 Morphology and optical properties . . . . . . . . . . . . . . . . . . 26

2.4 Device characterization . . . . . . . . . . . . . . . . . . . . . . . . . 28


3 A thienyl analogue of PCBM for bulk heterojunction solar cells 33

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

3.2 Electron transport in pristine [60]ThCBM . . . . . . . . . . . . . . 35

3.3 Bulk heterojunction solar cells using [60]ThCBM as acceptor . . . 37

3.3.1 [60]ThCBM and MDMO-PPV blends . . . . . . . . . . . . . 37

3.3.2 [60]ThCBM and P3HT blends . . . . . . . . . . . . . . . . . 39

3.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43


vii

viii CONTENTS

4 Effect of the P3HT:ThCBM growth rate on solar cell performance 494.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 504.2 Morphology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 514.3 Hole transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 544.4 Solar cells characterization . . . . . . . . . . . . . . . . . . . . . . . 554.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58References and notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

5 Stability tests of P3HT:methanofullerence solar cells 635.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 645.2 Accurate efficiency measurements . . . . . . . . . . . . . . . . . . 655.3 Effect of ageing on the device performance . . . . . . . . . . . . . 66

5.3.1 Charge carrier transport . . . . . . . . . . . . . . . . . . . . 675.3.2 Solar cells parameters . . . . . . . . . . . . . . . . . . . . . 685.3.3 External quantum efficiency (EQE) . . . . . . . . . . . . . . 71

5.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73References and notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

6 Efficient PV blends comprising more than two semiconductors 776.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 786.2 Mixture of two n-type semiconductors in the active layer . . . . . 816.3 Improving open-circuit voltage using bis-adducts fullerenes . . . 846.4 Effect of impurities in the fullerenes on cell efficiency . . . . . . . 856.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88References and notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

Summary 91

Samenvatting 95

1Introduction

1.1 Photovoltaics: A historical perspective

Photovoltaics literally means light-electricity: photo comes from the Greek phos,meaning light, and volt from the Italian scientist Alessandro Volta; a pioneer inthe study of electricity. This technology has many advantages: it is modular,clean, easy to maintain, and can be installed almost anywhere to suit the needsof the user. The electricity produced can be used directly, stored locally or fedinto an existing electricity grid.

The discovery of the photovoltaic effect is ascribed to the French physicistEdmund Becquerel in 1839, who reported a photocurrent when a silver coatedplatinum electrode was illuminated in aqueous solution [1]. Forty years later,in 1873 and 1876, respectively, Smith and Adams were the first to report on ex-periments with selenium [2, 3]. By that time the first solid state photovoltaicdevices were constructed. In 1906 Pochettino [4], and in 1913 Volmer [5], dis-covered photoconductivity in anthracene. However, it was not the photovoltaicproperties of materials, like selenium which excited researchers, but the photo-conductivity. In the late 1950s and 1960s the potential use of organic materials asphotoreceptors in imaging systems was recognized [6]. The scientific interest aswell as the commercial potential led to increased research into photoconductiv-ity and related subjects. In the early 1960s it was discovered that many commondyes, as methylene blue, had semiconducting properties [7]. Later, these dyes

1

2 Chapter 1: Introduction

were among the first organic materials to exhibit the photovoltaic effect [8, 9].Most of the earlier understanding of the photovoltaic effect in organic photocellscomes from the study of devices fabricated from many biological molecules likecarotenes, chlorophylls and other porphyrins, and the structurally related ph-talocyanines.

The first fabricated inorganic solar cell was reported by Chapin, Fuller andPearson in 1954 at Bell Laboratories [10]. It was a crystalline silicon solar cell,which converted sunlight into electric current with an efficiency of 6%, six timeshigher than the best previous attempt. Over the years the efficiency of crys-talline silicon cells has reached more than 24% in a laboratory setting [11], whichis close to the theoretical upper limit of 31% [12, 13]. The technology usedto make most of the silicon solar cells fabricated so far, benefited greatly fromthe high standard of silicon technology developed originally for transistors andlately for integrated circuits.

Today, a range of photovoltaic (PV) technologies are commercially availableor are under development in the research laboratories. An overview of the cur-rent PV technologies and the efficiencies reached are presented in Table 1.1.

Wafer-based crystalline silicon has dominated the photovoltaic industrysince the dawn of the solar PV area [13, 14]. It is widely available, has a con-vincing track-record in reliability and its physical characteristics are well under-stood, in part thanks to its use in the half-century-old microelectronics industry.Purified silicon (polysilicon) is the basic ingredient of the silicon modules. It ismelted and solidified using a variety of techniques to produce ingots or ribbonswith different degrees of crystal perfection. The ingots are shaped into bricksand sliced into thin wafers. Wafers are processed into solar cells and intercon-nected in weather-proof packages designed to last for at least 25 years. For thepast few years the availability of polysilicon feedstock has been a critical issuefor the rapidly growing PV industry. The tight supply has caused very highpolysilicon spot market prices and has limited production expansion for part ofthe industry [15].

The highest solar cell efficiency has been demonstrated using III-V semicon-ductors [51]. GaAs, InP and their alloys have been used in single-junction andmulti-junction configurations. The attractive features of GaAs are its high ab-sorbtion coefficient (devices only few µm thick absorb most of the light) and itshigh temperature coefficient compared with silicon (GaAs solar cells performsbetter in situations where the cells operates at high temperature and in space).GaAs cells have been developed primarily for use in space applications wherethe priority is to reduce the cell weight. However, GaAs cells are used in ter-

1.1. Photovoltaics: A historical perspective 3

Table 1.1: Current photovoltaic technologies and their efficiencies.

Technology Photovoltaic device Conversion efficiencies and notes

Monocrystalline solarcells

Silicon solar cells 24.7% laboratory efficiency [11];20.1% commercial module effi-ciency [17]

Galium arsenide (GaAs) 25.1% laboratory efficiency [51]

Thin-film technologies

Thin-film silicon (TFSi) 9.5% (a-Si) [18]; 12% (tandem a-Si/µc-Si) [19]; 13% (triple junctionusing SiGe alloys) [20]. All these arelaboratory efficiencies

Cu(In, Ga)(S, Se)2 and re-lated I-III-VI compounds(CIGS)

19.9% laboratory efficiency [21];13.4% commercial module effi-ciency [22]

Cadmium telluride (CdTe) 16.5% laboratory efficiency [23];10.7% commercial module effi-ciency [24]

GaAs 24.5% [52]

Organic-based solarcells

Bulk-heterojunction solarcells

6.5% laboratory efficiency [26]

Dye-Sensitized cells (Graet-zel cell)

10.4% laboratory efficiency [27]

Novel PV technologies:Novel active layers

Quantum wells, Quantumwires, Quantum dots,Nanoparticle inclusion inhost semiconductor

Theoretical efficiency limits are 50-60% [28, 29]

Novel PV technologies:Boosting the structureat the periphery of thedevice

Up-down converters > 10% efficiency improvement rel-ative to baseline should be demon-strated in the coming decade

Exploitation of plasmoniceffects

> 10% efficiency improvement rel-ative to baseline should be demon-strated in the coming decade

Concentrator photo-voltaic technologies(CPV)

Si concentrator cells, III-Vmulti-junction cells

Laboratory efficiencies: 26.8% @ 96suns (Si cells) [30]; 40.7% @ 240suns (III-V cells) [31]

restrial applications for power generation under concentrated light. The mainchallenges for GaAs solar cells relate to minimize front surface and junctionrecombination, minimize series resistance and substrates costs (yet depositingGaAs on GaAs substrates is prohibitively expensive).

Thin-film solar cell active layer materials are deposited directly on large areasubstrates, such as glass panels (square meter-sized and larger) or foils (severalhundred meters long). Thin-film PV has an inherent low-cost potential becauseits manufacture requires only a small amount of active (high cost) materials (filmthickness is typically 1 µm) and is suited for fully integrated processing and


high throughputs [14]. There are three major inorganic thin-film technologies:amorphous/microcrystalline silicon (TFSi), the polycrystalline semiconductorsCdTe, and Cu(In, Ga)(S, Se)2 (CIGS). The CIGS technology currently exhibits thehighest cell and module efficiencies of all inorganic thin film technologies [22].A main challenge specially facing CIGS thin-film technology is the reduction ofthe material costs: high cost materials (In, Ga) should be replaced, active layerthicknesses should be reduced, and tolerance impurity in the materials shouldbe increased. Another problematic issue is that the available resources of in-dium are very limited. The attractive features of CdTe are its chemical simplicityand stability. Because of its highly ionic nature, the surfaces and grain bound-aries tend to passivate and do not contain significant defects. Its ionic naturealso means that absorbed photons do not damage its stability. CdTe’s favorablethermo-physical properties and chemical robustness make the CdTe cells easyand cheap to manufacture, using a variety of deposition methods. The efficiencyof CdTe cells depends on how the CdTe layers are grown, the temperature atwhich the layers are deposited and the substrate on which they are deposited.The main concern with this technology is the toxicity of the materials involved,even though very small amounts are used in the modules.

Organic solar cells offer the prospect of low cost active layer material, low-cost substrates, low energy input and easy upscaling. Another advantage isthe possibility of printing active layers , thereby boosting production through-puts typically by a factor of 10 to 100 compared to other thin-film technologies.Within ”organic solar cells”, two technology branches exist. One is the hybridapproach in which organic solar cells retain an inorganic component (for exam-ple the Graetzel cell [27]). The other is the full organic approach (for examplebulk-heterojunction solar cells [26]). More details about organic solar cells willfollow in the next section. The main challenges for both approaches relate to theincrease of efficiency, stability improvement and the development of a technol-ogy adapted for these materials.

The novel PV technologies can be characterized as high-efficiency approaches.Within this category, a distinction is made between approaches that tailor theproperties of the active layer to better match the solar spectrum and approachesthat modify the incoming solar spectrum without fundamentally modifying theactive layer properties. By introducing quantum wells or quantum dots consist-ing of a low-band gap semiconductor within a host semiconductor with a widerband gap, the current might be increased while retaining (part of) the higheroutput voltage of the host semiconductor [32, 33]. Tailoring the incoming solarspectrum for maximum conversion efficiency in the active semiconductor layer

1.2. Organic solar cells 5

relies on up- and down-conversion layers and plasmonic effects [34–36]. Theapplication of such effects in PV is definitely still at a very early stage, but thefact that they can be ”bolted-on” to conventional solar technologies (crystallinesilicon, thin films) may reduce their time-to-market considerably. An improve-ment of at least 10% (relative) of the performance of existing solar cells technolo-gies thanks to up- and down-convertors or the exploitation of plasmonic effectsshould be demonstrated in the coming decade.

The concentrator technologies (CPV) are based on the idea of concentratingsunlight to generate photovoltaic electricity, which is as old as the science ofphotovoltaics itself. Concentrating the sunlight by optical devices like lenses ormirrors reduces the area of solar cells and/or of the modules that house them,and increases their efficiency. CPV’s reliance on beam irradiation and the ne-cessity to track the sun’s motion across the sky by moving the system, is partlycompensated by longer exposure of the cells to sunlight during the day. Themost important benefit of this technology is the possibility to reach system effi-ciencies beyond 30%, which can not be achieved by single-junction 1-sun (nonconcentrating) photovoltaic technology.

Although PV systems are commercially available and widely developed, fur-ther development of PV technology is crucial to enable PV to become a ma-jor source of electricity. Currently the research is primarily aimed at reduc-ing the generation cost of PV electricity which at present may vary from 4 to8 euro/watt peak [15]. Considering the present development rate the cost of PVelectricity is expected to be competitive with consumer electricity (grid parity)in Southern Europe by 2015. Specifically, this means reaching PV generationcosts of 0.15 euro/kWh, or a PV system price of 2.5 euro/watt peak.

1.2 Organic solar cells

An alternative approach for inorganic semiconductor photovoltaic devices is theuse of organic, molecular semiconductors, which can be processed over large ar-eas at relatively low temperatures, either by vacuum sublimation of molecularmaterials, or, preferably, by processing from solution of film-forming materialssuch as polymers. Compared with Si, high optical absorption coefficients arepossible with these materials, which allows production of thin solar cells withmany advantages and opportunities as a result. An additional feature of an or-ganic semiconductor is that they are less expensive compared with inorganicsemiconductors like Si. The challenges in developing organic semiconductorsfor use in photovoltaic applications are considerable; requiring new materials,


new methods of manufacture, new device architecture, new substrates, and en-capsulation materials.

Until the seventies, carbon-based molecules and polymers have been con-sidered to be insulating materials, and as such have been exploited as electricalinsulators in many applications. A step forward for the molecular electronic ma-terials was the discovery in 1977 by Heeger et al. that the conductivity of the con-jugated polymer polyacetylene can be increased by several orders of magnitudeupon doping with iodine [37]. This discovery was followed by the observationof electroluminescence in poly(p-phenylene vinylene)(PPV) in 1990 [38, 39], thatinitiated an exciting and rapidly expanding field of research into these materials.The novel electronic properties of both molecular and polymeric semiconduc-tors arise from their conjugated chemical structure. This means that the bondsbetween carbon atoms in the framework of the polymer are alternating singleor double (called conjugation). In conjugated materials three sp2 hybrid orbitalsform covalent bonds: one with each of the carbon atoms next to it, and the thirdwith a hydrogen atom or other group. The remaining electron occupies a pz

orbital. The mutual overlap of pz orbitals creates π bonds along the conjugatedbackbone, by that delocalizing π electrons along the entire conjugated path. Thehighest filled π orbital is named the highest occupied molecular orbital (HOMO)and the empty lowest π∗ orbital is named the lowest unoccupied molecular or-bital (LUMO). The energy gap between the HOMO and the LUMO ranges from1 to 4 eV which makes most semiconducting polymers ideally suited for appli-cations in optoelectronic devices operating in the visible light range.

An organic solar cell consists of a photoactive layer sandwiched betweentwo different electrodes. One of the electrodes must be (semi-)transparent sincethe cell will be exposed to light. The underlying principle of a light-harvestingorganic PV cell consists of six distinct processes: (a) absorption of a photon creat-ing an exciton (bound electron-hole pair); (b) exciton diffusion; (c) charge trans-fer at donor/acceptor interface; (d) dissociation and separation of the carrierpair; (e) charge carrier transport to the corresponding electrodes; (f) collectionof charges by the electrodes.

The photoactive layer is based on a single, a bi-layer, or a mixture of two(or more) semiconductors. PV devices based on dyes or polymers yield limitedpower conversion efficiencies, typically well below 0.1%, mainly because of thelow relative dielectric constant of organic materials. Due to this low dielectricconstant strongly bound excitons are formed upon light absorption. Thereforethe electric field in PV device, due to the work function difference between theelectrodes, is much too weak to dissociate the excitons. Tang et al. demonstrated

1.2. Organic solar cells 7

that bringing a donor and an acceptor together in one cell dramatically increasesthe power conversion efficiency to 1% [40]. This concept of the heterojunctionhas since been widely exploited in a number of donor-acceptor cells: dye sen-sitized solar cells [41–43], planar organic semiconductor cells [40, 44–46], andbulk heterojunction solar cells [47–50]. The idea behind a heterojunction is touse two semiconductors with different electron affinities and ionization poten-tials. This favors exciton dissociation at the interface: the electron is acceptedby the semiconductor with the larger electron affinity and the hole by the ma-terial with the lower ionization potential. In the planar heterojunction (bi-layerdevice) charge separation is much more efficient at the donor/acceptor (D/A)interface than at the electrode interface in the single layer device. In this devicethe excitons should be generated in the close proximity to the D/A interface,within the exciton diffusion length. Otherwise, the excitons decay instead of tocontribute to the photocurrent.

One of the most used acceptors in planar heterojunction solar cells is thefootball-shaped fullerene molecule C60. The sixty electrons from the pz orbitalsgive rise to a delocalized π system similar to that in conjugated polymers. Theelectron mobility of 0.08 cm2V−1s−1 [51], reported from field-effect measure-ments on thin films of evaporated C60, and mobilities of 0.5 cm2V−1s−1, mea-sured by time-of-flight experiments [52] for C60 single-crystals grown from thevapor phase, makes it ideal for use as an electron acceptor in organic PV cells.In 1992 Sariciftci et al. report the photoinduced electron transfer from a semi-conducting polymer onto C60 [53].

It is clear that exciton dissociation is most effective at the interface in theplanar heterojunction cells, thus the exciton should be formed within its diffu-sion length from the interface. Since typically diffusion lengths are only 5-7 nm[54, 55], this limits the effective light-harvesting layer. However, for most or-ganic semiconductors the film thickness should be more than 100 nm in order toabsorb most of the light. Therefore thick film layers increase light absorption butonly a small fraction of the excitons will reach the interface and dissociate. Thisproblem was solved by Yu et al. [49] who intimately mixed the donor and accep-tor. Therefore the interfacial area is greatly increased and the distance excitonshave to travel in order to reach the interface is reduced. This device structure iscalled bulk heterojunction, and has been extensively used since 1995. One of theinherent problems with bulk heterojunction is that of solid-state miscibility. Yuet al. showed that the solvent used plays an important role on the film quality,and implicitly on the performance of the bulk heterojunction solar cell. Furtherexperiments performed by Shaheen et al. strengthened the fact that the solid


state morphology has an important effect on the power conversion efficiencyof solar cells. In 2000 [56] they obtained the first truly promising results forbulk heterojunction solar cells when mixing MDMO-PPV and [60]PCBM (1:4by weight) and optimizing the nanoscale morphology of the film, yielding toa power conversion efficiency of 2.5%. Nowadays the power conversion effi-ciency of polymer/fullerene bulk heterojunction solar cells approaches 6% [26].

1.3 Characterization of organic solar cells

1.3.1 Determining the charge carrier mobility

The current density-voltage characteristics in the dark are the result of thebulk charge carrier transport properties and the electrical properties of theorganic-electrode interface. Generally, electrical characteristics are essentiallyinterface-dependent in the low voltage/low current regime, while they are bulk-dependent in the high voltage/high current regime (Figure 1.1). In the lat-ter case, when the applied voltage (Vbias) exceeds the built-in voltage (Vbi) thedark current density scales quadratically with de voltage indicative of a space-charge limited (SCL) transport. One of the frequently used tools to determinethe charge carrier mobilities is to examine the space charge limited conductionin the dark from current-voltage measurements. The SCL dark conduction oc-curs when the contacting electrodes are capable of injecting either electrons intothe conduction band or holes into the valence band of a semiconductor or aninsulator, and when the initial rate of such carrier injection is higher than therate of recombination or extraction, so that the injected carrier will form a spacecharge, limiting the current flow. Therefore the SCL current is bulk limited.

The mobility µ of carriers in molecularly doped polymers is empirically de-scribed by a stretched exponential dependence [57, 58]:

µ = µ0 exp

(γ

√V

L

)(1.1)

where µ0 is the zero-field electron mobility, γ the field activation factor, and V/L

is the applied electric field.For a trap free semiconductor, assuming that the injecting contact is Ohmic,

the SCL current density is given by the Mott-Gurney equation: [59]

JDark =98εµ

V 2

L3(1.2)

where ε is the dielectric constant of material, V is the internal voltage of the

1.3. Characterization of organic solar cells 9

Dar

k cu

rren

t den

sity

Voltage 0.0

J V/L(local leakage current)

J exp (qV/kT)Diffusion current

J V2/L3

Space-charge limitedcurrent

Figure 1.1: The J − V characteristic in the dark with the three different domains.

device, µ is the charge carrier mobility and L is the thickness of the active layer.In the foregoing discussion, only the dependance of the charge carrier mobilityon the electric field was taken into account. Thus the SCLC is given by: [60]

JDark =98εµ0 exp

(0.89γ

√V

L

)V 2

L3(1.3)

Here V is related to the applied voltage (Vbias) as:

V = Vbias − Vbi − VRS(1.4)

where the built-in voltage (Vbi) is the voltage at which the J − V characteristicbecome quadratic. The VRS

is the voltage drop across the series resistance of thesubstrate.

Recent developments shown that the application of Equation 1.1 to describethe electric field dependence of the charge carrier mobility in low mobility me-dia is not fully correct due to the fact that the density dependence of chargecarrier mobility has been neglected [61–63]. It should be noted, however, thatthe charge carrier density in solar cells is fairly modest [64] and the density de-pendence of charge carrier mobility is negligible.

One can choose the electrodes materials in such a way that the injection ofeither electrons or holes can be suppressed or enhanced, thereby enabling toselectively determine the charge carrier mobility of either electrons or holes.This can be achieved if the work function of one of the electrodes is close to the


Voltage

Cur

rent

den

sity

FF =

Jmax

x Vmax

VOC

JSC

0.0

0.0J

Dark

JLight

Maximum power point

Figure 1.2: Typical J − V curves of an organic bulk heterojunction solar cell in the dark(empty symbols) and illumination (full symbols) conditions. The short-circuit currentdensity (JSC ) and open-circuit voltage (VOC ) are shown. The maximum output power isgiven by the rectangle Jmax × Vmax .

energy level of the transport band under investigation, while a high barrier forinjection of the other carrier type into the material exists. An energy diagramfor such an electron-only and hole-only device will follow in Chapter 2.

1.3.2 Characterization under illumination

In order to determine the performance of a solar cell device, as well as its elec-trical behavior, current density-voltage (J − V ) measurements in the dark andunder illumination are performed. Figure 1.2 shows the typical J − V curveof a solar cell in the dark (empty symbols) and under illumination (full sym-bols). The photovoltage developed when the terminals are isolated is called theopen circuit voltage (VOC). The photocurrent drawn when the terminals areconnected together is the short-circuit current (JSC). The operating regime of asolar cell is the range of bias from 0 to VOC in which the cell delivers power.

The cell power density is given by:

P = J × V (1.5)

P reaches a maximum at the cell’s operating point or maximum power point. Themaximum of the obtained electrical power Pmax is located in the fourth quad-rant where the product of current density J and voltage V reaches its maximum

1.4. Motivation and outline of the thesis 11

value :Pmax = Jmax × Vmax (1.6)

The fill factor (FF ) measures the quality of the shape of the J − V curve and isdefined as:

FF =Jmax × Vmax

JSC × VOC, (1.7)

which is the ratio between the Pmax that can be drawn from the device and theproduct of JSC and VOC . The power conversion efficiency η of a solar cell isrelated to these three quantities by:

η =JSC × VOC × FF

Plight(1.8)

where Plight is the power of the incident light.These four quantities JSC , VOC , FF and η are the key performance charac-

teristic parameters of a solar cell. All these photovoltaic parameters should bedefined for particular illumination conditions. The Standard Test Conditions(STC) include a constant temperature of the PV cells (25 oC), the intensity ofradiation (1000 W/m2), and the spectral distribution of the light (air mass 1.5or AM 1.5 global, which is the spectrum of sunlight that has been filtered bypassing through 1.5 thickness of the earth atmosphere).

The photocurrent generated by a solar cell under illumination at short circuitis dependent on the incident light intensity. The relation between the photocur-rent density and the incident spectrum is defined as the cell’s external quantumefficiency (EQE). The EQE is the ratio between the photocurrent and incidentphoton flux at one particular wavelength. EQE depends upon the absorptioncoefficient of the solar cell material, the efficiency of charge separation and theefficiency of charge collection in the device. The EQE does not depend on theincident spectrum, and it is therefore a key quantity describing solar cell perfor-mance under different conditions.

1.4 Motivation and outline of the thesis

Photovoltaic technology converts the sun’s rays directly into electricity withoutmoving parts or emitting pollutants. If its costs can be lowered, photovoltaicelectricity could become a competitive source of energy. It would help to com-bat the global threat of climate change and improve the security of energy sup-ply. Although a lot of progress has been made with organic bulk heterojunctionsolar cells, which were briefly discussed in the previous sections, the power


conversion efficiency needs to be improved further if these cells are to be com-mercialized. Since 2000 the bulk heterojunction solar cells are prepared fromconjugated polymers and, with only few exceptions, always using the samefullerene derivative ([60]PCBM). So far [60]PCBM remains the best perform-ing soluble fullerene derivative used in polymer:fullerene blends. This thesisaddresses the possibility of using new fullerene derivatives in organic bulk het-erojunction photovoltaic devices since one of the challenges for organic solarcells relates to developing of new materials.

Chapter 2 presents an overview of the materials used in the active layer ofbulk heterojunction solar cells, and the protocols that we developed for repro-ducible device preparation and characterization.

In Chapter 3, experimental studies of the morphology, charge transport andsolar cell performance in blends of P3HT and a newly developed analogue of[60]PCBM ([60]ThCBM) are presented. [60]ThCBM was designed with the aimof improving miscibility with polythiophene donors, especially (regioregular)P3HT.

The influence of film organization on the performance of the P3HT:[60]ThCBMbulk heterojunction photovoltaic cells is discussed in Chapter 4. The photo-voltaic performance of the fast drying and subsequent annealing of the activelayer is compared with the one of the cells with a slow growth photoactive layer.The benefit of a well balanced and enhanced charge carrier transport for slowgrowing films makes it possible to fabricate thicker films maximizing the lightabsorption in these blends.

Chapter 5 addresses the electrical stability of P3HT:Methanofullerene bulkheterojunction solar cell devices which has been investigated under standardcondition of illumination and temperature. Electrical and optical properties ofthese devices have been monitored for a test period of 1000 hours of continuousillumination.

Finally, Chapter 6 describes the solar cells performance in blends of P3HT anda mixture of two n-type semiconductors with approximately the same electronaccepting ability (as measured by the reduction potential) such as C60 and C70

fullerene derivatives. The use of two or more different n-type semiconductorsconcomitantly in the same device is essentially unheard of in the field of organicelectronics, let alone solar cells. Furthermore, the effect that certain impuritiesin the n-type semiconductor composition (such as pure C60 or higher fullerenesthan C70) have on the electrical properties of the solar cell devices is quantified.

REFERENCES AND NOTES 13

References and notes

[1] A. E. Becquerel, Comptes Rendus de l’Academie des Sciences 9 (1839), 561.

[2] W. Smith, Nature 7 (1873), 303.

[3] W. G. Adams, R. E. Day, Proceedings of the Royal Society London 25 (1876), 113.

[4] A. Pochettino, Academy Lincei Rendus 15 (1906), 355.

[5] M. Volmer, Annales Physik 40 (1913), 775.

[6] P. M. Borsenberger, D. S. Weiss, Organic photoreceptors for imaging systems, MarcelDekker, New York, (1993).

[7] R. H. Bube, Photoconductivity of solids, Wiley, New York, (1960).

[8] G. A. Chamberlain, Organic solar cells: a review, Solar Cells 8 (1983), 47.

[9] S. Antohe, Organic photovoltaic cells: a review, Romanian Republic Physics 53, (2001),427.

[10] D. M. Chapin, C. S. Fuller, G. L. Pearson, Journal of Applied Physics 25, (1954), 676.

[11] J. Zao, A. Wang, M. A. Green, F. Ferrazza, 19.8% efficient ”honeycomb” textured mul-ticrystalline and 24.4% monocrystalline silicon solar cells, Applied Physics Letters 73(1998), 1991.

[12] W. Shockley, H. J. Queisser, Detailed balance limit of efficiency of p-n junction solar cells,Journal of Applied Physics 32 (1961), 510.

[13] M. A. Green, Solar cells - Operating principles, technology and system applications, Uni-versity of New South Wales, Sydney, (1992).

[14] A. Goetzberger, C. Hebling, H. W. Schock, Photovoltaic materials, history, status andoutlook, Materials Science & Engineering R-Reports 40 (2003), 1.

[15] R. M. Swanson, The wafered silicon roadmap, moderator, Proceedings of the 12th NRELSilicon Workhop, Breckenridge, CO, (2002).

[51] R. P. Gale, R. W. McClelland, D. B. Dingle, J. V. Gormley, R. M. Burgess, N. P. Kim,R. A. Mickelsen, B. F. Stanbery, High-efficiency GaAs/CuInSe2 and AlGaAs/CuInSe2thin-film tandem solar cells, Conf. Record, 21st IEEE Photovoltaic Specialists Confer-ence, Kissimimee, May (1990) 53-57.

[17] D. Rose, O. Koehler, N. Kaminar, B. Mulligan, D. King, Mass production of PV moduleswith 18% total-area efficiency and high energy delivery per peak Watt, IEEE 4th WorldConference on Photovoltaic Energy Conversion, Waikoloa, HI, May (2006), 2018.

[18] J. Meier, J. Sitznagel, U. Kroll, C. Bucher, S. Fay, T. Moriarty, A. Shah, Potential ofamorphus and microcrystalline silicon solar cells, Thin Solid Films 451-452, (2004), 518.

[19] M. Yoshimi, T. Sasaki, T. Sawada, T. Suezaki, T. Meguro, T. Matsuda, K. Santo,K. Wadano, M. Ichikawa, A. Nakajima, K. Yamamoto, High efficiency thin film sili-


con hybrid solar cell module on 1 m2-class large area substrate, Conf. Record, 3rd WorldConference on Photovoltaic Energy Conversion, Osaka, (2003), 1566.

[20] J. Yang, A. Banerjee, S. Sugiyama, S. Guha, Recent progress in amorphus silicon alloyleading to 13% stable cell efficiency, Conf. Record, 26th IEEE Photovoltaic SpecialistsConference, Anaheim, (1997), 563.

[21] M. A. Contreras, K. Ramanathan, J. AbuShama, F. Hasoon, D. Young, B. Egaas,R. Noufi, Diode characteristics in state-of-the art ZnO/CdS/CuIn(1−x)GaxSe2 solar cells,Progress in Photovoltaics:Research and aplications 13 (2005), 209.

[22] Y. Tanaka, N. Akema, T. Morishita, D. Okumura, K. Kushiya, Improvement of VOC

upward of 600 mV/cell with CIGS-based absorber prepared by selenization/sulfurization,Conf. Proceedings, 17th EC Photovoltaic Solar Energy Conference, Munich, Octo-ber, (2001), 989.

[23] X. Wu, J. C. Keane, R. G. Dhere, C. DeHart, A. Duda, T. A. Gessert, S. Asher,D. H. Levi, P. Sheldon, 16.5%-efficient CdS/CdTepolycrystalline thin-film solar cell,Conf. Proceedings, 17th European Photovoltaic Solar Energy Conference, Munich,October (2001), 995.

[24] D. Cunningham, K. Davies, L. Grammond, E. Mopas, N. O’Connor, M. Rubcich,M. Sadeghi, D. Skinner, T. Trumbly, Large area ApolloTM module performance and relia-bility, Conf. Record, 28th IEEE Photovoltaic Specialists Conference, Alaska, Septem-ber (2000), 13.

[52] G. J. Bauhuis, P. Mulder, J. J. Schermer, E. J. HaverKamp, J. van Deelen, P. K. Larsen,High efficiency thin film GaAs solar cells with improved radiation hardness, 20th EuropeanPhotovoltaic Solar Energy Conference, Barcelona, June (2005), 468-471.

[26] J. Y. Kim, K. Lee, N. E. Coates, D. Moses, T. -Q. Nguyen, M. Dante, A. J. Heeger, Effi-cient tandem polymer solar cells fabricated by all-solution processing, Science 317, (2007),222.

[27] Y. Chiba, A. Islam, K. Kakutani, R. Komiya, N. Koide, L. Han, High efficiency dyesensitized solar cells, Techincal Digest, 15th International Photovoltaic Science andEngineering Conference, Shanghai, October (2005), 665.

[28] A. Marti, L. Cuadra, A. Luque, Quasi-drift diffusion model for the quantum dot interme-diate band solar cell, Electron Devices, IEEE Transations on Volume 49 (2002), 1632.

[29] A. Luque, A. Marti, N. Lopez, E. Antolin, E. canovas, C. Stanley, C. Farmer, L. J. Ca-ballero, L. Cuadra, J. L. Balenzategui, Experimental analysis of the quasi-Fermi levelsplit in quantum dot intermediate-band solar cells, Applied Physics Letters 87 (2005),083505.

[30] P. J. Verlinden, R. M. Swanson, R. A. Crane, K. Wickham, J. A. Perkins, 26.8% efficientconcentrator point-contact solar cell, Conf. Record, 13th European Photovoltaic SolarEnergy Conference, Nice, October (1995), 1582.


[31] R. R. King, R. A. Sherif, D. C. Law, J. T. Yen, M. Haddad, Z. M. Fetzer, K. M. Ed-mondson, G. S. Kinsey, H. Yoon, M. Joshi, S. Mesropian, H. L. Kotal, D. D. Krut,J. H. Ermer, N. H. Karam, New horizons in III-V multijunction terestrial concentratorcell research, 21st European PV solar Energy conference, Dresden, September (2006).

[32] A. G. Norman, M. C. Hanna, P. Dipo, D. H. Levi, R. C. Reedy, J. C. Ward, M. M. Al-Jassim, InGaAs/GaAs QD superlattices: MOVPE growth, structural and optical charac-terization, and application in intermediate-band solar cells, Photovoltaic Specialists Con-ference (2005). Conference Record of the 31st IEEE, January (2005).

[33] S. Suraprapapich, S. Thainoi, S. Kanjanachuchai, S. Panyakeow, Quantum dotmolecules for photovoltaic cell application, Photovoltaic Specialists Conference (2005).Conference Record of the 31st IEEE, January (2005), 98.

[34] C. Stumpel, M. M. Cann, C. del Canizo, I. Tobias, P. Fath, Erbium-doped UP-converterson silicon solar cells: assessment of the potential, 20th European PV Solar Energy con-ference, Barcelona, Spain, 43.

[35] M. Sendova-Vassileva, A. Vuchkov, O. Angelov, D. Dimova-Malinovska, Sm3+ pho-toluminescence on co-sputtered SiO2 thin films, Journal of Material Sciences: Materialsin Electronics 14 (2003), 853.

[36] J. C. Pivin, M. Sendova-Vassileva, M. Nikolaeva, D. Dimova-Malinovska, A. Mar-tucci, Optical extinction resonance of Au and Ag clusters formed by ion irradiation in SiO2

and Al2O3, Applied Physics A 75 (2002), 338.

[37] C. K. Chiang, C. R. Fincher, Y. W. Park, A. J. Heeger, H. Shirakawa, E. J. Louis,S. C. Gau, A. G. MacDiarmid, Electrical Conductivity in Doped Polyacetylene, PhysicalReview Letters 39 (1977), 1098.

[38] R. H. Friend, R. W. Gymer, A. B. Holmes, J. H. Burroughes, R. N. Marks, C. Tal-iani, D. D. C. Bradley, D. A. Dos Santos, J. L. Bredas, M. Loglund, W. R. Salaneck,Electroluminescence in conjugated polymers, Nature 397 (1999), 121.

[39] J. H. Burroughes, D. D. C. Bradley, A. R. Brown, R. N. Marks, K. MacKay,R. H. Friend, P. L. Burns, A. B. Holmes, Light-emitting diodes based on conjugated poly-mers, Nature 347 (1990), 539.

[40] C. W. Tang,Two-layer organic photovoltaic cell, Applied Physics Letters 48 (1986), 183.

[41] M. Gratzel, Photoelectrochemical cells, Nature 414 (2001), 338.

[42] B. O’Regan, M. Gratzel, A low-cost, high-efficiency solar-cell based on dye-sensitized col-loidal TiO2 films, Nature 353 (1991), 737.

[43] A. K. Jana, Solar cells based on dyes, Journal of Photochemistry and PhotobiologyA-Chemistry 132 (2000), 1.

[44] P. Peumans, V. Bulovic, S. R. Forrest, Efficient photon harvesting at high optical in-tensities in ultrathin organic double-heterostructure photovoltaic diodes, Applied Physics


Letters 76 (2000), 2650.

[45] B. A. Gregg, Bilayer molecular solar cells on spin-coated TiO2 substrates, ChemicalPhysics Letters 258 (1996), 376.

[46] M. Granstrom, K. Petritsch, A. C. Arias, A. Lux, M. R. Andersson, R. H. Friend,Laminated fabrication of polymeric photovoltaic diodes, Nature 395 (1998), 257.

[47] S. E. Shaheen, R. Radspinner, N. Peyghambarian, G. E. Jabbour, Fabrication of bulkheterojunction plastic solar cells by screen printing, Applied Physics Letters 79 (2001),2996.

[48] J. J. M. Halls, C. A. Walsh, N. C. Greenham, E. A. Marseglia, R. H. Friend, S. C.Moratti, A. B. Holmes, Efficient photodiodes from interpenetrating polymer networks, Na-ture 376 (1995), 498.

[49] G. Yu, J. Gao, J. C. Hummelen, F. Wudl, A. J. Heeger, Polymer photovoltaic cells -enhanced efficiencies via a network of internal donor-acceptor heterojunctions, Science 270(1995), 1789.

[50] C. J. Brabec, N. S. Sariciftci, J. C. Hummelen, Plastic solar cells, Advanced FunctionalMaterials 11 (2001), 15.

[51] R. C. Haddon, A. S. Perel, R. C. Morris, T. T. M. Palstra, A. F. Hebard, R. M. Fleming,C-60 thin-film transistors, Applied Physics Letters 67 (1995), 121.

[52] E. Frankevich, Y. Maruyama, H. Ogata, Mobility of charge carriers in vapor-phase grownC60 single crystal, Chemical Physiscs Letters 214 (1993), 39.

[53] N. S. Sariciftci, L. Smilowitz, A. J. Heeger, F. Wudl, Photoinduced electron-transferfrom a conducting polymer to buckminsterfullerene, Science 258 (1992), 1474.

[54] D. E. Markov, C. Tanase, P. W. M. Blom, J. Wildeman, Simultaneous enhancement of thecharge transport and exciton diffusion in poly(p-phenylene vinylene) derivatives, PhysicalReview B 72 (2005), 045217.

[55] D. E. Markov, E. Amsterdam, P. W. M. Blom, A. B. Sieval, J. C. Hummelen, Accuratemeasurement of the exciton diffusion length in a conjugated polymer using a heterostruc-ture with a side-chain cross-linked fullerene layer, Journal of Physical Chemistry A 109(2005), 5266.

[56] S. E. Shaheen, C. J. Brabec, N. S. Sariciftci, F. Padinger, T. Fromherz, J. C. Hummelen,2.5% efficient organic plastic solar cells, Applied Physics Letters 78 (2001), 841.

[57] L. B. Schein, A. Peled, D. Glatz, The electric field dependence of the mobility in molecu-larly doped polymers, Journal of Applied Physics 66 (1989), 686.

[58] P. M. Borsenberger, Hole transport in tri-p-tolylamine-doped bisphenol-A-polycarbonate,Journal of Applied Physics 68 (1990), 6263.

[59] K. C. Kao, W. Hwang, Electrical transport in solids 14, 151, Pergamon Press, Oxford


(1981).

[60] P. N. Murgatroyd, Theory of Space-Charge-Limited Current Enhanced by Frenkel Effect,Journal of physics. D, applied physics 3 (1970), 151.

[61] C. Tanase, E. J. Meijer, P. W. M. Blom, D. M. de Leeuw, Unification of the hole transportin polymeric field-effect transistors and light-emmiting diodes, Physical Review Letters 91(2003), 216601.

[62] C. Tanase, P. W. M. Blom, D. M. de Leeuw, E. J. Meijer, Charge carrier density depen-dence of the hole mobility in poly(p-phenylene vinylene), Physica Status Solidi A-AppliedResearch 201 (2004), 1236.

[63] W. F. Pasveer, J. Cottaar, C. Tanase, R. Coehoorn, P. A. Bobbert, P. W. M. Blom,D. M. de Leeuw, M. A. J. Michels, Unified Description of Charge-Carrier Mobilitiesin Disordered Semiconducting Polymers, Physical Review Letters 94 (2005), 206601.

[64] L. J. A. Koster, E. C. P. Smits, V. D. Mihailetchi, P. W. M. Blom, Device model forthe operation of polymer/fullerene bulk heterojunction solar cells, Physical Review B 72(2005), 085205.

2Materials and experimentaltechniques

A short overview of the materials used and the protocols that we developedfor reproducible device preparation and characterization will be presented. Theoperational characteristics of devices based on organic semiconductors are verysensitive with respect to the exposure to oxygen and water. Therefore the prepa-ration and characterization of such devices should be carried out in a controlledatmosphere. For this purpose we performed our experiments in a glove-boxsystem equipped with preparation facilities such as a spin-coater, a vacuum de-position system, and electrical characterization set-up.

19

20 Chapter 2: Materials and experimental techniques

2.1 Introduction

Since 1977 when the field of electronics based on conjugated polymers started [1]a number of a new processing technologies such as spin-coating, ink-jet print-ing, screen printing, and soft lithography techniques have been developed forthese flexible materials. In all of these cases, special measures need to be takenduring the preparation and characterization of the devices. For instance waterand oxygen molecules can easily penetrate in the organic films, reacting withthe employed electrodes, modifying the chemical nature of the molecules, act-ing as traps, doping agent, or quenching sites [2–5]. As a result, the transportproperties of a semiconductor change in time by exposure to air or moisture andthis limits the performance of the solar cell devices. This illustrates the impor-tance of controlled environment during the preparation and characterization ofthe organic-based devices. We present the protocols that were developed forreproducible preparation and characterization of solar cell devices. Differentdevice structures, used for the experimental evaluation of the transport prop-erties, such as electron and hole mobilities are described. Finally, the experi-mental setup for the study of the degradation of the cells upon light exposure isdescribed.

2.2 Device preparation

2.2.1 Materials and solutions

Absorption of light in organic materials results in the production of stronglybound electron-hole pairs (Frenkel type excitons), which have quasi-particlesizes smaller than the ones produced in inorganic solar cells. This finds its ori-gins in the weak van der Waals type intermolecular forces in organic materialsthat localize the excitons within molecules.

In general two semiconductors with different electron affinities are neededto dissociate these excitons in free charge carriers in organic solar cells. Oneof the organic semiconductors used, the electron-donating material (the donor,p-type semiconductor), is a conjugated polymer. The conjugated polymersused through this thesis are: poly(2-methoxy-5-(3’,7’-dimethyloctyloxy)-1,4-phenylene vinylene) (MDMO-PPV) and regioregular poly(3-hexylthiophene)(P3HT) (See Figure 2.1). By replacing MDMO-PPV with P3HT higher solar cellefficiencies were obtained, as will be demonstrated in Chapters 3 and 4. MDMO-PPV was purchased from Covion and used here as received. Regioregular P3HT,

2.2. Device preparation 21

Figure 2.1: Chemical formulae of the conjugated polymers used in this thesis.

electronic grade ( regioregularity more than 98.5 %, purchased from Rieke Met-als Inc.) was further purified by Bert de Boer in our laboratory [6]. The typicalaverage molecular weight, suggested by SEC (size exclusion chromatography),was ≈ 50,000 g/mol.

The other semiconductor used in the active layer of these solar cells isan electron-accepting material (the acceptor, n-type semiconductor), whichis in our study a methanofullerene. Methanofullerenes possess many ben-efits compared to the native (un-derivatized) fullerene in organic electron-ics. One benefit is their increased processability (increased solubility in aro-matic solvents) compared to native fullerenes, while maintaining much of thedesirable electronic properties of the native fullerene. The most commonlyused methanofullerene is [6,6]-phenylC61 butyric acid methyl ester ([60]PCBM).Another methanofullerene derivative is [6,6]-thienylC61 butyric acid methylester([60]ThCBM) (used in the Chapter 3 and Chapter 4 of this thesis). In Chap-ter 5 a higher methanofullerene analog of [60]PCBM such as [70]PCBM ([6,6]-phenylC71 butyric acid methyl ester) was used as acceptor in the photoactivelayer of the solar cells. In the last chapter a new library of methanofullerenessuch as: [70]ThCBM ([6,6]-thienylC71 butyric acid methyl ester), [60]ThCBM-bis adducts and [70]ThCBM-bis adducts, and non-methanofullerene derivativescalled fulleropyrrolidines such as: 2-(4-Butyloxyphenyl)N-methyl[60]fullero-pyrrolidine ([60]BPMF) and 2-(4-Butyloxyphenyl)N-methyl[70]fulleropyrrolidi-ne ([70]BPMF) were used as n-type semiconductors. All the n-type semicon-ductors used throughout this thesis were obtained from Solenne B.V. and theirchemical formulae are presented in Figure 2.2.

To study the device performance, blend solutions of polymer:methanofulle-rene were prepared. When the MDMO-PPV was the donor material, the sol-vents used were chlorobenzene (CB) or ortho-dichlorobenzene (ODCB). The op-


O

OCH3

[60]PCBM

O

OCH3S

[60]ThCBM

O

OCH3

[70]PCBM

O

OCH3S

[70]ThCBM

O

OCH3S

O

H3CO

S

[60]ThCBM-bis adducts

O

OCH3S

O

H3CO

S

[70]ThCBM -bis adducts

NOC4H9

[60]BPMF

NOC4H9

[70]BPMF

Figure 2.2: The chemical structures of the methanofullerenes used in this thesis.


Figure 2.3: Sandwich structure of a solar cell (left) and the layout of a photovoltaic cell(right). One substrate contains four devices with different active areas.

timum weight ratio of MDMO-PPV:methanofullerene ([60]ThCBM or [60]PCBM)was in this case 1:4 [7, 8]. When the donor material was P3HT, the solvents usedwere either chloroform, methylthiophene, or ODCB with an optimum weightratio of P3HT:metanofullerene of 1:1 [9]. Furthermore, all solutions preparedfor the experiments which are presented in Chapters 5 and 6 were made from amixture of P3HT:methanofullerene of 1:1 in ODCB. The solutions were preparedin a N2 atmosphere and rigorously stirred for more than 12 hours on a hot platemaintained at 50 oC, except for the solutions in chloroform which were stirredat room temperature.

2.2.2 Preparation of various device structures used

A plastic solar cell typically consists of a photoactive layer sandwiched be-tween a glass substrate covered with a patterned layer of indium tin oxide(ITO)(from Philips Research Laboratories) and a top electrode (as shown inFigure 2.3). Prior to the spin-coating of the photoactive layer on ITO, surfacetreatments were used to remove the contaminants and to reduce the rough-ness of the ITO layer. These cleaning steps consist of: rubbing the ITO glasssubstrates with soap and water, rinsing with de-mineralized water, followedby ultrasonic treatment in acetone (for 5 minutes) in order to remove the or-ganics (soap) from the surface. Next, the substrates were immersed in cas-cade water (running water with nitrogen bubbling through) for 5 minutes, fol-lowed by ultrasonic treatment in isopropanol for 5 minutes, and finally 20 min-utes of plasma treatment. Subsequently, a 60 nm transparent layer of poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS) (Baytron P VP Al4083 grade ) was spin coated from an aqueous dispersion under ambient condi-


Figure 2.4: The Karl Suss spin coater that we used in all our experiments.

tions, in order to improve device performance and stability. After spin coatingthe layer was baked at 200 oC, in ambient atmosphere, for 5 minutes in order toremove the remaining water from the PEDOT:PSS layer, that can cause degra-dation of the device. The thickness of PEDOT:PSS layer was typically 60 nm.

The photoactive layer was spin-coated on top of the PEDOT:PSS layer usinga Karl Suss spin coater (see Figure 2.4). The spin-coating process was dividedin three steps: in the first step the solution is spread over the surface whenthe substrate was accelerated to its final rotation speed. Subsequently, in thesecond step, the substrate was spinning at a constant rate where viscous forcesdominate the thinning behavior of the active layer. In this second step the filmis gradually formed and it is often possible to see color change in the reflectedlight. In the third step, the active layer is drying while the substrate is spinningat a constant lower rate and the solvent evaporation rate dominates the thinningbehavior. This last spin-coating step was not performed for slow growth activelayers (as will be discussed in the next paragraph). We refer to layers obtainedby the process above as fast growth active layers.

The spin-coating of the organic layer was performed inside a nitrogen glovebox with O2 and H2O levels below 0.1 ppm. The thicknesses of the MDMO-PPV:methanofullerene ([60]PCBM or [60]ThCBM) blend film was typically 100nm. The fast grown P3HT:[60]ThCBM layer thicknesses varied from 85 to262 nm. Thermal annealing was performed on complete P3HT:ThCBM de-vices in a nitrogen (N2) atmosphere on a hot plate at 150 oC for 4 minutes.This annealing temperature and duration was found to give optimum perfor-


Figure 2.5: The metal evaporation system with two sources for simultaneous deposition.The sample holder which could accommodate nine samples is located above the sources.

mance for the solar cells fabricated using chloroform as solvent [10]. The slowgrowth P3HT:methanofullerene films were spun from either 2-methylthiopheneor ODCB solution on top of the PEDOT:PSS layer. Subsequently, the wet filmswere dried overnight at room temperature, inside a N2 glove box, in a closedPetri dish [11]. The active layer thicknesses ranged for these devices from 150to 500 nm. All the film thicknesses were determined with a Dektak 6M Stylusprofiler (Veeco).

The top electrode was deposited by thermal evaporation through a shadowmask at a base pressure of 10−7 mbar. All the top electrode materials used in thisthesis were obtained from Sigma Aldrich. The metal deposition system (see Fig-ure 2.5) includes two resistive heating evaporation sources located at the bottomof the evaporating chamber. Boats made of a refractory metal (tungsten or tan-talum) were used as deposition sources. The evaporation rate can be controlledby the applied electrical power. The sample holder, which could accommodatenine substrates with the typical size of 3×3 cm2 each, was positioned above thetwo sources. Each substrate consisted of four independently addressable de-vices with different active areas: (0.096± 0.002), (0.172± 0.001), (0.378± 0.001),and (1.047±0.02) cm2. These areas were measured at Energy research Centre ofthe Netherlands (ECN) using an optical microscope. For all the solar cell devicespresented throughout this thesis, the top contact was either LiF (1 nm)/Al (100nm) or Sm (5 nm)/Al (100 nm). Chapter 3 addresses the variation of the pho-tovoltaic parameters of P3HT:[60]ThCBM solar cell devices as a function of the


Figure 2.6: Schematic representation of an electron-only device (left) and of a hole-onlydevice (right).

annealing temperature for these two different top contacts (Sm/Al and LiF/Al).

The electron-only (see Figure 2.6 left) devices, used to investigate the elec-tron transport in the pristine [60]ThCBM films, were fabricated in the same wayas was presented above. For such a device it is expected that LiF/Al will forman Ohmic contact for electron injection into [60]ThCBM. The work function ofPEDOT:PSS (5.2 eV) does not match the HOMO level of [60]ThCBM (≈ 6.1eV), thus the hole injection current from PEDOT:PSS into [60]ThCBM can be ne-glected. Consequently, only the electrons are expected to flow into [60]ThCBMunder forward bias conditions, therefore in the dark only the electron current ismeasured [13].

In order to investigate the hole transport through the polymer phase (MDMO-PPV or P3HT) in a blend with [60]ThCBM, the electron current through the[60]ThCBM (or other methanofullerene) phase has to be blocked. This can beachieved by choosing proper electrodes which suppress the injection of elec-trons into the device, resulting in a hole-only device (see Figure 2.6 right) [14].The hole-only devices were fabricated with the following structure: ITO/ PE-DOT:PSS/Polymer (MDMO-PPV or P3HT):[60]ThCBM/Pd. The work functionof PEDOT:PSS is about 5.2 eV. Hence, PEDOT:PSS serves as an Ohmic contactfor hole injection into the HOMO level of the polymer (MDMO-PPV or P3HT).The work function of Pd is about 5.12 eV, which leads to a large mismatch withthe LUMO level of [60]ThCBM [15] and therefore it prevents injection of theelectrons into [60]ThCBM [16](or other methanofullerene with similar LUMOlevel).

2.3 Morphology and optical properties

Atomic Force Microscopy (AFM) probes the topography of a surface. The oper-ation of an AFM is based on the measurement of the force between the sample

2.3. Morphology and optical properties 27

Figure 2.7: Solar cell set-up (left) that consist of a light source (1), sample table, and elec-trical connectors (2). At the right is a solar cell (with the four active areas) during illumi-nation from the halogen lamp.

surface and a sharp tip located at the free end of a flexible cantilever. Whilethe tip is scanned with respect to the sample, the interaction between the tipand the surface causes the cantilever to deflect. The magnitude of the deflec-tion is measured with high precision by the reflection of a laser beam from theback side of the cantilever on a photodiode detector. Depending on the distancebetween the tip and surface, three main AFM operation modes can be distin-guished: contact, non-contact and tapping mode. One of the main advantagesof AFM is that any type of sample can be scanned, although the resolution thatcan be achieved with AFM is lower compared with Scanning Tunneling Micro-scope (STM). However, scanning soft samples (like polymers) in contact modemay cause damage to the sample. It is preferable in this case to use the tappingmode, which is a semi-contact mode. In this mode the cantilever oscillates atthe resonance frequency of the quartz crystal it is mounted on. In this modethere is only a intermediate contact between tip and the sample. The (AFM)height and phase images of polymer:methanofullerene films were taken using aDI NanoScope IV scanning probe microscope operating in a tapping mode witha silicon cantilever, under ambient conditions.

Absorption spectra of the active layers were collected using a Perkin-Elmer


Lambda 900 UV-VIS/NIR spectrometer. All films were measured in the trans-mission mode on glass/ITO/PEDOT:PSS substrates and subsequently correctedfor substrate absorption.

2.4 Device characterization

The electric characterization of the devices was performed under a dry nitrogenatmosphere (inside glove box). Herein the anode (ITO) was taken as the posi-tively biased electrode and the cathode (LiF/Al, Sm/Al or Al) as the negativelybiased electrode. The devices were illuminated by a white light halogen lampwith an output power of 100 mW/cm2, calibrated with a silicon diode receivedfrom ECN (see the experimental set-up in Figure 2.7). The typical emission spec-trum of this halogen lamp is shown in Figure 2.8. However, for accurate pho-tovoltaic measurements, performed at ECN, a solar simulator class A was used,with a light spectrum that approximates the AM1.5 global (AM1.5 G) spectrum(air mass 1.5 or AM1.5G, which is the spectrum of sunlight that has been filteredby passing through 1.5 thickness of the earth atmosphere). A monocrystallineSilicon diode, which was initially calibrated at ISE Callab (Freiburg, Germany),was used as a reference cell in order to set the light intensity of the solar sim-ulator. In all these measurements the mismatch factor correction for the solarsimulator was taken into account. The detailed procedure of the STC measure-ments is described elsewhere [12].

The stability tests of polymer:methanofullerene solar cells (as investigatedin Chapter 5) were carried out inside a N2 glove box, with O2 and H2O lev-els below 0.1 ppm. These tests were carried out in two different locations: atECN and in our laboratory (RuG). Four halogen lamps were used for the sta-bility tests at the ECN facility. The output power of these lamps was typically100 mW/cm2 as controlled by a calibrated silicon diode. During the ageingtest the cells were kept at maximum power point operation (≈ 0.4 V forwardbias), which is similar to the operational conditions of the normal outdoor solarcell. Cooling of the cells was ensured by ventilation by electrical fans in orderto maintain the temperature at the surface of the samples constant at approxi-mately room temperature. In some tests, however, the cooling was switched offand the degradation tests were performed at an elevated temperature (≈ 50 oC).At the RuG laboratory the degradation tests were performed at a temperatureof ≈ 50 oC. A single halogen lamp was used for these tests, and the cells wereeither kept under open circuit operation (without any bias) or under maximumpower point conditions. The characteristic spectrum of the halogen lamps used

2.4. Device characterization 29

400 600 800 1000 1200 1400 1600 18000.0

0.2

0.4

0.6

0.8

1.0

Phot

on fl

ux [a

rb.u

nits

]

Wavelength [nm]

AM 1.5 Global Halogen lamp

Figure 2.8: Typical emission spectrum of the halogen lamp used in our experiments to-gether with the AM 1.5G solar spectrum.

in both laboratories is shown in Figure 2.8. Current-voltage measurements wereperformed using a computer-controlled Keithley 2400 Source Meter, sweepingin steps of 0.05 V with a delay of 0.10 sec, in the dark or under illumination.



[1] C. K. Chiang, C. R. Fincher, Y. W. Park, A. J. Heeger, H. Shirakawa, E. J. Louis,S. C. Gau, A. G. MacDiarmid, Electrical Conductivity in Doped Polyacetylene, PhysicalReview Letters 39 (1977), 1098.

[2] A. Hamed, Y. Y. Sun, Y. K. Tao, R. L. Meng, P. H. Hor, Effects of oxygen and illuminationon the in situ conductivity of C60 thin films, Physical review. B, Solid state 47 (1993),10873.

[3] M. Yan, L. J. Rothberg, F. Papadimitrakopoulos, M. E. Galvin, T. M. Miller, DefectQuenching of Conjugated Polymer Luminescence, Physical Review Letters 73 (1994),744.

[4] J. Steiger, S. Karg, R. Schmechel, H. von Seggern, Aging induced traps in organic semi-conductors, Synthetic Metals 122 (2001), 49.

[5] D. Kolosov, D. S. English, V. Bulovic, P. F. Barbara, S. R. Forrest, M. E. Thompson, Di-rect observation of structural changes in organic light emitting devices during degradation,Journal of Applied Physics 90 (2001), 3242.

[6] V. D. Mihailetchi, H. Xie, B. de Boer, L. J. A. Koster, P. W. M. Blom, Chargetransport and photocurrent generation in poly(3-hexylthiophene):methanofullerene bulk-heterojunction Ssolar cells, Advanced Functional Materials 16 (2006), 699.

[7] S. E. Shaheen, C. J. Brabec, N. S. Sariciftci, 2.5% efficient organic plastic solar cells,Applied Physics Letters 78 (2001), 841.

[8] V. D. Mihailetchi, L. J. A. Koster, P. W. M. Blom, C. Melzer, B. de Boer, J. K. J. vanDuren, R. A. J. Janssen, Compositional dependence of the performance of poly(p-phenylenevinylene):methanofullerene bulk-heterojunction solar cells, Advanced Functional Mate-rials 15 (2005), 795.

[9] D. Chirvase, J. Parisi, J. C. Hummelen, V. Dyakonov, Influence of nanomorphology onthe photovoltaic action of polymer-fullerene composites, Nanotechnology 15 (2004), 1317.

[10] L. M. Popescu, P. van’t Hof, A. B. Sieval, H. T. Jonkman, J. C. Hummelen, Thienylanalog of 1-(3-methoxycarbonyl)propyl-1-phenyl- [6,6]-methanofullerene for bulk hetero-junction photovoltaic devices in combination with polythiophenes, Applied Physics Let-ters 89 (2006), 213507.

[11] G. Li, V. Shrotriya, J. Huang, Y. Yao, T. Moriarty, K. Emery, Y. Yang, High-efficiency so-lution processable polymer photovoltaic cells by self-organization of polymer blends, NatureMatererials 4 (2005), 864.

[12] J. M. Kroon, M. M. Wienk, W. J. H. Verhees, J. C. Hummelen, Accurate efficiencydetermination and stability studies of conjugated polymer/fullerene solar cells, Thin SolidFilms 403-404 (2002), 223.


[13] 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, Electron transport in amethanofullerene, Advanced Functional Materials 13 (2003), 43.

[14] C. Melzer, E. J. Koop, V. D. Mihailetchi, P. W. M. Blom, Hole transport in poly(phenylenevinylene)/methanofullerene bulk-heterojunction solar cells, Advanced Functional Mate-rials 14 (2004), 865.

[15] M. Al-Ibrahim, H. K. Roth, U. Zhokhavets, G. Gobsch, S. Sensfuss, Flexible largearea polymer solar cells based on poly(3-hexylthiophene)/fullerene, Solar Energy Materi-als&Solar Cells 85 (2005), 13.

[16] V. D. Mihailetchi, P. W. M. Blom, J. C. Hummelen, M. T. Rispens, Cathode dependenceof the open-circuit voltage of polymer:fullerene bulk heterojunction solar cells, Journal ofApplied Physics 94 (2003), 6849.

3A thienyl analogue of PCBM for bulkheterojunction solar cells∗

An analogue of [6,6]-phenylC61 butyric acid methyl ester ([60]PCBM) was de-signed with the aim of improving miscibility with polythiophene donors, espe-cially poly(3-hexylthiophene) (P3HT). In the title compound the phenyl groupof [60]PCBM is replaced by a thienyl group, resulting in a compound named[6,6]-thienylC61 butyric acid methyl ester ([60]ThCBM). In this chapter, experi-mental studies of the morphology, charge transport and solar cell performancein blends of P3HT: [60]ThCBM are reported. [60]ThCBM is a first example ofmore tailored fullerene acceptor materials, which may optimize the morphol-ogy of bulk heterojunction photovoltaic devices with respect to their transportproperties in combination with a specific kind of donor polymer.

∗The main results of this Chapter have been published as: Lacramioara M. Popescu, Patrickvan’t Hof, Alexander B. Sieval, Harry T. Jonkman, and Jan C. Hummelen, Applied Physics Letters89 (2006), 213507.

33

34 Chapter 3: A thienyl analogue of PCBM for bulk heterojunction solar cells

3.1 Introduction

Organic photovoltaic devices consisting of three dimensional interpenetratingnetworks of conjugated polymers and a C60 derivative represent a challengingalternative for renewable sources of electrical energy [1]. In the last few years,several efforts to improve the efficiencies of these devices were undertaken [2–4]. One major improvement to this device structure was obtained by optimiz-ing the morphology of the blend by casting the polymer and [60]PCBM froma solvent that prevents large scale phase separation and enhances the polymerchain packing. Bulk-heterojunction solar cells made from poly [2-methoxy-5-(3’,7’-dimethyloctyloxy)-1,4-phenylene vinylene] (MDMO-PPV) and [60]PCBMreached a power conversion efficiency of 2.5% [2]. A step forward towardshigher efficiency was reached by replacing MDMO-PPV with a high-mobilitypolymer as the light absorber and the electron donating material. For pristineregioregular P3HT, the highest reported hole mobility in a diode configurationamounts to 10−4 cm2V−1s−1 [5], whereas the hole mobility of pristine MDMO-PPV is 5×10−7 cm2V−1s−1 [6–8]. Moreover, P3HT has the advantage of an en-hanced photostability and an improved optical absorption in the visible regionwhich results in a better overlap with the solar emission spectrum compared toMDMO-PPV. The most promising polymer solar cells developed at the start ofthis investigation in terms of efficiency and stability was based on a combina-tion of P3HT as donor and [60]PCBM as acceptor [9]. The efficiency of solar cellsbased on P3HT and [60]PCBM was shown to depend strongly on the processingconditions, and, in particularly, it shows an improvement when the devices arethermally annealed [9, 36]. Up till now [60]PCBM is the remains the best per-forming soluble fullerene derivative in combination with regioregular P3HT.Attempts have been made to develop new fullerenoids for bulk heterojunctionsolar cells by attaching solubilizing groups to C60, resulting in [60]PCBM ana-logues with altered substitution patterns on the solubilizing phenyl ring or byreplacement of the methano bridge present in both [60]PCBM and many otherderivatives [10–14]. Herein, an analogue of [60]PCBM was designed with theaim of improving miscibility with polythiophenes donors, especially P3HT. Inthis compound the phenyl group from [60]PCBM is replaced by a thienyl group,resulting in [60]ThCBM (the chemical structure shown in Figure 3. 1). Subtleas this modification may seem, we have previously observed that even slightmodifications of the [60]PCBM structure may lead to significant changes in itssolubility, the morphology of the spin coated films, and charge carrier trans-port. The synthesis of this compound was done using procedures as described

3.2. Electron transport in pristine [60]ThCBM 35

for [60]PCBM [15]. Cyclic voltametry measurements of [60]ThCBM show thatthe energy levels of the highest occupied molecular orbital (HOMO) and lowestunoccupied molecular orbital (LUMO) are identical to those of [60]PCBM.

In this chapter we report experimental studies of the morphology, chargetransport and solar cell performance in blends of MDMO-PPV:[60]ThCBMand P3HT:[60]ThCBM. The results indicate that [60]ThCBM provides the samecharge carrier mobility as [60]PCBM and that a highly functional bulk hetero-junction morphology with P3HT can be obtained.

3.2 Electron transport in pristine [60]ThCBM

Prior to investigating solar cells made with this methanofullerene as active layercomponent, it is important to determine its electron transport properties, sincethis is a relevant parameter which strongly influences the solar cell performance[16]. For films of pristine [60]PCBM, spin coated from a chlorobenzene (CB)solution, an electron mobility of 2×10−3 cm2V−1s−1 has been observed [17].[60]ThCBM is not as soluble as [60]PCBM in CB but it dissolves readily in ortho-dichlorobenzene (ODCB). Hence, for a better comparison, the electron mobili-ties of [60]PCBM and [60]ThCBM were determined in films spun from ODCBsolution.

Among the various methods to determine the charge carrier mobility weregard as the most appropriate one, the analysis of the space-charge limitedcurrent (SCLC) by investigating the (J-V ) characteristics in the dark. For a trap-free semiconductor, assuming that the injecting contact is Ohmic, the SCLC isgiven by the Mott-Gurney[26] equation: J ∝ µV 2L−3, where µ is the chargecarrier mobility, V is the applied voltage, and L is the thickness of the activelayer. Only the dependence of the charge carrier mobility on the electric fieldwas taken into account here. Thus, the SCLC is given by Eq. (1):[27]

JDark =98ε0εrµe exp

(0.89γ

√V

L

)V 2

L3, (3.1)

where µe is the zero-field electron mobility, ε0 is the vacuum permittivity, εr

is the relative dielectric constant, γ the field activation factor, and V/L is theapplied electric field (E). Recent developments shown that the application ofEquation 3.1 to describe the electric field dependence of the charge carrier mo-bility in low mobility media is not fully correct due to the fact that the densitydependence of charge carrier mobility has been neglected [28–30]. It should benoted, however, that the charge carrier density in solar cells is fairly modest


O

OCH3S

[60]ThCBM

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4

101

102

103

104

-1 0 1 2 3 410-310-210-1100101102103104

J D [A

/m2 ]

Vbias [V]

J Dar

k [A/m

2 ]V

bias-V

bi-V

RS [V]

Figure 3.1: Chemical structure of [60]ThCBM (left). Experimental (symbols) and calcu-lated (solid line) JDark-V characteristics for ITO/PEDOT:PSS/[60]ThCBM/LiF/Al elec-tron-only device with a thickness L=102 nm (right). JDark is plotted against the ap-plied voltage (Vbias) corrected for the built-in voltage (Vbi) and the voltage drop over theITO/PEDOT:PSS layers (VRS ; RS=18 Ω). The inset shows the experimental data withoutany voltage corrections.

[31] and the density dependence of charge carrier mobility is negligible. Theεr for [60]ThCBM as well as for [60]PCBM (≈3.9) [17] was determined fromimpedance spectroscopy measurements. Hence, the only unknown parameterin Eq. (1) is the charge carrier mobility.

The JDark-V characteristic of an electron-only pristine [60]ThCBM device,measured at room temperature is depicted in Figure 3. 1. The active area was0.1 cm2. Based on the work function of ITO/PEDOT:PSS (≈5.2 eV) and LiF/Al(≤3.7 eV), compared with the HOMO (6.1 eV) and LUMO (3.7 eV) levels of[60]ThCBM, one can assume that the current through the device is dominatedby the electrons, since only electrons are injected from LiF/Al electrode in for-ward bias (when ITO is positively biased). The value of µe was extracted byfitting Eq. (1) to the experimental data, as shown in Fig. 3. 1. [60]ThCBMshows a µe of 1.8(±0.8)×10−3 cm2V−1s−1 at room temperature. For [60]PCBMfilms, spin coated from ODCB solution, a µe of 1.4(±0.5)×10−3 cm2V−1s−1 wasobtained accordingly. These values of µe are similar to those observed pre-viously for [60]PCBM films spun from chlorobenzene [17]. Moreover, bothmethanofullerenes have the same electron mobility, which indicates that theelectron transport properties of [60]PCBM are not altered upon replacing thephenyl group by a thienyl moiety.

3.3. Bulk heterojunction solar cells using [60]ThCBM as acceptor 37

Figure 3.2: The 1.0×1.0 µm AFM height and simultaneously taken phase images of MD-MO-PPV:[60]ThCBM (1:4 w/w) films spin cast from: a CB solution (a) and an ODCBsolution (b).

3.3 Bulk heterojunction solar cells using [60]ThCBM asacceptor

3.3.1 [60]ThCBM and MDMO-PPV blends

It has been shown in literature that different solvents can result in different ef-ficiencies for MDMO-PPV:methanofullerene bulk heterojunction solar cells, asa result of a changed morphology [2, 18, 19]. Tapping mode atomic force mi-croscopy (AFM) was employed, to gain insight in the molecular packing of thecompounds. For the MDMO-PPV:[60]ThCBM (1:4 w/w) blends processed fromCB solutions the height images reveal a rough surface with domain sizes of ap-proximately 50-200 nm and a surface roughness (root-mean-square [rms]) of ≈3nm for a 1.0 µm scan size. Contrary, the samples fabricated from ODCB showa smooth surface with a roughness of ≈ 0.7 nm, and domain sizes of approx-imately 10 nm (as seen in Figure 3. 2). Because the exciton diffusion length in


-0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0-60

-40

-20

0

20

J Ligh

t [A/m

2 ]

Vbias

[V]

[60]PCBM [60]ThCBM

Figure 3.3: JLight-V characteristics of ITO/PEDOT:PSS/MDMO-PPV:[60]Fullerene (1:4w/w))/ LiF/Al devices, spin coated from CB (full symbols) and ODCB solutions (emptysymbols). The active layer thickness is ≈100 nm. All the devices were illuminated byhalogen lamp with intensity of 100 mW/cm2, calibrated using a reference silicon diode.

organic materials is typically small, 5-7 nm [20–22], the latter morphology favorsexciton dissociation and subsequently yields a higher photocurrent. The largedifferences in morphology induced by the solvent are also reflected by the pho-tovoltaic behavior of the MDMO-PPV:[60]ThCBM devices. The solar cells weremade by sandwiching the MDMO-PPV:[60]ThCBM (1:4 w/w) blend betweenan ITO/PEDOT:PSS bottom electrode and a LiF/Al top electrode.

The solar cell devices were illuminated with a halogen lamp, calibratedto an intensity of approximately 1 sun with a Si diode. To compare the roleof [60]ThCBM on the photovoltaic device performance, reference devices ofMDMO-PPV:[60]PCBM (1:4 w/w) were made and characterized in the samemanner. Figure 3. 3 shows the J-V characteristics of the solar cell devices un-der illumination. The short-circuit current density (Jsc) is increasing for cellswith MDMO-PPV:[60]ThCBM spin coated from ODCB solution, as a result ofan enhanced charge carrier generation rate due to an increased donor/acceptorinterface area, which facilitates exciton dissociation. The open-circuit voltage(V oc) of 0.75 V and a fill factor (FF ) of 52 % are somewhat lower, resulting in aoverall power conversion efficiency (η) of approximately 1.6 %. Similar powerconversion efficiencies were obtained for solar cells with a photoactive layer ofMDMO-PPV:[60]PCBM spin-coated from the same solvent.

Figure 3. 4 shows the experimental JDark-V of an ITO/PEDOT:PSS/MDMO-PPV:[60]ThCBM (1:4 w/w)/Pd hole-only device with a thickness of 260 nm, at


-1 0 1 2 3 4 510-5

10-4

10-3

10-2

10-1

100

101

102

103

0.1 1 1010-3

10-2

10-1

100

101

102

103

J Dar

k [A/m

2 ]

Vbias

-Vbi

-VRS

[V]

J Dar

k [A/m

2 ]

Vbias

[V]

Figure 3.4: JDark-V characteristics of ITO/PEDOT:PSS/MDMO-PPV:[60]ThCBM (1:4w/w)/Pd hole-only device with a thickness of 260 nm, at a temperature T = 298 K. Theinset shows the JD plotted against the applied voltage (Vbias), corrected for the built-involtage (Vbi) and the voltage drop over the ITO/PEDOT:PSS layers (VRS ; RS=18 Ω) in adouble logarithmic scale.

room temperature.

In the forward bias, when holes are injected from PEDOT:PSS, the slope ofthe log(JDark) versus log(Vbias − Vbi − VRS) plot, shown in the inset of Figure3. 4, indicates that JDark depends quadratically on the voltage. This is a char-acteristic for a space-charge limited current [23]. In reverse bias, however, theinjection of holes from Pd is suppressed due to the offset between its work func-tion and the HOMO level of the polymer [24]. From the fit of the SCLC to theexperimental data (solid line) we have extracted a hole mobility of MDMO-PPVphase, in the order of 2.5(±1.0)×10−5 cm2V−1s−1. A reason for the poor overallperformance of the solar cell might be the presence of a large mobility mismatchbetween the electrons and holes in MDMO-PPV:[60]ThCBM blend. In contrastto MDMO-PPV:[60]PCBM blends, the hole mobility of MDMO-PPV is not asstrongly enhanced when blended with [60]ThCBM [25].

3.3.2 [60]ThCBM and P3HT blends

Next we studied the solar cell performance of devices with a P3HT:[60]ThCBMphotoactive layer and the influence of the thermal annealing on the power con-


20 40 60 80 100 120 140 160 1800.0

0.8

1.6

2.4

3.2

20

40

60

80

100

0.56

0.58

0.60

0.62

36

48

60

72

[%]

Sm/Al LiF/Al

Annealing temperature [ oC]

J SC [A

/m2 ]

VO

C [V

]FF

[%]

Figure 3.5: Device performance of P3HT:[60]ThCBM (1:1 w/w) blends, under illumina-tion from a halogen lamp, as a function of annealing temperature and for different topcontacts (see the legend). The thicknesses of all P3HT:[60]ThCBM devices are ≈100nm.

version efficiencies. The devices consist of P3HT:[60]ThCBM (1:1 w/w) blendfilms, spin coated from chloroform solution on an ITO/PEDOT:PSS layer, fol-lowed by evaporation of the top contacts Sm (5 nm) or LiF (1 nm) and Al (100nm). The annealing was performed on a complete device in N2 atmosphere, ona hotplate for 5 minutes. The active layer thicknesses is ≈ 100 nm. Figure 3. 5shows the variation of the principal photovoltaic parameters as a function ofthe annealing temperature. The V oc for the cells with Sm/Al as top contact is≈ 30 mV higher compared with the V oc of the cells with LiF/Al as top contact,and also the FF is increased when Sm/Al is used as top contacts, which is an


-1 0 1 2 3 4 510-6

10-4

10-2

100

102

104

106

-0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0-120

-100

-80

-60

-40

-20

0

20

J Ligh

t [A/m

2 ]

Vbias

[V]

as cast annealed

J [A

/m2 ]

Vbias

[V]

Figure 3.6: J-V characteristics of ITO/PEDOT:PSS/P3HT:[60]ThCBM (1:1 w/w)/Sm/Aldevices as cast (circles) and after thermal annealing (squares) is shown in a semilogarith-mic plot. The devices were measured at room temperature (295 K) in the dark (full sym-bols) and under illumination (empty symbols) through the transparent ITO electrode.The devices under illumination were measured by a halogen lamp with intensity of 100mW/cm2, calibrated using a reference silicon diode. The inset shows the J-V character-istics under illumination in a linear scale.

indication that Sm/Al is a better top contact compared to LiF/Al.

The J-V characteristics of the solar cell devices, in the dark and under il-lumination, as cast and after thermal annealing are represented in Figure 3. 6.A substantial increase in the current is observed upon annealing, reflecting ahigher overall mobility. For the as cast devices the performance characteristicswere: Jsc = 14.35 A/m2, V oc = 0.59 V, FF = 36.5%, and η = 0.31%. After ther-mal annealing the Jsc increased to 83.65 A/m2. Due to annealing, the overallpolymer crystallinity has improved and the hole mobility is increased, as will bedemonstrated in the following. The devices show an increase in FF from 36.6%to 61.4% upon thermal annealing. This suggests an increase in carrier transportand it results in η of 3.03%.

In order to investigate the hole transport through the P3HT phase in ablend with [60]ThCBM, the electron current through the [60]ThCBM phase hasto be blocked. This can be done by choosing proper electrodes which sup-press the injection of electrons into the device, resulting in a hole-only de-vice. Herein the hole-only devices were fabricated with the following structure:


0 1 2 3 4 510-2

100

102

104

as cast annealed SCLC fit

J Dar

k [A/m

2 ]

Vbias

-Vbi

-VRS

[V]

Figure 3.7: Experimental (symbols) and calculated (solid lines) JDark-V for anITO/PEDOT:PSS/ P3HT:[60]ThCBM (1:1 w/w)/Pd hole-only device, as cast and afterthermal annealing (see legend). The active layer thickness is L=145 nm.

ITO/PEDOT:PSS/P3HT:[60]ThCBM (1:1 w/w)/Pd. The work function of PE-DOT:PSS is about 5.2 eV.

Hence, PEDOT:PSS serves as an Ohmic contact for hole injection into HOMOof P3HT (≈4.9 eV [32]). The work function of Pd is about 5.12 eV, which leadsto a large mismatch with the LUMO of [60]ThCBM [24] and this prevents in-jection of the electrons into [60]ThCBM. When the applied voltage exceeds thebuilt-in voltage (Vbi) in this hole-only device, the transport of holes through theP3HT is limited by the space-charge that accumulates, and consequently thecurrent is described by Eq. (1). The experimental JDark of P3HT:[60]ThCBMblends that were measured in hole-only device configuration, as cast and afterthermal annealing, are shown in Figure 3. 7. Assuming that the device is holedominated, the JDark-V measurements provide information on the zero fieldhole mobility (µh) of the P3HT phase in a blend. For as cast films, the µh inblend of P3HT:[60]ThCBM (1:1 w/w) is approximately 2×10−6 cm2V−1s−1 andthe field activation factor γ is 1.5×10−4 m0.5V−0.5. Upon thermal annealing ofP3HT:[60]ThCBM films an enhancement of the hole mobility by a factor of 100was observed (µh≈2×10−4 cm2V−1s−1).

The AFM images of the P3HT:[60]ThCBM films for as cast and after thermalannealing, are shown in Figure 3. 8. It has been shown that a whisker morphol-ogy is present in P3HT [33–35]. In pristine P3HT, this predominantly polycrys-talline film may be formed already during the coating process and the thermal

3.4. Conclusion 43

Figure 3.8: The 500×500 nm AFM height and simultaneously taken phase images ofP3HT:[60]ThCBM (1:1 w/w) films, as cast (a) and after thermal annealing (b).

annealing does not further improve the crystallinity. A phase AFM image of aP3HT:[60]ThCBM (1:1 w/w) spin coated film as cast [see Figure 3.8(a)] revealsan amorphous structure. In a blend film the presence of another molecule, inthis case [60]ThCBM, might prevent crystallization of P3HT. Upon thermal an-nealing, more structures can be seen in the spin-coated layer [see Figure 3. 8(b)],demixing of the two components is increased, and the crystallinity of the P3HTis enhanced [36]. At this point many researchers are tempted to believe that theresulting interpenetrating network provides continuous pathways in the entirephotoactive layer for efficient charge carrier transport.

3.4 Conclusion

In conclusion, we have investigated a new methanofullerene for application asan acceptor material in bulk heterojunction solar cells. The electrical characteri-zation of pristine [60]ThCBM films reveals that its electron transport properties


equal those of [60]PCBM, which indicates that the electron transport propertiesare not altered upon replacing the phenyl group by a thienyl moiety. The dimin-ished solubility of [60]ThCBM resulted in undesirable morphologies of the ac-tive layer when the MDMO-PPV:[60]ThCBM (1:4 w/w) mixture was spun fromCB. However, spin coating the mixture from ODCB solvent gave smooth filmsand an increase in the current output which makes it similar to the MDMO-PPV:[60]PCBM current. But as result of a low V oc of the devices, the overallpower conversion efficiencies was increasing only to 1.6%. The FF of 52%,slightly lower compared to the FF of the MDMO-PPV:[60]PCBM devices, isin accordance to the fact that the hole mobility of MDMO-PPV phase in a blendis not enhanced by [60]ThCBM. In contrast, using [60]ThCBM in combinationwith P3HT as the donor counterpart, the improved morphology, after thermalannealing and the optimized absorber composition, leads to a power conver-sion efficiency up to 3.0%. The increase of efficiency is due to the increasingcrystallinity of P3HT and hence the enhancement of the hole mobility in theP3HT phase of the blend by two orders of magnitude relative to as cast devices.

In the next Chapter more detailed studies of the morphology and electricalproperties of photovoltaic devices consisting of blends of P3HT:[60]ThCBM willbe presented.



[1] C. J. Brabec, V. Dyakonov, J. Parisi, N. S. Sariciftci (Eds), Organic Photovoltaics:Concepts and Realisations, Springer-Verlag, Heidelberg, 60 (2003), 159.


[3] N. S. Sariciftici, L. Smilowitz, A. J. Heeger, F. Wudl, Photoinduced electron-transferfrom a conducting polymer to buckminsterfullerene, Science 258 (1992), 1474.

[4] G. Yu, J. Gao, J. C. Hummelen, F. Wudl, A. J. Heeger, Polymer photovoltaic cells-enhanced efficiencies via a network of internal donor-acceptor heterojunctions, Science 270(1995) 1789.

[5] C. Goh, R. J. Kline, M. D. McGhee, E. N. Kadnikova, F. J. M. J. Frechet,Molecular-weight-dependent mobilities in regioregular poly(3-hexyl-thiophene) diodes, Ap-plied Physics Letters 86 (2005), 122110.

[6] P. W. M. Blom, M. J. M. de Jong, M. G. van Munster Electric-field and temperature de-pendence of the hole mobility in poly(p-phenylene vinylene), Physical Review. B 55 (1997),R656.

[7] M. C. J. M. Vissenberg, P. W. M. Blom, Transient hole transport in poly(-p-phenylenevinylene) LEDs, Synthetic Metals 102 (1999), 1053.

[8] H. C. F. Martens, H. B. Brom, P. W. M. Blom, Frequency-dependent electrical response ofholes in poly(p-phenylene vinylene), Physical Review. B 60 (1999), R8489.

[9] F. Padinger, R. S. Rittberger, N. S. Sariciftci, Effects of postproduction treatment on plas-tic solar cells, Advanced Functional Materials 13 (2003), 85.

[10] F. B. Kooistra, J. Knol, F. Kastenberg, L. M. Popescu, W. J. H. Verhees, J. M. Kroon,J. C. Hummelen, Increasing the open circuit voltage of bulk-heterojunction solar cells byraising the LUMO level of the acceptor, Organic Letters 9 (2007), 551.

[11] S. A. Backer, K. Sivula, D. F. Kavulak, J. M. J. Frechet, High efficiency organic photo-voltaics incorporating a new family of soluble fullerene derivatives, Chemistry of materials19 (2007), 2927.

[12] I. Riedel, E. von Hauss, J. Parisi, N. Martin, F. Giacalone, V. Dyakonov, Diphenyl-methanofullerenes: new and efficient acceptors in bulk-heterojunction solar cells, AdvancedFunctional Materials 15 (2005), 1979.

[13] I. Riedel, N. Martin, F. Giacalone,J. L. Segura, D. Chirvase, J. Parisi, V. Dyakonov,Polymer solar cells with novel fullerene-based acceptor, Thin Solid Films 451-452 (2004),43.

[14] X. Wang, E. Perzon, F. Oswald, F. Langa, S. Admassie, M. R. Andersson, O. Inganas,Enhanced photocurrent spectral response in low-bandgap polyfluorene and C70-derrivative-


based solar cells, Advanced Functional Materials 15 (2005), 1665.

[15] J. C. Hummelen, B. W. Knight, F. LePeg, F. Wuddl, J. Yao, C. L. Wilkins, Prepara-tion and characterization of fulleroid and methanofullerene derivatives, Journal of OrganicChemistry 60 (1995), 532.


[17] 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, Electron transport in amethanofullerene, Advanced Functional Materials 13 (2003), 43.

[18] M. M. Wienk, J. M. Kroon, W. J. H. Verhees, J. Knol, J. C. Hummelen, P. A. vanHal, R. A. J. Janssen, Efficient Methano[70]fullerene/MDMO-PPV bulk heterojunctionphotovoltaic cells, Angewandte Chemie 115 (2003), 3493.

[19] F. B. Kooistra, V. D. Mihailetchi, L. M. Popescu,D. Kronholm,P. W. M. Blom,J. C. Hummelen, New C84 derivative and its application in a bulkheterojunction solar cell, Chemistry of materials 18 (2006), 3068.

[20] J. M. M. Halls, K. Pichler, R. H. Friend, S. C. Moratti, A. B. Holmes, Exciton disso-ciation at a poly(p-phenylenevinylene) C60 heterojunction, Applied Physics Letters 68(1996), 3120.

[21] D. E. Markov, C. Tanase, P. W. M. Blom, J. Wildeman, Simultaneous enhancement ofcharge transport and exciton diffusion in poly(p-phenylene vinylene) derivatives, PhysicalReview B 72 (2005), 045217.

[22] D. E. Markov, E. Amsterdam, P. W. M. Blom, A. B. Sieval, J. C. Hummelen, Accuratemeasurement of the exciton diffusion length in a conjugated polymer using a heterostruc-ture with a side-chain cross-linked fullerene layer, Journal of Physical Chemistry A, 109(2005), 5266.

[23] J. K. J. van Duren, V. D. Mihailetchi, P. W. M. Blom, T. van Woudenbergh, J. C. Hum-melen, M. T. Rispens,R. A. J. Janssen, M. M. Wienk, Injection-limited electron currentin a methanofullerene, Journal of Applied Physics 94 (2003), 4477.



[26] K. C. Kao, W. Hwang, Electrical transport in solids 14, 151, Pergamon Press, Oxford


(1981).


[28] C. Tanase, E. J. Meijer, P. W. M. Blom, D. M. de Leeuw, Unification of the hole transportin polymeric field-effect transistors and light-emmiting diodes, Physical Review Letters 91(2003), 216601.

[29] C. Tanase, P. W. M. Blom, D. M. de Leeuw, E. J. Meijer, Charge carrier density depen-dence of the hole mobility in poly(p-phenylene vinylene), Physica Status Solidi A-AppliedResearch 201 (2004), 1236.

[30] W. F. Pasveer, J. Cottaar, C. Tanase, R. Coehoorn, P. A. Bobbert, P. W. M. Blom,D. M. de Leeuw, M. A. J. Michels, Unified Description of Charge-Carrier Mobilitiesin Disordered Semiconducting Polymers, Physical Review Letters 94 (2005), 206601.

[31] L. J. A. Koster, E. C. P. Smits, V. D. Mihailetchi, P. W. M. Blom, Device model forthe operation of polymer/fullerene bulk heterojunction solar cells, Physical Review B 72(2005), 085205.

[32] Y. Kim, S. A. Choulis, J. Nelson, D. D. Bradley, S. Cook, J. R. Durrant, Device annealingeffect in organic solar cells with blends of regioregular poly(3-hexylthiophene) and solublefullerene, Applied Physics Letters 86 (2005), 063502-1.

[33] J. K. Ihn, J. Moulton, P. Smith, Whiskers of poly(3-Alkylthiophene)s, Journal of polymerscience. Part B, Polymer physics 31 (1993), 735.

[34] E. Mena-Osteritz, A. Meyer, B. M. W. Langeveld-Voss, R. A. J. Janssen, E. W. Meijer,P. Bauerle, Two-dimensional crystals of poly(3-alkylthiophene)s: Direct visualization ofpolymer folds in submolecular resolution, Angewandte Chemie-International Edition39 (2000), 2680.

[35] S. Malik, A. K. Nandi, Crystallization mechanism of regioregular poly(3-alkyl thiophene)s,Journal of Polymer Science Part B-Polymer Physics 40 (2002), 2073.

[36] X. Yang, J. Loos, S. C. Veenstra, W. J. H. Verhees, M. M. Wienk, J. M. Kroon,M. A. J. Michels, R. A. J. Janssen, Nanoscale morphology of high-performance polymersolar cells, Nano Letters 5 (2005), 579.

4Effect of the P3HT:ThCBM growth rateon solar cell performance

We present a study of the morphology and electrical properties of photovoltaicdevices consisting of blends of regioregular poly(3-hexylthiophene) (P3HT) aselectron donor and [6,6]-thienylC61 butyric acid methyl ester([60]ThCBM) as anelectron acceptor. The solar cells were fabricated by depositing the photoac-tive layer from solvents with different boiling points. In this way we controlthe growth rate of the films. Atomic force microscopy analysis reveals a moregranular structure in case of fast grown films, as well as a well-defined fibri-lar and crystalline structure for slowly grown films. Electrical characterizationof the solar cell devices indicates a clear advantage of the slow growth proce-dure. This results in an order of magnitude enhancement in the hole mobilitythrough the P3HT phase. As a consequence of a more balanced transport ofelectrons and holes in the blend and an 0.4% (absolute) increase in power con-version efficiency. Additionally, it is shown experimentally that optimum deviceperformance, for the slow growth procedure, is obtained at film thicknesses ofapproximately 300 nm, making it possible to optimize light absorption in theseblends.

49

50 Chapter 4: Effect of the P3HT:ThCBM growth rate on solar cell performance

4.1 Introduction

Solar cells based on blends of conjugated polymers and methanofullerenederivatives continue to be of interest as promising candidates for low cost andlarge area photovoltaic devices. Up to now the ’state of the art’ polymer so-lar cells are based on a combination of regioregular P3HT and [60]PCBM. Forthese cells it was demonstrated that a thermal annealing process of the com-plete devices resulted in an enhancement of the solar cell performance [1]. Thisenhancement of the power conversion efficiency (η) has been attributed to: anincrease of the hole transport in the blend due to crystallization of P3HT, an im-proved film morphology, and a red shift of the absorption spectrum that resultsin a better overlap with the solar spectrum [2–4]. Li et al. have recently reportedplastic solar cells based on a rr-P3HT:PCBM mixture with a power conversionefficiency up to 4%, which was realized by controlling the growth rate of thephotoactive layer [5]. An efficient acceptor-type fullerene that can be used insolar cells based on rr-P3HT and other donor-type materials was synthesized inour group [6]. This fullerene derivative was designed to improve the compat-ibility with polythiophene donors by replacing the phenyl group of [60]PCBMwith a thienyl one, resulting in [60]ThCBM. Previous results, shown in Chapter3, indicate that [60]ThCBM gives the same charge carrier mobility as [60]PCBMand that a highly functional bulk heterojunction morphology can be readily ob-tained with rr-P3HT [6] as donor counterpart.

Previously it was demonstrated that high hole mobility can be obtained infield effect transistors with P3HT by using high boiling point solvents in the filmcasting process [7]. The microstructure of such devices was found to be highlycrystalline, with a needle-like morphology. In the solar cell devices, high elec-tron and hole mobility is desired to obtain good fill factors and high conversionefficiencies.

In this chapter, we study the influence of processing conditions by investi-gating the effect of the boiling point of the casting solvent on the morphologyand transport properties of P3HT:[60]ThCBM photoactive layer. As is demon-strated in the following, this study provides insight in the relation between themicroscopic morphology of the organic photoactive layer, the charge transportproperties, and the performance of the bulk heterojunction solar cells. Addi-tionally the thickness dependence of the solar cell performance was analyzed indetail.

4.2. Morphology 51

4.2 Morphology

The morphology of the active layer film is a key parameter in understandingthe device performance of the organic bulk heterojunction solar cells. The mostcommon and practical tool to investigate the morphology of the organic filmsis AFM. Here AFM was used to investigate the morphology of the active layerof regioregular P3HT:[60]ThCBM blends that were grown at different rates byusing solvents having different boiling points. Figure 4. 1 shows the AFMimages of the P3HT:[60]ThCBM blend films cast from three solvents with dif-ferent boiling points: chloroform (bp 61 oC), 2-methylthiophene (bp 113 oC),and ODCB (bp 173 oC). The films spin coated from chloroform solution werenamed fast growth active layers. Spin coating the photoactive layer from 2-methylthiophene and ODCB results in wet films, which were dried overnight atroom temperature, in a closed Petri dish in nitrogen glove box. We refer to thelayers obtained by the later method as slow growth active layers (more detailscan be found in Chapter 2).

The tapping mode AFM images of the fast grown films [see Figure 4. 1(a)]indicate that these films are less crystalline and have a granular surface to-pography without apparent whisker-like structures. The domain sizes aresmall, approximately 10 nm, and the domains consist of small crystallineP3HT chain like structures together with either pure [60]ThCBM or mixedP3HT:[60]ThCBM domains. Even though thermal annealing of the fast grownP3HT:methanofullerene films was found to enhance the P3HT crystallinity tosome extent [4], the degree of crystallization remains limited by the time avail-able for chain alignment, due to rapid evaporation of the solvent during thespin-coating process. Hence the interchain interactions, and consequently thedevice performance, are not optimal for fast grown films. For films spin castfrom methylthiophene, however, one can see [Figure 4. 1(b)] that well-definedwhisker-like structures with a width of approximately 15 nm and a length ofmore than 600 nm are formed during drying. Such well organized whisker-likestructures are present over the whole surface and we attribute them, in agree-ment with literature, to crystalline P3HT domains [4, 8]. The vibronic structureof the absorbtion peaks (see Figure 4. 2) is much more pronounced for the slowlygrown films, indicating a higher degree of ordering [9]. Compared with the fastgrown film, the absorption shifts a little to the red in the slowly grown film. Ared shift of approximately 37 meV appears in the better ordered films due toa better interchain π system overlap. It was previously shown that in a P3HTwhisker morphology, the P3HT chains are lying normal to the long axis of the


Figure 4.1: The 1 × 1µm AFM height and simultaneously taken phase images ofP3HT:[60]ThCBM blend films after the following treatment: fast grown spun from CHCl3(a), slow grown spun from methylthiophene, as inset the schematic representation for thewhisker formation (b), and slow grown spun from ODCB (c). This pictures are taken for200 nm active layer thickness of the fast grown film, and 300 nm active layer thickness ofthe slow grown films.

whisker and are folded with a period of approximately 15 nm [8, 10, 11]. Thisis in a good agreement with the topographical measurements shown in Figure4. 1(b). Moreover, the films prepared from an ODCB solution have a different

4.2. Morphology 53

1.5 1.8 2.1 2.4 2.7 3.0 3.30.00.10.20.30.40.50.60.70.80.91.0

fast grown (CHCl3)

slow grown (ODCB)

Nor

mal

ised

Energy [eV]

Figure 4.2: Normalized absorbtion coefficient (α) for fast and slow grownP3HT:[60]ThCBM blend films.

morphology compared to those prepared from methylthiophene, even thoughthey were both grown under slow drying conditions. In case of films spun fromODCB, more pronounced [60]ThCBM crystallites are observed at the surface,and a phase separation into relatively large domains occurs. This indicates thatthe boiling point of the solvent is a critical factor in the polymer crystalliza-tion process as well as for finding the optimum film morphology in order torealize the best device performance [7]. ODCB has a higher boiling point (173oC) than methylthiophene (113 oC). The drying time for films spin coated fromODCB solution is substantially longer under the same (atmospheric) conditions.It is known that [60]PCBM tends to crystallize in the pristine form and evencrystallizes in blends with MDMO-PPV or P3HT [12, 13]. This is also the casefor [60]ThCBM when blended with P3HT and spun from ODCB solution. Theround shaped structures, which can be seen in the AFM image in Figure 4. 1(c),are between 20-30 nm in size and are presumed to represent the [60]ThCBMcrystallites. The P3HT whiskers form a matrix, which is clearly visible in thephase images of Figures 4. 1(b) and 4. 1(c). The [60]ThCBM molecules diffuseout of the matrix and form large crystalline clusters for slow grown films fromhigher boiling point solvents (i.e., ODCB). Consequently, slower solvent evapo-ration facilitates the growth of highly crystalline films, which enhances polymerinterchain interactions and thus improves the hole mobility significantly, as willbe shown in the following section. However, this may have a negative effect


0 1 2 3 4 5

100

101

102

103

slow grown (Methylthiophene) slow grown (ODCB) fast grown (CHCl

3)

SCLC fit

J Dar

k [A/m

2 ]

Vbias

-Vbi

-VRS

[V]

Figure 4.3: Experimental (symbols) and calculated (solid lines) JDark-V characteristicsfor ITO/PEDOT:PSS/P3HT:[60]ThCBM/Pd hole-only devices. JDark is plotted againstthe applied voltage (Vbias), corrected for the built-in voltage (Vbi) and the voltage dropover the ITO/PEDOT:PSS layers (VRS). The thicknesses of the slow grown films are 270nm, whereas the thickness of the fast grown film is 160 nm.

on the photocurrent of solar cells since phase separation into large domains willreduce charge generation as a result of the short exciton diffusion lengths, char-acteristic for conjugated polymers.

4.3 Hole transport

In organic bulk heterojunction solar cells charge carrier mobility in the activelayer is determining the power conversion efficiency. Before discussing the so-lar cell performance of P3HT:[60]ThCBM blends, the charge carrier mobility inthe blends, prepared with different growth rates, is evaluated. As shown in theprevious section, slow drying of the P3HT:[60]ThCBM films using high boil-ing point solvents results in an enhanced crystallization of the P3HT phase, ascompared with a fast drying procedure. Electron and hole mobility values aredetermined experimentally using various techniques including time-of-flightmeasurements [5], field-effect transistor measurements [7], and the analysis ofthe space-charge limited current (SCLC) [14] by investigating current density-voltage (J-V ) characteristics in the dark. Figure 4. 3 shows the experimentaldark current densities (JDark) of a hole-only device for fast growth as well asfor slow growth of the active layer. The work function of PEDOT:PSS is 5.2eV and therefore the PEDOT:PSS serves as an Ohmic contact for hole injection

4.4. Solar cells characterization 55

into the HOMO of P3HT (4.9 eV [15]). On the other hand, the work function ofpalladium is 5.12 eV, which leads to a large mismatch with the LUMO level of[60]ThCBM [16], and it prevents injection of electrons into [60]ThCBM. Whenthe applied voltage exceeds the built-in voltage (Vbi), JDark scales quadrati-cally with the voltage V , indicative of space-charge limited (SCL) transport:JDark ∝ ε0εrµV 2/L3,where ε0εr is permittivity of the polymer, µ charge car-rier mobility, and L is the active layer thickness. For disordered semiconductors

the electric field dependence of the µ must be considered: µ ∝ µ0 exp(γ√

VL ),

where µ0 is the zero-field mobility, γ is the field activation parameter whichdepends on the energetic and positional disorder of localized transport states.The combined equations provide for a direct determination of the charge car-rier mobility [17–19]. The value of the hole mobility (µh), at zero electric field,was extracted by fitting the SCL dark current density (using the above equa-tions) to the experimental data, as shown in Figure 4. 3. The hole mobility ofP3HT in a blend with [60]ThCBM turned out to be the same for the films pre-pared by the slow growth method from methylthiophene and ODCB solution,and amounts to µh=2×10−3 cm2V−1s−1. For the fast dried and thermally an-nealed P3HT:[60]ThCBM blend films, we determined a hole mobility of P3HTphase of 2×10−4 cm2V−1s−1. Hence, applying slow drying, the hole mobility inthe P3HT phase increases one order of magnitude with respect to the thermallyannealed films. We conclude that during slow drying of the films, the P3HThas more time for reaching thermodynamic equilibrium, and the films have ahigh degree of ordering as can be seen from the absorption spectra, and there-fore the charge transport in the P3HT phase is increased.This leads to a morebalanced transport of electrons and holes in the device, since an electron mo-bility of ≈2×10−3 cm2V−1s−1 has been reported in P3HT:[60]PCBM blends forsimilar casting conditions [2, 20]. The fact that the hole mobility is identical inboth slow grown films, despite of a difference in film morphology shown in Fig-ures 4. 1(b) and 4. 1(c), is an indication that maximum crystallinity of P3HT isreached already at moderate growing times (moderate boiling point solvents).Most likely, further increasing the drying time of the blend films mainly resultsin the formation of more crystalline [60]ThCBM domains.

4.4 Solar cells characterization

After a study of the morphology and charge transport we proceeded with theanalysis of the performance of the solar cell with photoactive P3HT:[60]ThCBMblend layers prepared under fast and slow growing conditions. All devices were


-0.4 -0.2 0.0 0.2 0.4 0.6

-120-100-80-60-40-200

204060

fast grown (CHCl3)

slow grown (Methylthiophene) slow grown (ODCB)

J Ligh

t [A/m

2 ]

Vbias

[V]

Figure 4.4: JLight-V characteristics of ITO/PEDOT:PSS/P3HT:[60]ThCBM/Sm /Al de-vices with an active layer fast and slow grown. The slow grown films are 300 nm thick,and the fast grown film is 200 nm thick.

prepared as described in the experimental section. Figure 4. 4 shows the pho-tocurrent (JLight) of the solar cells in which the active layers were prepared byeither fast drying and subsequent annealing, or by slow drying. The best de-vice prepared by a fast drying procedure had a power conversion efficiency η =(3.1±0.2)%, a short-circuit current density Jsc = 90.8 A/m2, an open-circuit volt-age V oc = 0.59 V, and a fill factor FF = 59.3%. For the slowly dried films the bestperformance is obtained using an ODCB solution; with values of η = (3.5±0.3)%,Jsc = 99.5 A/m2, V oc = 0.55 V and FF = 63%. The best device prepared from2-methylthiophene showed a similar efficiency of η = (3.4±0.2)%. The betterperformance of the slowly grown devices, compared to the thermally annealedfast grown ones, most probably originates from an improved morphology andthe subsequent enhancement of the hole mobility through the crystalline P3HTphase in the blend. In order to improve the device performance we also studiedthe influence of the thickness of the active layer on the solar cells performance.Figure 4. 5 shows the variation of the photovoltaic parameters as a function ofactive layer thickness. The devices fabricated by the slow drying method wereall spun from ODCB. The Jsc is increasing rapidly with the active layer thick-nesses for both slow and fast growth films until approximately 200-230 nm, in-dicating an efficient charge generation in these devices. At a P3HT:[60]ThCBMfilm thickness of 200-230 nm, most of the incoming photons are absorbed by theactive layer. This is in agreement with previous results which showed that for

4.4. Solar cells characterization 57

0 100 200 300 400 50040

60

80

100

120

0.520.540.560.580.600.62455055606570

2.5

3.0

3.5

4.00 100 200 300 400 500

Jsc

[A/m

2 ]

Active layer thickness [nm]

Voc

[V]

[%]

FF [%

]

Figure 4.5: Device performance of the P3HT:[60]ThCBM blends, under illumination froma halogen lamp, as a function of the active layer thickness. The active layer was casted intwo different ways: fast grown spun form CHCl3 plus a subsequent thermal annealingstep (empty circles), and slow grown spun from ODCB (full circles). The full lines are aguide to the eye.

a P3HT film of 240 nm more than 95% of the incoming photons are absorbedby the active layer within its spectral range [21]. Above 200 nm film thicknessonly a small fraction of photons from the solar spectrum remains to be absorbedand therefore the Jsc saturates. A small decrease in Jsc for fast drying filmsfor thicknesses above 200 nm can be attributed to the limitation imposed on Jsc

by space-charge effects and/or charge recombination. These effects are a con-


sequence of an unbalanced electron and hole transport which is present in theblend and is caused by a lower hole mobility of P3HT phase [22, 23]. The V oc

is relatively constant as a function of active layer thickness, with a larger volt-age produced by the fast dried and annealed devices. This difference in voltageof approximately 40 mV is fully accounted by the red shift in absorbtion spec-tra (37 meV) as a result of the enhanced crystallinity of P3HT in the slow grownfilms (see Figure 4. 2). The reduction of the optical gap of P3HT is a consequenceof the decrease in ionization potential (HOMO level) by an enhancement of theintermolecular overlap. It appears to be in agreement with a simple model cal-culations which confirm the reduction of the band gap and find that about 3

4 ofthe shift occurs in the valence band [24]. Since the V oc of donor:acceptor bulkheterojunction solar cells is linearly related to the HOMOdonor-LUMOacceptor

energy difference [16], the V oc is lowered when the ionization potential of thedonor is decreased. Consequently the V oc is lowered in slow grown films.The FF is decreasing linearly with the layer thicknesses for both, fast and slowgrowth methods, due to increasing recombination of charge carriers by increaseof the mean distance they need to travel before they are extracted at the elec-trodes. Moreover the FF for the slow growth films is in average 10% higher forthe same film thickness, compared to the fast growth films, as a result of a betterand more balanced transport of electrons and holes in the device (see previoussection). Consequently, the difference in conversion efficiency between slow andfast growing films is mainly determined by the difference in FF and V oc for thesame film thickness. Although the V oc is lower for slow growth films its supe-rior fill factors well compensate the loss in V oc and therefore result in a betteroverall cell performance. The benefit of an improved charge carrier mobility forslow growing films together with a better balanced electron and hole transportmakes it possible to manufacture thicker active layers without significant loss inFF , and therefore it allows to maximize charge generation in these blends.

4.5 Conclusion

In conclusion, we have shown that slow growing of the active layer films inP3HT:[60]ThCBM blends, by using solvents with a high boiling point, facili-tates growing of highly crystalline films. This enhances the interchain interac-tions in the P3HT phase, and improves the hole transport in the blend by anorder of magnitude. This leads to a well balanced charge carrier transport inthe blend. The difference in power conversion efficiency between slow and fastgrown films is mainly determined by the difference in the fill factor and open-

4.5. Conclusion 59

circuit voltage of the cells. Although the open circuit voltage tends to be low-ered (with approximately 40 mV) in devices with active layers with a higher de-gree of molecular organization, it is the significant increase in the fill factor (upto 70%) which gives much higher (13% relative) power conversion efficiencies.The benefit of better charge transport for slow growing films makes it possibleto fabricate thicker films without significant loss in fill factor, and therefore itallows for maximizing light absorption in these blends.

In the next Chapter a study of the lifetime of the bulk heterojunction photo-voltaic devices, with a slow growth P3HT:methanofullerene photoactive layer,will be presented.





[3] W. Ma, C. Yang, X. Gong, K. Lee, A. J. Heeger, Thermally Stable, Efficient Polymersolar cells with nanoscale control of the interpenetrating network morphology, AdvancedFunctional Materials 15 (2005), 1617.

[4] X. Yang, J. Loos, S. C. Veenstra, W. J. H. Verhees, M. M. Wienk, J. M. Kroon,M. A. J. Michels, R. A. J. Janssen, Nanoscale morphology of high-performance polymersolar cells, Nano Letters 5 (2005), 579.


[6] L. M. Popescu, P. van’t Hof, A. B. Sieval, H. T. Jonkman, J. C. Hummelen, Thienylanalog of 1-(3-methoxycarbonyl)propyl-1-phenyl- [6,6]-methanofullerene for bulk hetero-junction photovoltaic devices in combination with polythiophenes, Applied Physics Let-ters 89 (2006), 213507.

[7] J. F. Chang, B. Sun, D. W. Breiby, M. M. Nielsen, T. I. Solling, M. Giles, I. McCulloch,H. Sirringhaus, Enhanced mobility of poly(3-hexylthiophene) transistors by spin-coatingfrom high-boiling-point solvents, Chemistry of materials 16 (2004), 4772.

[8] J. K. Ihn, J. Moulton, P. Smith, Whiskers of poly(3-Alkylthiophene)s, Journal of polymerscience. Part B, Polymer physics 31 (1993), 735.

[9] M. Sundberg, O. Inganas, S. Stafstrom, G. Gustafsson, B. Sjogren, Optical absorptionof poly(3-alkylthiophenes) at low temperatures, Solid State Communications 71 (1989),435.

[10] S. Malik, A. Nandi, Crystallization mechanism of regioregular poly(3-alkyl thiophene)s,Journal of polymer science. Part B, Polymer physics 40 (2002), 2073.

[11] E. Mena-Osteritz, A. Meyer, B. M. W. Langeveld-Voss, R. A. J. Janssen, E. W. Mei-jer, P. Bauerle, Two-dimensional crystals of poly(3-alkylthiophene)s: direct visualization ofpolymer folds in submolecular resolution, Angewandte Chemie. International edition39 (2000), 2679.

[12] X. Yang, J. K. J. van Duren, M. T. Rispens, J. C. Hummelen, R. A. J. Janssen,M. A. J. Michels, J. Loss, Crystalline organization of a methanofullerene as used for plasticsolar-cell applications, Advanced Materials 16 (2004), 802.


[13] D. Chirvase, J. Parisi, J. C. Hummelen, V. Dyakonov, Influence of nanomorphology onthe photovoltaic action of polymer-fullerene composites, Nanotechnology 15 (2004), 1317.

[14] V. D. Mihailetchi, H. Xie, B. de Boer, L. M. Popescu, J. C. Hummelen, P. W. M. Blom,Origin of the enhanced performance in poly(3-hexylthiophene): [6,6]-phenyl C61-butyricacid methyl ester solar cells upon slow drying of the active layer, Applied Physics Letters89 (2006), 012107.

[15] Y. Kim, S. A. Choulis, J. Nelson, D. D. Bradley, S. Cook, J. R. Durrant, Device annealingeffect in organic solar cells with blends of regioregular poly(3-hexylthiophene) and solublefullerene, Applied Physics Letters 86 (2005), 063502-1.





[20] V. Shrotriya, Y. Yao, G. Li, Y. Yang, Degradation mechanism of organic light-emittingdevice investigated by scanning photoelectron microscopy coupled with peel-off technique,Applied Physics Letters 89 (2006), 063503.

[21] K. M. Coakley, M. D. McGehee, Conjugated polymer photovoltaic cells, Chemistry ofmaterials 16 (2004), 4533.

[22] L. J. A. Koster, V. D. Mihailetchi, H. Xie, P. W. M. Blom, Origin of the light intensitydependence of the short-circuit current of polymer/fullerene solar cells, Applied PhysicsLetters 87 (2005), 203502.

[23] V. D. Mihailetchi, J. Wildeman, P. W. M. Blom, Space-Charge Limited Photocurrent,Physical Review Letters 94 (2005), 126602.

[24] R. A. Street, J. E. Northrup, A. Salleo, Transport in polycrystalline polymer thin-filmtransistors, Physical Review B 71 (2005) 165202.

5Stability tests ofP3HT:methanofullerence solar cells

In spite of the high potential of polymer photovoltaic cells, the main challengeremains the improvement of their stability under operational conditions. Thestability of P3HT:methanofullerene solar cell devices was investigated in thischapter. Electrical and optical properties of these devices were monitored fora test period up to 1000 hours under continuous illumination. The results in-dicate that solar cells with [60]ThCBM, as acceptor, show a faster degradationcompared to [70]PCBM and [60]PCBM solar cell devices. External quantum ef-ficiencies and absorption spectra indicate that the fullerene degradation is moreprominent than that of the polymer. The hole transport through the blend (P3HTphase) remains unaffected during the stability test, therefore the hole mobilityis not the cause for the degradation observed in the solar cell devices.

63

64 Chapter 5: Stability tests of P3HT:methanofullerence solar cells

5.1 Introduction

A key issue faced by the organic electronics community in general is the stabilityof the devices. The term ”stability” used here deals with the intrinsic stabilityof the organic materials used in the active layer, the nanomorphology and thestability of the contact between metal conductors and organic semiconductors.Device degradation arises from changes in morphology, loss of interfacial ad-hesion, and interdiffusion of components, as well as chemical decompositionunder operational conditions. Thus, careful design and material engineeringcan substantially improve device lifetimes.

In the last couple of years organic photovoltaics based on conjugated poly-mers and methanofullerenes have been suggested as a low cost alternative to theinorganic semiconductors due to their advantages of light weight and low-costplastic-based technologies. Most promising devices to date are blends of poly-thiophenes [such as regioregular P3HT] as donors and metanofullerene [suchas [60]PCBM] as acceptors [1–5]. P3HT possesses some unique properties com-pared to other polymers, including its self organization, high hole mobility, andextended absorption in the red region. For this type of blends, some approacheshave been used to improve the efficiency, including thermal annealing with si-multaneously applied external voltage [6], postproduction annealing at hightemperatures [2], and controlling the growth rate of the photoactive layer inthe device preparation [1]. The power conversion efficiency [1] and the lifetimeof organic photovoltaic devices still do not warrant commercialization. Manyefforts made in the past decade were focussed on optimizing the device param-eters (film thickness, layer materials, etc) of donor:acceptor bulk heterojunctionsolar cells, whereas only limited research exist on their stability and the degra-dation of the electrical characteristics [7–17].

In this chapter we report on the power conversion efficiency values, mea-sured under a solar simulator class A with a light spectrum that approxi-mates the AM1.5 global spectrum, for blends of regioregular P3HT and vari-ous methanofullerenes. A monocrystalline Silicon diode, which was initiallycalibrated at ISE Callab (Freiburg, Germany), was used as a reference cell in or-der to set the light intensity of the solar simulator. In all these measurementsthe mismatch factor correction for the solar simulator was taken into account.Further we present experimental studies on the stability under continuous il-lumination of the solar cells fabricated by slow drying of the active layer ofP3HT:methanofullerene.

5.2. Accurate efficiency measurements 65

Table 5.1: Photovoltaic parameters of the best P3HT:methanofullerene solar cells with anactive area of 0.0958(±0.0018) cm2. Here L is the active layer thickness.

L JSC VOC FF ηMethanofullerene (nm) (mA/cm2) (mV) (%) (%)[60]PCBM 300 10.5 563 60.7 3.6[70]PCBM 250 10.2 573 63.3 3.7[60]ThCBM 300 10.1 566 65.9 3.8

5.2 Accurate efficiency measurements

Prior to the investigation of the stability of the various P3HT:methanofullerenesolar cell devices, an accurate efficiency determination is necessary. The de-vice fabrication procedure is described in detail in Chapter 2 of this thesis. Theslow grown P3HT:methanofullerene blend films were spin cast from an ortho-dichlorobenzene (ODCB) solution on top of a 60 nm thick PEDOT:PSS layer.Subsequently, the wet films were dried overnight inside a closed Petri dish atroom temperature in a N2 glove box. To complete the devices Sm (5nm)/Al (100nm) or only Al (100 nm) top electrode was deposited by thermal evaporation.

Table 5.1 lists the photovoltaic parameters of the best fabricated P3HT:methanofullerene solar cells, measured at ECN. The active layer thicknessesis approximately 300 nm for [60]PCBM and [60]ThCBM cells and 250 nm for[70]PCBM cell. These thicknesses have been found to be the optimum for thesetype of blends (as shown in Chapter 4). The short-circuit current density (JSC)varies less than 4% for these solar cells and no significant improvement is ob-served with respect to the [70]PCBM [18, 19], which will be elaborated more inthe next chapter. From the data it can be concluded that the difference in thepower conversion efficiency (η) between the investigated solar cells finds its ori-gin in the variation of the fill factor (FF ) and where [60]ThCBM shows a slightlyincreased performance.

In figure 5.1 (bottom) we compare the external quantum efficiencies (EQE)of the solar cells listed in Table 5.1. A small shoulder that is present at 480 nmin the absorption of P3HT:[70]PCBM solar cell (as shown in the top part of Fig-ure 5.1) is a result of the fullerene contribution [18], but there is no contributionpresent in the EQE at the same wavelength range. This absorption seems not tocontribute to the EQE in the relevant wavelength range, although this may becaused by the lesser sensitivity of the EQE measurements. Nevertheless up to70% of the incoming photons are converted to current at the maximum absorp-tion range in these blends.


420 480 540 600 660 720 7800

20

40

60

0

1x1042x1043x1044x1045x104

[60]PCBM [70]PCBM [60]ThCBM

EQE

[%]

Wavelength [nm]


[cm

-1]

Figure 5.1: Absorption coefficient α (top) and the external quantum efficiency EQE (bot-tom) of ITO/PEDOT:PSS/P3HT:Methanofullerene/Sm/Al devices.

5.3 Effect of ageing on the device performance

After the initial current-voltage characterization of the P3HT:methanofullerenesolar cells presented in the previous section, stability test were carried out inorder to investigate their useful lifetime under continuous illumination. Thesetests were carried out inside a N2 glove box, with O2 and H2O levels below0.1 ppm in two different locations: at ECN and in our laboratory (RUG). Fourhalogen lamps were used for the stability tests at the ECN facility. The totaloutput power of these lamps was typically 100 mW/cm2 (one sun) as controlledby a calibrated Si diode. Cooling of the cells was ensured using ventilation byelectrical fans in order to maintain the temperature at the surface of the samplesaround room temperature. For some tests, however, cooling was not appliedand the stability tests were performed at an elevated temperature (around 50oC). In this section the investigation was focused on the charge carrier mobilityand on the solar cell performance.

5.3. Effect of ageing on the device performance 67

5.3.1 Charge carrier transport

One of the most important parameters that determines the performance of abulk heterojunction solar cell is the charge carrier mobility of electrons and holesafter dissociation of the excitons at the donor/acceptor interface. Therefore inthis section aging tests were carried out to investigate the changes that mayoccur in the charge carrier transport properties of the solar cells. It has beenshown in various papers that one of the most reliable ways to obtain informa-tion about the hole transport in a polymer:fullerene blend is the investigationof the current-voltage characteristics of a hole-only device in the dark [20–23].Here hole-only devices were fabricated by depositing an electron blocking con-tact on an otherwise normal P3HT:[60]PCBM solar cell device (see Chapter 2for details). This hole-only device was then illuminated (at RUG), for about 200hours at 50 oC, in a nitrogen glove box by a halogen lamp adjusted to a power of100 mW/cm2. Subsequently the illumination was switched off and dark J − V

characteristics of the devices were measured. Figure 5. 2 shows the J − V char-acteristics of the hole-only devices (open symbols) before and after 200 hoursof constant illumination. No significant degradation occurs in the dark currentof the devices due to illumination and the rise in temperature. Moreover, thedark current of the hole-only device after 200 hours of continuous illuminationis slightly higher compared to the dark current before degradation test. This isconsistently observed for all devices investigated. The typical hole mobility ofthe P3HT phase in the blend before degradation test (t = 0 hours) is ≈ 2 × 10−3

cm2V−1s−1 and after 200 hours of continuous illumination the hole mobility isslightly higher (≈ 3.3 × 10−3 cm2V−1s−1). This increase in the hole-mobilityafter 200 hours of continuous illumination could be due to a more favorablemorphology for hole transport in the blend. This result indicates that the holemobility is not the cause for the degradation observed in the solar cell device.

The other important parameter that could affect current transport throughthe solar cells is the electron mobility. However, electron transport is difficult tobe measured independently during degradation due to the fact that an electron-only device will require a low work function transparent metal electrode (inorder to block hole injection). This type of device is difficult to realize and it isless likely to remain stable during illumination. Experimental trials using a thinand transparent Sm low work function electrode have failed to succeed due toa large increase in leakage current of such a electron-only device upon illumi-nation. A successful method exists to fabricate an electron-only device using abottom electrode consisting of a self assembled monolayer (SAM) on top of a


0.0 0.5 1.0 1.5 2.0 2.5 3.0100

101

102

103

104

solar cell (t = 0 hrs) solar cell (t = 200 hrs) hole-only (t = 0 hrs) hole-only (t = 200 hrs)

J Dar

k [A/m

2 ]

Vbias

-Vbi

[V]

Figure 5.2: Dark J-V characteristics of the P3HT:[60]PCBM bulk heterojunction solarcells (see legend)and of the P3HT:[60]PCBM hole-only devices measured at 0 and after200 hours of continuous illumination under V oc condition. The applied voltage (Vbias)is corrected for the built-in voltage (Vbi).

thick (> 50 nm) Ag layer [24]. However, this method can not be used here tomeasure degradation of the devices under illumination since a (semi) transpar-ent electrode is required for photocurrent generation. Thus, as an alternative,we have investigated the dark current of the solar cell (before and after contin-uous illumination) where both charge carriers are injected from the respectiveelectrodes. The results are represented by full symbols in Figure 5.2. Since thehole current through the solar cell is slightly higher after 200 hours of light ex-posure it is intuitively anticipated that the decline of the dark current of thesolar cell is a result of a change in the electron transport properties and/or inthe recombination rates of electrons and holes in the solar cell device since bothcharges are injected from the electrodes [25].

5.3.2 Solar cells parameters

Figure 5.3 shows the parameters of the P3HT:methanofullerene solar cells mea-sured under maximum power point operation during 1000 hours of constantillumination at 25 oC. The methanofullerenes used are [60]PCBM, [60]ThCBM,and [70]PCBM. The decay of the photovoltaic parameters after 2 hours of il-lumination, in a nitrogen atmosphere, is in the same range for all three typesof solar cells. The performance is well above 95% of the initial values. In themore prolonged tests, there is a clear difference in the photovoltaic parame-


0.6

0.8

1.0

0.6

0.8

1.0

0.6

0.8

1.0

0 200 400 600 800 1000

0.6

0.8

1.0

Jsc

(t)/Js

c (t=

0)


Voc

(t)/V

oc (t

=0)

FF (t

)/FF

(t=0)

(t)/

(t=0

)

Time [hours]

Figure 5.3: Normalized operational parameters for the P3HT:Methanofullerene bulk het-erojunction solar cells (see legend) measured during 1000 hours of constant illumination,under maximum power point conditions, at room temperature in a nitrogen atmosphere.

ters for the three different methanofullerene used. The short-circuit current ofthe [70]PCBM solar cells shows the smallest decrease in the ageing tests, withits value staying above 90% of the initial value after 1000 hours of continuousillumination. For the [60]PCBM and [60]ThCBM solar cells, 1000 hours of con-tinuous illumination resulted in a decrease to 85% respectively 80% of the ini-tial short-circuit current. The degradation of the open-circuit voltage lies in thesame range for all three types of solar cell devices and their final value amountsapproximately 90% of the initial values. The decay of the fill factor for all threetypes of solar cells roughly follows the same pattern as the decay in open-circuitvoltage and could be due to modification in contact resistance, increased recom-bination/trapping rates or poorer charge carrier mobility.

These results suggest that the overall efficiency, being proportional to theJSC , VOC , and FF is more stable for [70]PCBM-based solar cells (a decrease


0 200 400 600 800 10000.4

0.5

0.6

0.7

0.8

0.9

1.0

JSC

VOC

FF

Nor

mal

ized

par

amet

ers

Time [hours]

Maximum power point operation

0 200 400 600 800 10000.4

0.5

0.6

0.7

0.8

0.9

1.0

Nor

mal

ized

par

amet

ers

Time [hours]

JSC

VOC

FF

Open circuit operation

Figure 5.4: Normalized parameters of the P3HT:[60]PCBM bulk heterojunction solar cells(see legend) measured during 1000 hours of constant illumination at a temperature of50 oC in nitrogen atmosphere. The data in the left figure shows the solar cells aged atmaximum power point condition whereas the right figure shows the results of cell agedat open-circuit condition.

to ≈ 90% after 1000 hours of continuous illumination) compared to [60]PCBMand [60]ThCBM solar cells (a decrease to 80% respectively 75% after 1000 hoursof continuous illumination). The decrease in efficiency for solar cells contain-ing [70]PCBM, [60]PCBM, and [60]ThCBM can be explained by the fact thatduring illumination also fullerenes crystallize and form larger crystallites intime, which could result in a decrease of the electron mobility in the blend asa consequence of an interrupted percolation path. The [70]PCBM (which isan isomeric mixture) has less tendency to crystallize compared to [60]PCBMand [60]ThCBM, therefore this kind of degradation of cell performance as aresult of a phase separation with this higher fullerene derivative in the activelayer is very likely reduced. One can conclude that the crystallization of themethanofullerene phase is an important factor in the stability of the devices un-der continuous illumination.

Next we studied the degradation of devices under different operational con-ditions: maximum power point and open-circuit. These tests were carried outat the RUG laboratories in nitrogen glove box, at a temperature of 50 oC, bysimultaneous illumination of both devices with one halogen lamp. Figure 5.4shows that there is no significant difference for the solar cell parameters of thecells kept under open-circuit or at maximum power point conditions during thetests, indicating that the kinetics of the degradation are not influenced by the ap-plied bias (≈ 0.4 V forward bias). In both cases the degradation was significantlymore pronounced compared to cells exposed to light at room temperature (seeFigure 5.3). It was observed that most of the degradation occurs directly after


0 60 120 180 240 300 360

0.7

0.8

0.9

1.0

Al Sm/Al J

SC

VOC

FF

Time [hours]

Nor

mal

ized

sola

r cel

l par

amet

ers

Figure 5.5: Normalized parameters for the P3HT:[60]PCBM bulk heterojunction solarcells (see legend) measured during 360 hours of constant illumination, at open-circuitconditions, at a temperature of 50 oC in nitrogen atmosphere. Empty symbols are solarcells with Al top contact and full symbols are solar cells with Sm/Al top contact.

the illumination was switched on and consequently during the rise in temper-ature, and all cell parameters are affected. This initial degradation is attributedto the morphology changes at the elevated temperature of the cells, and it is notobserved for the degradation carried out at room temperature (as seen in Figure5.3).

Another issue is the effect of the top electrode on the degradation process.Since the contact resistance between the active layer and metal electrode couldbe affected by light exposure, we studied solar cells with different top electrodes.Figure 5. 5 shows the decay of the photovoltaic parameters (JSC , VOC , and FF )of P3HT:[60]PCBM solar cells having either Sm/Al or only an Al as top elec-trode. In both cases the same degradation was observed for all cell parameters,after 360 hours of continuous illumination at 50 oC. These results indicate thatthe top electrodes used in our devices do not significantly affect the stability ofthese devices under illumination. Therefore most probably the degradation incell parameters, as seen in this specific study, might be caused by the changesthat occur in the active layer of the solar cells.

5.3.3 External quantum efficiency (EQE)

In order to gain an additional understanding of the origin of solar cell degrada-tion upon illumination, we investigated the EQE of the P3HT:methanofullerene


010203040506070

400 450 500 550 600 650 700 750 800

0.2

0.4

0.6

0.8

1.0


EQE

[%]

EQE

(t=10

00)/E

QE

(t=0)

Wavelength [nm]

Figure 5.6: Influence of the ageing on the EQE of the P3HT:methanofullerene solar cells(see legend). The full symbols represent the initial EQE measurements (taken at t=0hours), whereas the empty symbols indicate the EQE measured after 1000 hours of illu-mination at room temperature and in nitrogen atmosphere.

solar cells. Figure 5.6 shows the EQE as a function of the incident wavelengthfor solar cells measured before and after 1000 hours of illumination at roomtemperature for blends of P3HT and the three methanofullerenes used in this

300 400 500 600 700 8000.0

0.2

0.4

0.6

0.8

1.0 [60]PCBM P3HT P3HT:[60]PCBM (1:1 w/w)

Wavelength [nm]

Nor

mal

ized

Figure 5.7: Normalized absorption coefficients of a pristine [60]PCBM, P3HT films andof a P3HT:[60]PCBM (1:1 w/w) blend film.

5.4. Conclusion 73

investigation. These measurements were performed at ECN laboratories. Thelower part of the Figure 5.6 shows that a much stronger degradation occurs forwavelengths below 500 nm irrespective of the fullerene used. For these shortwavelengths the EQE of the solar cells is very likely dominated by the responseof fullerene absorption (as can be seen from Figure 5.7) and the subsequent holetransfer at donor:acceptor interface. This could be an indication that after illu-mination less excitons generated in the fullerene domain are able to dissociateinto free charge carriers as a result of a more unfavorable morphology createdduring illumination.

5.4 Conclusion

The experimental results reported in this chapter show that the highest powerconversion efficiency (3.8%) for the slow growth P3HT:methanofullerene filmsis obtained for P3HT:[60]ThCBM solar cells, mostly due to its superior fill fac-tor. The P3HT:[70]PCBM cell was found to be the most stable one, with a powerconversion efficiency decreasing to about 82% of initial performance, after 1000hours of operation under continuous illumination at room temperature. The[70]PCBM is a mixture of isomers and has less tendency to crystallize comparedto [60]PCBM and [60]ThCBM therefore this kind of degradation of cell perfor-mance as a result of phase separation with this higher fullerene derivative inthe active layer is very likely reduced. This means that the crystallization ofthe methanofullerene phase might play an important factor in the stability ofdevices under continuous illumination. Moreover, for the case when degra-dation tests were performed at higher temperatures, most of the degradationoccurred directly after the illumination was switched on and thus during therise in temperature, and this affected all cell parameters. This initial degrada-tion might be attributed to morphology changes in the photoactive layer at theelevated temperature of the cells. Furthermore, the hole mobility through theP3HT phase in the blend is slightly higher after 200 hours of continuous illumi-nation, compared with the hole mobility before degradation test. The changesin morphology could favor hole transport and in the same time be unfavorablefor the electron transport leading to an unbalanced charge carrier mobility inthe devices after the stability test. Additional results presented in this chapterindicate that the top electrodes used do not significantly affect the stability ofP3HT:methanofullerene devices under continuous illumination. This is anotherindication that the degradation in the cell parameters, under specific test condi-tions used in this study, is caused by the changes that occur in the active layer


of the solar cells.In the last Chapter experimental studies of the bulk heterojunction solar cells

containing two different types of n-type semiconductors [60]PCBM and [70]PCBM(of which one as a mixture of isomers) in combination with rr-P3HT as a p-typesemiconductor will be presented.



[1] G. Li, V. Shrotriya, J. Huang, Y. Yao, T. Moriarty, K. Emery, Y. Yang, High-efficiency so-lution processable polymer photovoltaic cells by self-organization of polymer blends, NatureMaterials 4 (2005), 864.

[2] W. Ma, C. Yang, X. Gong, K. Lee, A. J. Heeger, Thermally Stable, Efficient Polymersolar cells with nanoscale control of the interpenetrating network morphology, AdvancedFunctional Materials 15 (2005), 1617.

[3] P. Schilinsky, C. Waldauf, C. J. Brabec, Recombination and loss analysis in polythiophenebased bulk heterojunction photodetectors, Applied Physics Letters 81 (2002), 3885.

[4] M. Reyes-Reyes, K. Kim, J. Dewald, R. Lopez-Sandoval, A. Avadhanula, S. Curran,D. L. Carrol, Meso-structure formation for enhanced organic photovoltaic cells, OrganicLetters 26 (2005), 5749.

[5] M. Reyes-Reyes, K. Kim, D. L. Carrol, High-efficiency photovoltaic devices basedon annealed poly(3-hexylthiophene) and 1-(3-methoxycarbonyl)-propyl-1-phenyl-(6,6)C61

blends, Applied Physics Letters 87 (2005), 083506.


[7] H. Neugebauer, C. Brabec, J. C. Hummelen, N. S. Sariciftci, Stability and photodegra-dation mechanisms of conjugated polymer/fullerene plastic solar cells, Solar Energy Mate-rials&Solar Cells 61 (2000), 35.

[8] J. Steiger, S. Karg, R. Schmechel, H. von Seggern, Aging induced traps in organic semi-conductors, Synthetic Metals 122 (2001), 49.

[9] J. C. Hummelen, J. Knol, L. Sanchez, Stability issues of of conjugated polymer/fullerenesolar cells from a chemical viewpoint, Proceedings of SPIE Vol.4108 (2001), 76.

[10] F. C. Krebs, J. E. Carle, N. Cruys-Bagger, M. Andersen, M. R. Lilliedal, M. A. Ham-mond, S. Hvidt, Lifetimes of organic photovoltaics: photochemistry, atmosphere effectsand barrier layers in ITO-MEHPPV:PCBM-aluminium devices, Solar Energy Materi-als&Solar Cells 86 (2005), 499.

[11] K. Kawano, R. Pacios, D. Poplavskyy, J. Nelson, D. D. C. Bradley, J. R. Durrant,Degradation of organic solar cells due to air exposure, Solar Energy Materials&SolarCells 90 (2006), 3520.

[12] R. De Bettignies, J. Leroy, M. Firon, C. Sentein, Accelerated lifetime measurements ofP3HT:PCBM solar cells, Synthetic Metals 156 (2006), 510.

[13] F. C. Krebs, Encapsulation of polymer photovoltaic prototypes, Solar Energy Materi-als&Solar Cells 90 (2006), 3633.

[14] S. Chambon, A. Rivaton, J. -L. Gardette, M. Firon, Photo- and thermal degradation of


MDMO-PPV:PCBM blends, Solar Energy Materials&Solar Cells 91 (2007), 394.

[15] E. A. Katz, S. Gevorgyan, M. S. Orynbayev, F. C. Krebs, Out-door testing and long-termstability of plastic solar cells, European physical journal. Applied physics 36 (2007),307.

[16] S. Bertho, I. Haeldermans, A. Swinnen, W. Moons, T. Martens, L. Lutsen, D. Van-derzande, J. Manca, A. Senes, A. Bonfiglio, Influence of thermal ageing on the stabilityof polymer bulk heterojunction solar cells, Solar Energy Materials&Solar Cells 91 (2007),385.

[17] S. Chambon, A. Rivaton, J. -L. Gardette, M. Firon, L. Lutsen, Aging of a donorconjugated polymer: photochemical studies of the degradation of poly[2-methoxy-5-(3’,7’-dimethyloctyloxy)-1,4-phenylenevinylene] Journal of polymer science. Part A, Polymerchemistry 45 (2007), 317.

[18] Y. Yao, C. Shi, G. Li, V. Shrotriya, Q. Pei, Y. Yang, Effects of C70 derivative in lowband gap polymer photovoltaic devices: spectral complementation and morphol-ogy optimization, Applied Physics Letters 89 (2006), 153507.

[19] M. M. Wienk, J. M. Kroon, W. J. H. Verhees, J. Knol, J. C. Hummelen, P. A. vanHal, R. A. J. Janssen, Efficient Methano[70]fullerene/MDMO-PPV bulk heterojunctionphotovoltaic cells, Angewandte Chemie 115 (2003), 3493.


[21] S. M. Tuladhar, D. Poplavskyy, S. A. Choulis, J. R. Durrant, D. D. C. Bradley,J. Nelson, Ambipolar charge transport in films of methanofullerene andpoly(phenylenevinylene)/methanofullerene blends, Advanced Functional Materials15 (2005), 1171.

[22] V. D. Mihailetchi, H. Xie, B. de Boer, L. M. Popescu, J. C. Hummelen, P. W. M. Blom,Origin of the enhanced performance in poly(3-hexylthiophene): [6,6]-phenyl C61-butyricacid methyl ester solar cells upon slow drying of the active layer, Applied Physics Letters89 (2006), 012107.

[23] V. Shrotriya, Y. Yao, G. Li, Y. Yang, Degradation mechanism of organic light-emittingdevice investigated by scanning photoelectron microscopy coupled with peel-off technique,Applied Physics Letters 89 (2006), 063503.


[25] L. J. A. Koster, V. D. Mihailetchi, P. W. M. Blom, Bimolecular recombination in poly-mer/fullerene bulk heterojunction solar cells, Applied Physics Letters 88 (2006), 052104.

6Efficient PV blends comprising morethan two semiconductors

In this chapter we present experimental studies of the bulk heterojunction solarcells containing, concomitantly in the same device, two different types of n-typesemiconductors [60]PCBM and [70]PCBM (of which one as a mixture of isomers,in this case) in combination with a p-type semiconductor, here rr-P3HT. Replac-ing the n-type with the [60]PCBM/[70]PCBM mixture with no change in pro-cessing conditions results in identical device performance within experimentalerror. Next, we show that this concept generalizes to other fullerene derivativetypes. The fullerene bis-adducts, used as n-type semiconductors in combina-tion with rr-P3HT, are promising candidates for increasing power conversionefficiency of the solar cells because they show a clear enhancement in the VOC

compared with the ”state of the art” P3HT:[60]PCBM solar cells. Furthermore,the effect that certain impurities in the n-type semiconductor composition (suchas pure C60 or PCBM derivatives of fullerenes higher than C70) have on the elec-trical properties of the solar cell devices is quantified.

77

78 Chapter 6: Efficient PV blends comprising more than two semiconductors

6.1 Introduction

Significant progress has been made in the development of thin-film organic-based electronic devices, such as solar cells [1–5], transistors [6–8], photodetec-tors [9], sensors, and other devices for commercial application. The two mainforces that drive this upcoming technology is the promise of low cost and pos-sibility of mass production. Many of these devices utilize solution-processablesemiconductors based on fullerene derivatives in pure form. The most com-monly used fullerene derivative is [60]PCBM [10], which is a methanofullerene.Another methanofullerene derivative is [60]ThCBM [11], as used in Chap-ter 3. Methanofullerenes possess many benefits compared to the native (un-derivatized) fullerene in organic electronics applications. One benefit is their in-creased processability compared to native fullerenes, while maintaining muchof the desirable electronic properties of the native fullerene. The increase inprocessability is related to an approximately ten-fold increase in solubility inaromatic solvents.

Very limited research exists on the potential of the cost aspect of future massproduction of plastic PV devices. The cost of production of the two componentsof the PV active layer, i.e. the donor and acceptor materials, should definitelybe relatively low compared to the total cost of the device. Further reduction ofthe materials prices is therefore needed and fortunately possible. Since new andimproved donor materials are being developed steadily, it is quite hard to makea good assessment of the cost aspect of the donors, at the moment. This doesnot seem to be the case of the acceptor component. Some alternative fullereneacceptors have been described more recently, but very little research seems toindicate (at least up to now) that acceptors outside the fullerene realm are truecompetitors. Hence it is of great importance for the field to investigate andoptimize the cost and availability aspects of fullerenes and their derivatives,and to assure that even up to a terawatt/year production this technology wouldindeed be feasible from the materials point of view. The basic atom ingredientcarbon, like its classical counterpart silicon, is readily available. Independentof the production method, fullerenes are always formed as a mixture of carboncages Cn with n = 60, 70, 76, 78, 84, and many other ”higher” fullerenes. Isolationand purification of C60 (by far the most abundant fullerene) or C70 (the secondin abundance) from mixed fullerenes remains a costly procedure, but it is morefeasible to separate these two fullerenes from the higher ones.

In the cost analysis of the fullerene derivatives commonly used in organicsolar cells, such as [60]PCBM, the fullerene raw material is at present the largest

6.1. Introduction 79

driver of final cost of the derivative. The high price of single componentfullerenes C60 and C70 is due to the high price of separation and purification,which is clearly seen by comparing the cost of pure fullerene grades to the as-produced ”raw” fullerene grades, which are mixtures primarily of C60 and C70.Since the composition of the active layer is typically at least half fullerene deriva-tive for this type of organic solar cells, there is a large incentive in final devicecost-savings to minimize fullerene purification as much as possible.

The photoactive layer of organic bulk heterojunction solar cells is based on ablend of an electron donating material (p-type semiconductor) and an electronaccepting material (n-type semiconductor) forming nanostructured bicontinu-ous interpenetrating network. So far all these solar cells utilize a single n-type(acceptor) semiconductor. Furthermore, the use of two or more different n-typesemiconductors concomitantly in the same device is essentially unheard of inthe field of organic solar cells let alone in the field of classical inorganic electron-ics.

In this chapter experimental studies of the solar cells performance of blendsof rr-P3HT and a mixture of two n-type semiconductors [60]PCBM and [70]PCBM(of which one as a mixture of isomers) are presented. It should be noted that[60]PCBM and [70]PCBM have a close to identical first reduction potential andthe onset of their optical absorption is also identical (≈ 695 nm). Hence, boththe HOMOs and LUMOs of these molecular semiconductors are (close to) iden-tical. Further, the potential of improving power conversion efficiency of the bulkheterojunction solar cells using [60]ThCBM bis-adducts and/or [70]ThCBM bis-adducts is investigated with respect to their superior VOC . In the end, the effectthat certain impurities in the n-type semiconductor composition (such as pureC60 or PCBM derivatives of fullerenes higher than C70) have on the electricalproperties of the solar cell devices is quantified.

The detailed fabrication procedure and the electrical characterization ofthe devices are described in detail in Chapter 2. Particulary here the sol-vent used for spin coating the active layers was ortho-dichlorobenzene (ODCB)and the weight to weight ratio of P3HT:acceptor was always 1:1. Spin coat-ing the photoactive layer from ODCB results in wet films, which were driedovernight at room temperature, in a closed Petri dish in nitrogen glove box.We refer to the layers obtained by this method as slow growth active layers.The acceptors comprise of [60]PCBM, [70]PCBM, [60PCBM/[70]PCBM mix-ture, [60]ThCBM, [70]ThCBM, [60]ThCBM/[70]ThCBM mixture, [60]ThCBMbis-adducts, [70]ThCBM bis-adducts, [60]ThCBM bis-adducts/[70]ThCBM bis-adducts mixture, and a [60]BPMF/[70]BPMF mixture. The chemical formulae of


O

OCH3

[60]PCBM

O

OCH3

[70]PCBM

O

OCH3S

[60]ThCBM

O

OCH3S

[70]ThCBM

O

OCH3S

O

H3CO

S

[60]ThCBM-bis adducts

O

OCH3S

O

H3CO

S

[70]ThCBM -bis adducts

NOC4H9

[60]BPMF

NOC4H9

[70]BPMF

Figure 6.1: The chemical formulae of the methanofullerenes used in this chapter.

all these fullerenes used here are shown in Figure 6.1. The design of the exper-iments and the analysis of the obtained data throughout this chapter was doneusing a confidence-interval and hypothesis-testing procedure for single factorexperiments [12]. The experimental factor chosen here was the acceptor mix-ture of methanofullerenes while the response variable was solar cell parameters

6.2. Mixture of two n-type semiconductors in the active layer 81

(JSC , VOC , FF , and η). Statistical models were used to describe the responsesfrom experiments, such as Tukey’s honestly significant difference (HSD) proce-dure [12, 13].

6.2 Mixture of two n-type semiconductors in the active layer

Figure 6.2 shows the mean values for VOC and η of the P3HT:acceptor (1:1 w/w)blend devices fabricated under identical processing conditions where the onlydifference was in the composition of the acceptor. Here the acceptor is a blendof [60]PCBM and [70]PCBM, each of 99% pure or higher. Fullerenes higher inmolecular weight than C70 and fullerene derivatives higher in molecular weightthan [70]PCBM were removed, since these compounds may act as electron trapsand diminish performance (see section 6.4). The intervals around each averagedvalue in Figure 6.2 are based on Tukey’s HSD multiple comparison procedure[12, 13]. Based on this statistical analysis of the experimental data, there is suf-ficient evidence to conclude that the means of all solar cell parameters (see Ta-ble 6.1) are not significantly different across various compositions of [60] and[70]PCBM in the acceptor mixture. [60]PCBM and [70]PCBM do have very dif-ferent solubilities and precipitation behaviors as pure components, but blendsof the two surprisingly do not seem to affect morphology or complicate the pro-cessing conditions for devices. It is known by now that significant optimization

0 20 40 60 80 100

480

500

520

540

560

580

2.85

3.00

3.15

3.30

3.45

3.60

3.75

3.90

4.05

VO

C [m

V]

Weight percentage of [70]PCBM

[%]

VOC

Figure 6.2: Mean VOC and η values of the P3HT:acceptor (1:1 w/w) solar cells asa function of the weight percentage of [70]PCBM in the acceptor mixture (given by[60]PCBM:[70]PCBM). The 95% Tukey HSD confidence intervals are shown for all av-eraged.


Table 6.1: Mean values and 95% Tukey HSD intervals of the P3HT:acceptor (1:1 w/w)solar cells parameters. The acceptor here is a mixture of [60]PCBM and [70]PCBM in theweight percentage of [70]PCBM indicated in the first column. The active layer thicknessof all devices ranges from 180 to 300 nm.

[70]PCBM Counts JSC VOC FF η(%) (-) (mA/cm2) (mV) (%) (%)0 6 9.0 (±0.54) 563 (±5.5) 66.4 3.3 (±0.13)20 6 9.4 (±0.54) 558 (±5.5) 65.3 3.4 (±0.13)30 6 9.4 (±0.55) 558 (±5.7) 65.1 3.4 (±0.13)50 6 9.0 (±0.57) 561 (±5.9) 65.2 3.3 (±0.14)70 6 9.4 (±0.53) 556 (±5.5) 64.8 3.4 (±0.13)80 6 9.0 (±0.53) 564 (±5.5) 64.0 3.3 (±0.13)90 6 8.9 (±0.54) 556 (±5.5) 65.3 3.2 (±0.13)100 6 9.7 (±0.54) 555 (±5.5) 63.2 3.4 (±0.13)

efforts are often required because small changes in processing conditions andmaterials can cause significant changes in device performance.

Table 6.2: Mean values and 95 percent Tukey HSD intervals of the photovoltaic pa-rameters of P3HT:acceptor (1:1 w/w) solar cells. The acceptor composition is either amethanofullerene or a mixture of methanofullerenes (3:1 w/w of [60]:[70] unless other-wise indicated in the table).

acceptor Counts JSC VOC FF η(-) (mA/cm2) (mV) (%) (%)

[60]ThCBM 6 9.8 (±0.46) 553 (±3.2) 62.5 3.4 (±0.22)[70]ThCBM 6 10.9 (±0.46) 557 (±3.2) 62.5 3.8 (±0.22)[60]:[70]ThCBM 5 11.1 (±0.50) 559 (±3.4) 62.5 3.9 (±0.30)[60]ThCBM bis-adducts

6 6.7 (±0.46) 635 (±3.2) 65.2 2.8 (±0.20)

[70]ThCBM bis-adducts

6 6.4 (±0.46) 627 (±3.2) 59.1 2.4 (±0.22)

[60]:[70]ThCBMbis-adducts

6 6.5 (±0.46) 628 (±3.2) 62.5 2.5 (±0.22)

[60]:[70]BPMF(4:1)

6 4.1 (±0.58) 515 (±7.1) 26.2 0.5 (±0.13)

[60]:[70]BPMF(3:2)

6 6.1 (±0.58) 562 (±7.1) 40.9 1.4 (±0.13)

The results shown in Figure 6.2 indicate that by replacing the n-type withthe [60]PCBM/[70]PCBM mixture with no change in processing conditions ordevice parameters results in identical device performance within experimentalerror.

6.2. Mixture of two n-type semiconductors in the active layer 83

One important thing, in terms of device operation, that [60]PCBM and[70]PCBM have in common is their electron accepting ability (as determinedby the reduction potential) or, in other words, their lowest unoccupied molec-ular orbital (LUMO) level. For donor/acceptor blends, the LUMO level of theacceptor determines the upper limit of the VOC [14]. In order to verify the gen-eral applicability of the mixtures of fullerene derivatives in the acceptor, newfullerenes with the same LUMO levels (such as [60]ThCBM and [70]ThCBM,[60]ThCBM bis-adducts and [70]ThCBM bis-adducts, [60]BPMF and [70]BPMF)were tested, see Figure 6.1. Figure 6.3 shows the J − V characteristics of theP3HT:acceptor solar cell devices under illumination. The weight to weight mix-ture of the C60 fullerene derivatives to C70 fullerene derivatives in the activelayer is 3:1. Very similar J − V characteristics are observed for ThCBM’s butalso for ThCBM bis-adducts, especially for VOC . The statistical analysis of thesolar cells parameters are listed in Table 6.2. As can be seen from the table,there is no statistical difference between [60] and [70]ThCBM and their mixturesor between [60] and [70]ThCBM bis-adducts and their mixture. This furtherstrengthens the results obtained in Figure 6.2 with PCBM. It should be notedthat the [60]BPMF and [70]BPMF mixtures indicated in Table 6.2 shows statis-tically different results, because the solar cells were not fully optimized with

-0.4 -0.2 0.0 0.2 0.4 0.6 0.8

-12

-10

-8

-6

-4

-2

0

2

4

6

8

[60]ThCBM [70]ThCBM [60]ThCBM:[70]ThCBM (3:1) [60]ThCBM bis-adducts [70]ThCBM bis-adducts [60]ThCBM:[70]ThCBM (3:1) bis-adducts

J Ligh

t [mA

/cm

2 ]

Vbias

[V]

Figure 6.3: J-V characteristics under illumination of the bestITO/PEDOT:PSS/P3HT:acceptor (1:1 w/w)/Sm/Al devices. The acceptor compo-sition here is either [60]ThCBM and [70]ThCBM or a mixture of them, and [60]ThCBMbis-adducts and [70]ThCBM bis-adducts or a mixture of them (in the weight ratiosindicated in the legend).


respect to the solubility and film crystallinity.

6.3 Improving open-circuit voltage using bis-adductsfullerenes

One of the basic photovoltaic parameters that limits the overall performance ofthe P3HT:[60]PCBM bulk heterojunction photovoltaic devices is the VOC . Thisparameter received much attention in the last couple of years [10, 14–17]. Incase of ohmic contacts, meaning that the work function of the positive elec-trode matches the HOMO level of the donor and the negative electrode matchesthat of the LUMO level of the acceptor, the VOC is governed by the HOMO-LUMO difference of the donor and the acceptor [14]. Consequently, the VOC

can be increased by either lowering the HOMO level of the polymer (the donor)or raising the LUMO level of the fullerene (the acceptor) until remaining suf-ficient offset between the LUMOs of the donor and the acceptor necessary forcharge transfer [18] . In our laboratory, a lot of effort has been put to influencethe LUMO level of [60]PCBM by placing electron donating and electron with-drawing substituents on the phenyl ring, allowing further optimization of theVOC of polymer:fullerene organic solar cells [19]. The [60]ThCBM bis-adducts(see molecular structure depicted in Figure 6.1) have the first reduction poten-tials (as measured with cyclic voltametry) ≈ 100 mV more negative comparedto [60]PCBM. Therefore these bis-adducts of fullerene derivatives could be areal candidates for further improving the power conversion efficiencies of bulkheterojunction solar cells when blended with rr-P3HT (Lenes et al., AdvancedMaterials in print).

Figure 6.4 shows the VOC values of a series of six photovoltaic devices fab-ricated under the same processing conditions and containing rr-P3HT as theelectron donor and different methanofullerenes as the electron acceptor. Themethanofullerenes comprise [60]ThCBM, [70]ThCBM, [60]ThCBM/[70]ThCBMmixture, [60]ThCBM bis-adducts, [70]ThCBM bis-adducts, and [60]ThCBM bis-adducts/[70]ThCBM bis-adducts mixture. The device structure is ITO/PEDOT:PSS/P3HT:methanofullerene (1:1 w/w)/Sm/Al. As seen in Figure 6.4, all de-vices fabricated using bis-adducts fullerenes show a clear enhancement in theVOC of the cells by an average of 75 mV (in close agreement with the expec-tation from their reduction potentials). All parameters of the solar cells areshown in Table 6.2. However, in spite of the enhancement in VOC , the solarcells fabricated using ThCBM bis-adducts do not exhibit superior cell efficien-cies (solely because of much lower JSC) as compared with ThCBM. It should be

6.4. Effect of impurities in the fullerenes on cell efficiency 85

A C A:B C:D B D

550

555

560

565

570620

625

630

635

640

A = [60]ThCBMB = [70]ThCBMC = [60]ThCBM bis-adductsD = [70]ThCBM bis-adducts

Methanofullerenes

VO

C [m

V]

Figure 6.4: Mean VOC values of the ITO/PEDOT:PSS/P3HT:Methanofullerene (1:1w/w)/Sm/Al solar cell devices. The methanofullerenes used and their 3:1 w/w mix-tures are shown in the legend. Each data point is an average of 6 devices.

noted that the solar cells using ThCBM bis-adducts were fabricated using theprocess conditions optimized for [60]ThCBM, and almost certain the film mor-phology may vary significantly from that with ThCBM bis-adducts (which arebigger molecules). As the JSC of the bulk heterojunction solar cells is stronglyrelated to the solid-state morphology of the photoactive layer, the optimizationof the process condition is required in order to maximize the efficiency for thesenew molecules. The recently obtained results with PCBM bis-adducts were ob-tained after substantial optimization efforts (Lenes et al., Advanced Materials inprint).

6.4 Effect of impurities in the fullerenes on cell efficiency

The impact that impurities present in the fullerene acceptors can have on deviceperformance is linked, at least in part, to the degree of energetic disorder (orlack of crystallinity) that influences the electron mobility [20, 21]. Impuritiesmay affect the morphology, the electronic properties, and device performanceof a solar cell device in an unpredictable manner. For example trace amounts ofhigher fullerenes in the [60] or [70]PCBM may alter significantly the VOC of thedevice, since the reduction potentials of fullerenes are not the same. Therefore itis important to understand and predict the behavior of the solar cells upon thepresents of impurities in fullerene mixture.


-0.4 -0.2 0.0 0.2 0.4 0.6 0.8-10-8-6-4-202468

[60]PCBM:[70]PCBM [60]PCBM:[70]PCBM + C

60 (13%)

[60]PCBM:[70]PCBM + C60

(39%)

[60]PCBM:[70]PCBM + [Cn]PCBM (4%)

J Ligh

t [mA

/cm

2 ]

Vbias

[V]

Figure 6.5: JLight-V characteristics of the best ITO/PEDOT:PSS/P3HT:acceptor (1:1w/w)/Sm/Al solar cell devices. Here the acceptor is a [60]PCBM/[70]PCBM mix-ture (3:1 w/w) which is contaminated with molar percentage of C60 (see thelegend).[Cn]PCBM describes all the PCBM derivatives of fullerenes higher than C70.

Herein we have investigated the device performance of P3HT:fullerene (1:1w/w) blends by intentionally contaminating the fullerenes acceptors with vari-ous amounts of pure C60 or other fullerenes with higher molecular weight. Theacceptors used here are [60]PCBM, [70]PCBM or [60]PCBM/[70]PCBM mixturewith a purity of about 99% or higher. Including C60 up to 2.5% (molar) in theP3HT:[60]PCBM (1:1 w/w) active layer does not degrade the photovoltaic per-formance even if the VOC is decreasing with ≈ 17 mV. The decrease of the VOC

Table 6.3: Mean values and 95 percent Tukey HSD intervals of the photovoltaic parame-ters of P3HT:acceptor (1:1 w/w) solar cells. The acceptor composition is either [60]PCBM(denoted as A) or a 3:1 w/w mixture of [60]PCBM and [70]PCBM (denoted as B). All themethanofullerenes used here are intentionally contaminated with various (molar) per-centages of C60. [Cn]PCBM describes all PCBM derivatives of fullerenes higher than C70.

acceptor Counts JSC VOC FF η(-) (mA/cm2) (mV) (%) (%)

A 3 9.6 (±1.66) 582 (±4.2) 58.7 3.3 (±0.45)A + 2.5 % C60 6 9.5 (±1.17) 566 (±3.0) 62.4 3.3 (±0.32)B 5 8.3 (±1.28) 550 (±3.3) 62.5 2.8 (±0.35)B + 13% C60 6 8.1 (±1.17) 504 (±3.0) 60.7 2.5 (±0.32)B + 39% C60 6 6.3 (±1.17) 479 (±3.0) 56.9 1.7 (±0.32)B + 4% [Cn]PCBM 9 1.0 (±0.18) 289 (±2.7) 42.1 0.12 (±0.09)

6.4. Effect of impurities in the fullerenes on cell efficiency 87

B B+13% B+39% A A+2.5%

480490500510520530540550560570580590

2

4

6

8

10

12

14

Molar percentage of C60

VO

C [m

V]

A = [60]PCBMB = [60]PCBM:[70]PCBM

VOC

JSC

J SC [m

A/c

m2 ]

Figure 6.6: Operational parameters for the ITO/PEDOT:PSS/P3HT:acceptor (1:1w/w)/Sm/Al solar cell devices as a function of the molar percentage of C60 in theacceptor. Here the acceptor is either [60]PCBM (denoted as A) or [60]PCBM/[70]PCBMmixture (3:1 w/w) (denoted as B). The circles represent the VOC and the triangles repre-sent the JSC (see the legend).

is compensated by the increase in the FF , and in this way the solar cell effi-ciency (η) stays the same as compared with the bulk heterojunction solar cellswith an active layer without contamination with C60 (see Table 6.3). The influ-ence of pure C60 that is added to an acceptor mixture of [60]PCBM:[70]PCBM(3:1 w/w) was further investigated. Figure 6.5 shows the photocurrent (JLight)of these solar cells for different levels of pure C60 in the acceptor mixture. FromFigure 6.5 and Table 6.3 it can be seen that 39% molar of pure C60 in the activelayer leads to a decrease of 1.1% (absolute) of the power conversion efficiencyas compared with uncontaminated cells (our reference cells, denoted as B in theTable 6.3). Nevertheless, the introduction of up to 39% molar of pure C60 inthe acceptor mixture did not completely degrade the device performance. Onthe other hand, however, the addition of only 4% (molar) of PCBM derivativesgreater than C70 (which are deep traps), resulted in a strong decrease in the cellefficiency. This is understood to be due to a combined effect of a decrease inVOC , as a result of lower lying LUMO level for higher fullerenes, but also by adecrease in JSC as a result of deep traps introduced in the active layer [21].

Figure 6.6 shows the variation of the VOC and JSC as a function of the molarpercentage of C60 added in the active layer. The VOC (represented by the cir-cles) is decreasing rapidly with the amount of C60 introduced in the active layerindicating that traps lower the quasifermilevel therefore the LUMO level of the


resulted acceptor is altered. Furthermore, the JSC (represented by the triangles)is staying constant till C60 reached a concentration of 39 %. The fact that theJSC is staying constant at low concentrations of C60 in the active layer confirmsthat we do not introduce electron traps by introducing C60. We simply obtainthe VOC of P3HT:C60 cell, but of one in which morphology is influenced by thepresence of the methanofullerene. At 39 % (molar) of C60 in the active layer onemay expect percolation in the C60 phase and the JSC starts to decrease.

6.5 Conclusion

In conclusion, we have shown that blends of [60]PCBM and [70]PCBM can beused as n-type semiconductor in combination with rr-P3HT for bulk hetero-junction solar cells. By replacing the n-type semiconductor with a [60]PCBM/[70]PCBM mixture, without any optimization in the processing conditions, de-vice performance of these type of solar cells shows very little or no sensitivityto the blend composition. The use of a mixture of a C60 fullerene derivative anda C70 fullerene derivative as the semiconductor composition in photovoltaic de-vices is advantageous because it is much less expensive to prepare a mixture ofa C60 fullerene derivative and a C70 fullerene derivative compared to a pure C60

or C70 fullerene derivative. Preparation of a pure C60 or C70 fullerene derivativenecessarily requires expensive purification processes at some stage in the syn-thesis. This purification is necessary because known processes for producingnative fullerene feedstocks produce mixtures of fullerenes.

The fullerenes bis-adducts, used as a n-type semiconductors in combinationwith rr-P3HT, are real candidates for increasing the power conversion efficiencyof the bulk heterojunction solar cells because they show a clear enhancementin VOC (with approximately 80 mV) compared with P3HT:[60]ThCBM solar celldevices.

Finally, we demonstrate that by introducing some trace amounts of pure C60

in a given fullerene derivative n-type semiconductor does not strongly reducethe VOC of the solar cells. Furthermore, by including in the n-type semiconduc-tor trace amounts of PCBM derivatives of fullerenes higher than C70 ([Cn]PCBM, where n>70), which are deep traps, the device performance is drastically al-tered due to the significant decrease of the VOC as a result of a lower lyingLUMO level for these PCBM derivatives.






[4] K. Kim, J. Liu, M. A. G. Namboothiry, D. L. Carroll, Roles of donor and acceptor nan-odomains in 6% efficient thermally annealed polymer photovoltaics, Applied Physics Let-ters 90 (2007), 163511.

[5] J. Y. Kim, K. Lee, N. E. Coates, D. Moses, T. -Q. Nguyen, M. Dante, A. J. Heeger, Effi-cient tandem polymer solar cells fabricated by all-solution processing, Science 317, (2007),222.

[6] T. D. Anthopoulos, C. Tanase, S. Setayesh, E. J. Meijer, J. C. Hummelen,P. W. M. Blom, D. M. de Leeuw, Ambipolar organic field-effect transistors based on asolution-processed methanofullerene, Advanced Materials 16 (2004), 2174.

[7] T. D. Anthopoulos, D. M. de Leeuw, E. Cantatore, S. Setayesh, E. J. Meijer, C. Tanase,J. C. Hummelen, P. W. M. Blom, Organic complementary-like inverters employingmethanofullerene-based ambipolar field-effect transistors, Applied Physics Letters 85(2004), 4205.

[8] C. Waldauf, P. Schillinsky, M. Perisutti, J. Hauch, C. J. Brabec, Solution-processed or-ganic n-type thin-film transistors, Advanced Materials 15 (2003), 2084.

[9] H. Wang, O. Hofmann, R. Das, E. M. Barrett, A. J. deMello, J. C. deMello,D. D. C. Bradley, Integrated thin-film polymer/fullerene photodetectors for on-chip mi-crofluidic chemiluminescence detection, Lab on a chip 7 (2007), 58.

[10] M. C. Scharber, D. Muhlbacher, M. Koppe, P. Denk, C. Waldauf, A. J. Heeger,C. J. Brabec, Design rules for donors in bulk-heterojunction solar cells towards 10% energy-conversion efficiency, Advanced Materials 18 (2006), 789.

[11] L. M. Popescu, P. van’t Hof, A. B. Sieval, H. T. Jonkman, J. C. Hummelen, Thienyl ana-log of 1-(3-methoxycarbonyl)propyl-1-phenyl- [6,6]-methanofullerene for bulk heterojunc-tion photovoltaic devices in combination with polythiophenes, Applied Physics Letters 89(2006), 213507.

[12] D. C. Montgomery, Design and Analysis of Experiments, 6th ed. (John Wiley & Sons,Inc., New York, 2005), 34-43.

[13] R. L. Mason, R. F. Gunst, J. L. Hess, Statistical Design and Analysis of Experiments with


Application to Engineering and Science, 2nd ed. (John Wiley & Sons, Inc., New Jersey,2003).


[15] K. L. Mutolo, E. I. Maio, B. P. Rand, S. R. Forrest, M. E. Thompson, Enhanced open-circuit voltage in subphthalocyanine/C60 organic photovoltaic cells, Journal of the Amer-ican Chemical Society 128 (2006), 8108.

[16] A. Gadisa, M. Svensson, M. R. Anderson, O. Inganas, Correlation between oxida-tion potential and open-circuit voltage of composite solar cells based on blends of polythio-phenes/fullerene derivative, Applied Physics Letters 84 (2004), 1609.

[17] C. M. Ramsdale, J. A. Barker, A. C. Arias, J. D. MacKenzie, R. H. Friend, N. C. Green-ham, The origin of the open-circuit voltage in polyfluorene-based photovoltaic devices, Jour-nal of Applied Physics 92 (2002), 4266.

[18] L. J. A. Koster, V. D. Mihailetchi, P. W. M. Blom, Ultimate efficiency of polymer/fullerenebulk heterojunction solar cells, Applied Physics Letters 88 (2006), 093511.

[19] F. B. Kooistra, J. Knol, F. Kastenberg, L. M. Popescu, W. J. H. Verhees, J. M. Kroon,J. C. Hummelen, Increasing the open circuit voltage of the bulk-heterojunction solar cellsby raising the LUMO level of the acceptor, Organic Letters 9 (2007), 551.


[21] M. M. Mandoc, F. B. Kooistra, J. C. Hummelen, B. de Boer, P. W. M. Blom, Effectof traps on the performance of bulk heterojunction organic solar cells, Applied PhysicsLetters 91 (2007), 263505.

Summary

The direct conversion of the sunlight into electricity is the most elegant processto generate environmentally-friendly renewable energy. Plastic solar cells offerthe prospect of flexible, lightweight, lower cost of manufacturing, and hope-fully an efficient way to produce electricity from sunlight. Since the discoveryof photo induced charge transfer from a conjugated polymer to C60, followedby introduction of the bulk heterojunction concept, this material combinationhas been extensively studied in organic solar cells leading to a power conver-sion efficiency approaching 6% nowadays. A typical bulk heterojunction solarcell (see Figure 1) consists of a photoactive layer sandwiched between two dif-ferent electrodes. Clearly, if the cell is to be exposed to light, at least one ofthe electrodes must be (semi-)transparent. The photoactive layer is based on ablend of an electron donating material (p-type semiconductor) and an electronaccepting material (n-type semiconductor) forming nanostructured bicontinu-ous interpenetrating networks. The fundamental steps in molecular solar powerconversion are: absorption of a photon creating an exciton (bound electron-holepair); exciton diffusion; charge transfer at donor/acceptor interface; dissociationand separation of the carrier pair; charge carrier transport to the correspondingelectrodes; collection of charges by the electrodes.

Figure 1: Device configuration of a bulk heterojunction solar cell with a photoactive layerconsisting of a blend of rr-P3HT (the donor) and methanofullerene (the acceptor).

91

92 Summary

Although significant progress has been made for bulk heterojunction pho-tovoltaic devices, the current efficiency of these solar cells does not guarantee(large scale) commercialization. The most commonly used n-type semiconduc-tor in the bulk heterojunction solar cells is [60]PCBM. Up till now [60]PCBMremains the best performing soluble fullerene derivative in combination withrr-P3HT. Improving the performance and lifetime of bulk heterojunction solarcells requires careful design and material engineering, and more insight in theoperation of these devices. This thesis addresses the possibility of using newfullerene derivatives in organic bulk heterojunction photovoltaic devices, anddiscusses the preparation, and the morphological and electrical characterizationof devices made from rr-P3HT and a library of new methanofullerenes.

In Chapter 2, the experimental set-up used to study the device characteristicsof bulk heterojunction solar cells based on different n-type semiconductors isoutlined. A short overview of the materials used and the protocols that we de-veloped for reproducible device preparation and characterization is presented.

The preparation of an analogue of [60]PCBM with the aim of improving mis-cibility with polythiophenes donors, especially rr-P3HT is described in Chapter3. In this compound the phenyl group from [60]PCBM is replaced by a thienylgroup, resulting in [60]ThCBM (see Figure 1). The electrical characterizationof pristine [60]ThCBM films reveals that its electron transport properties equalthose of [60]PCBM, which indicates that the electron transport properties arenot altered upon replacing the phenyl with a thienyl group. When [60]ThCBMwas blended with MDMO-PPV (1:4 w/w) the power conversion efficiency waslower compared to that of the MDMO-PPV:[60]PCBM solar cells, due to an un-balanced charge transport (the hole mobility of MDMO-PPV is not strongly en-hanced upon blending with [60]ThCBM). In contrast, when using [60]ThCBMin combination with P3HT (1:1 w/w) as the donor counterpart, the improvedmorphology, after thermal annealing and the optimized absorber composition,leads to a power conversion efficiency up to 3.0%. The increase of efficiency,after thermal annealing, is due to the increasing crystallinity of P3HT and hencethe enhancement of the hole mobility in the P3HT phase of the blend by twoorders of magnitude relative to as cast devices.

In Chapter 4 we present a more detailed study of the morphology and electri-cal properties of photovoltaic devices consisting of blends of rr-P3HT as electrondonor and [60]ThCBM as an electron acceptor. The solar cells were fabricatedby depositing the photoactive layer from solvents with different boiling points.In this way the growth rate of the films was controlled. It is shown that slowgrowing of the active layer films in P3HT:[60]ThCBM (1:1 w/w ) blends, by

93

using solvents with a high boiling point, facilitates the formation of highly crys-talline films. This enhances the interchain interactions in the P3HT phase, andimproves the hole transport in the blend by an order of magnitude. This leadsto a well balanced charge carrier transport in the blend. The difference in powerconversion efficiency between slow and fast grown films is mainly determinedby the difference in the fill factor and open-circuit voltage of the cells. Althoughthe open circuit voltage tends to be lowered (with approximately 40 mV) in de-vices with active layers with a higher degree of molecular organization, it isthe significant increase in the fill factor (up to 70%) which gives much higher(13% relative) power conversion efficiencies. The benefit of better charge trans-port for slow growing films makes it possible to fabricate thicker films (≈ 300nm) without significant loss in fill factor, and therefore it allows for maximizinglight absorption in these blends.

In spite of the high potential of polymer photovoltaic cells, a major chal-lenge remains the improvement of their stability under operational conditions.Therefore, in Chapter 5 the stability of the slow growth P3HT:methanofullerenesolar cells devices has been investigated. The methanofullerenes used in thischapter are [60]PCBM, [60]ThCBM , and [70]PCBM (see Figure 1). Electricaland optical properties of these devices were monitored for a test period up to1000 hours under continuous illumination. The experimental results reportedin this chapter show that the P3HT:[70]PCBM cell was found to be the moststable one, with a power conversion efficiency decreasing to about 82% of ini-tial performance, after 1000 hours of operation under continuous illuminationat room temperature. The hole transport through the blend (P3HT phase) re-mains unaffected during the stability test, therefore the hole mobility is not thecause for the degradation observed in the solar cell devices. The decrease inefficiency for solar cells containing [70]PCBM, [60]PCBM, and [60]ThCBM canbe explained by the fact that during illumination also fullerenes crystallize andform larger crystallites in time, which could result in a decrease of the electronmobility in the blend as a consequence of interrupted percolation paths. The[70]PCBM (as a mixture of isomers) has less tendency to crystallize comparedto [60]PCBM and [60]ThCBM, therefore this kind of degradation of cell perfor-mance as a result of phase separation with this higher fullerene derivative in theactive layer is very likely reduced. Additional results presented in this chap-ter indicate that the top electrodes used do not significantly affect the stabilityof P3HT:methanofullerene devices under continuous illumination. The samedegradation for photovoltaic parameters, after 360 hours of continuous illumi-nation, was observed upon using Al as top electrode instead of Sm/Al. This is

94 Summary

another indication that the degradation in the cell parameters, as seen in thisspecific study, might be caused by the changes that occur mainly within the ac-tive layer of the solar cells during continuous illumination.

In Chapter 6 we present experimental studies of the bulk heterojunction solarcells containing, concomitantly in the same device, two different types of n-typesemiconductors [60]PCBM and [70]PCBM (of which one as a mixture of iso-mers, in this case) in combination with a donor material, here rr-P3HT. It shouldbe noted that both the HOMOs and LUMOs of [60]PCBM and [70]PCBM are(close to) identical. These two molecular semiconductors do have very differentsolubilities and precipitation behaviors as pure components, but using blendsof the two surprisingly does not seem to affect morphology and does not com-plicate the processing conditions for devices. Simply replacing the n-type withthe [60]PCBM/[70]PCBM mixture with no change in processing conditions ordevice parameters results in identical device performance within experimentalerror. Next, we show that this concept generalizes to other fullerene derivativetypes. The fullerene bis-adducts, used as n-type semiconductors in combina-tion with rr-P3HT, are promising candidates for increasing power conversionefficiency of the solar cells because they show a clear enhancement in the VOC

compared with the ”state of the art” P3HT:[60]PCBM solar cells. Finally, theeffect that certain impurities in the n-type semiconductor composition (such aspure C60 or PCBM derivatives of fullerenes higher than C70) have on the electri-cal properties of the solar cell devices is quantified. By including in the n-typesemiconductor trace amounts of PCBM derivatives of fullerenes higher than C70

the device performance is drastically influenced because they act as deep elec-tron traps.

Samenvatting

De directe omzetting van zonlicht in elektriciteit met behulp van zonnecellenis het elegantste proces voor de productie van duurzame energie. Zonnecellengemaakt van plastics kunnen in principe flexibel, licht en goedkoop zijn, en kun-nen in de toekomst hopelijk met voldoende efficientie licht omzetten in elek-triciteit. De ontdekking van ladingsoverdracht van een geleidend polymeernaar C60 onder invloed van licht, gevolgd door de introductie van het conceptvan de bulkheterojunctie, heeft geleid tot uitvoerige studies van deze combi-natie van materialen in organische (plastic) zonnecellen. Momenteel wordenrendementen tot 6% gehaald voor de omzetting van het invallende lichtvermo-gen in elektrisch vermogen. Een plastic zonnecel volgens het principe van debulkheterojunctie is afgebeeld in Figuur 1. Gewoonlijk bestaat deze uit een fo-toactieve laag tussen twee verschillende elektroden, waarvan er een licht moetkunnen doorlaten. De fotoactieve laag zelf bestaat uit een mengsel van een elek-trondonerend materiaal (p-type halfgeleider) en een elektronaccepterend ma-teriaal (n-type halfgeleider) die samen een driedimensionaal netwerk vormenwaarin de beide materialen tot op nanometerschaal met elkaar gemengd zijn.

Figure 1: Structuur van een zonnecel volgens het principe van de bulkheterojunctie, meteen fotoactieve laag die bestaat uit rr-P3HT (de donor) en een methanofullereen (de ac-ceptor).

95

96 Samenvatting

De fundamentele stappen in het moleculaire proces dat in deze cel leidt tot deomzetting van licht in elektrische stroom zijn: absorptie van een foton in eenvan beide materialen genereert een exciton (een koppel van een elektron en eengat, een positieve lading); diffusie van het exciton naar een grensvlak tussende beide materialen; ladingsoverdracht bij het donor/acceptor grensvlak; dis-sociatie en scheiding van de beide ladingen; transport van de beide ladingen,ieder door een van beide materialen, naar de juiste elektrode; verzamelen vande ladingen door de elektrodes.

Alhoewel behoorlijke vooruitgang is geboekt voor de fotovoltaısche syste-men gebaseerd op bulkheterojuncties is het succesvol commercialiseren ervanmomenteel nog niet goed mogelijk doordat het rendement van de cellen nog niethoog genoeg is. De meest gebruikte n-type halfgeleider in de bulkheterojunc-tiezonnecellen is [60]PCBM, omdat dit tot op heden het fullereenderivaat is datvoldoende oplosbaar is en in combinatie met rr-P3HT de beste resultaten geeft.Het verder verbeteren van de levensduur en het rendement van de bulkhetero-junctiezonnecellen vereist, in combinatie met gedetailleerde studie van de werk-ing van het systeem, een systematische aanpak voor wat betreft het onderzoekaan en het varieren van de gebruikte materialen in de fotoactieve laag. In ditproefschrift worden diverse nieuwe fullereenderivaten getest in de organischezonnecellen gebaseerd op de combinatie rr-P3HT/methanofullereen, en diverserelevante eigenschappen van deze cellen, zoals morfologie en geleiding van deactieve laag, worden onderzocht.

De experimentele opstelling die gebruikt is om de eigenschappen van debulkheterojunctie zonnecellen, met daarin de diverse n-type halfgeleiders, teonderzoeken wordt besproken in Hoofdstuk 2. Tevens wordt een kort overzichtgegeven van de gebruikte materialen, van de meettechnieken en van de metho-den die zijn ontwikkeld om de assemblage van de zonnecellen reproduceerbaarte maken.

Hoofdstuk 3 beschrijft de ontwikkeling van een variant van [60]PCBM metals doel het verbeteren van de menging van het methanofullereen met polythio-pheen donoren, zoals rr-P3HT. In deze variant is de benzeenring in [60]PCBMvervangen door een thiofeengroep, waardoor [60]ThCBM ontstaat (Zie Figuur1). De elektrische metingen aan films van [60]ThCBM laten zien dat het elek-tronengeleidende vermogen van beide verbindingen gelijk is, dus dat de elek-trontransporterende eigenschappen niet veranderen door het vervangen van debenzeengroep door een thiofeenring. Een mengsel van [60]ThCBM met MDMO-PPV in een gewichtsverhouding van 1:4 gaf zonnecellen met een lager rende-ment dan die met [60]PCBM, omdat er een te groot verschil was in het transport

97

van de beide ladingen. Het transport van gaten in de MDMO-PPV fase bleekniet substantieel te stijgen na menging met [60]ThCBM, dit in tegenstelling totwat gebeurt in mengsels met [60]PCBM. De combinatie van [60]ThCBM metde donor P3HT (gewichtsverhouding 1:1) gaf echter zonnecellen met een ren-dement van 3.0% na optimalisatie van de samenstelling van de actieve laag enhet verkrijgen van een verbeterde morfologie na verwarmen van deze laag. Destijging van het rendement na verwarmen is het gevolg het meer kristallijn zijnvan de P3HT fase. Hierdoor treedt een stijging van twee ordes van grootte opvan de gatengeleiding in de P3HT fase ten opzichte van die in onbehandeldeactieve lagen.

Een meer gedetailleerde studie van de morfologie en de elektrische eigen-schappen van fotovoltaische systemen bestaande uit mengsels van het elek-trondonerende rr-P3HT en de electronacceptor [60]ThCBM wordt beschrevenin Hoofdstuk 4. Oplossingen van de organische materialen in een aantal oplos-middelen met verschillende kookpunten werden gebruikt voor het maken vande actieve laag van de zonnecellen. Dit maakt het mogelijk om de groeisnel-heid van de film te controleren. De resultaten toonden aan dat in het geval vanP3HT:[60]ThCBM mengsels (in een 1:1 gewichtsverhouding) het langzaam latengroeien van de actieve laag de vorming van meer kristallijne films bevordert.Deze films worden verkregen als een oplosmiddel met een hoog kookpuntwordt gebruikt.In dergelijke films is er een betere interactie tussen de polymeer-ketens in de P3HT fase, waardoor het gatentransport in het mengsel met eenorde van grootte verbetert en een goede balans wordt verkregen in het transportvan de beide ladingen. Het verschil in rendement tussen films die langzaam ofsnel gegroeid zijn wordt voornamelijk bepaald door het verschil in ”fill factor”en open klemspanning tussen de beide soorten films. De open klemspanningdaalt gewoonlijk met ongeveer 40 mV in films die meer moleculair geordendzijn, maar deze daling wordt gecompenseerd door een significante stijging vande fill factor, waardoor een netto toename van ongeveer 13% (relatief) in ren-dement wordt bereikt. Het verbeterde ladingstransport in de actieve lagen dielangzaam gevormd worden maakt het ook mogelijk om dikkere films (300 nm)te maken zonder dat de fill factor al te veel kleiner wordt, zodat de lichtabsorptiein de films kan worden geoptimaliseerd.

De huidige resultaten die worden verkregen met de plastic fotovoltaıschecellen, zoals de hier beschreven zonnecellen, zijn weliswaar veelbelovend, maargaan voorbij aan de noodzaak om ook de stabiliteit van de zonnecellen overeen langere operationele periode te verbeteren. De stabiliteit van de zon-necellen met langzaam gevormde P3HT:methanofullereen lagen is bestudeerd

98 Samenvatting

in Hoofdstuk 5. De gebruikte methanofullerenen zijn [60]PCBM, [60]ThCBM en[70]PCBM (zie Figuur 1). De elektrische en optische eigenschappen van decellen werden gevolgd gedurende een periode van maximaal ongeveer 1000uur, waarbij de cellen continu belicht werden. De resultaten laten zien dat indit geval de P3HT[70]PCBM zonnecel de stabielste is, na 1000 uur werkzaamte zijn onder continue belichting, met een daling van het rendement naar82% van de beginwaarde. Het gatentransport in de P3HT fase bleef stabielgedurende de testperiode, wat betekent dat verandering in de mobiliteit vande gaten niet de oorzaak kan zijn van de verslechtering van de werking vande zonnecel. De rendementsdaling van de zonnecellen kan wel worden verk-laard door langzame kristallisatie van een deel van de aanwezige fullereen-moleculen. Hierdoor kan de elektronmobiliteit in het mengsel dalen doordateen deel van de aanwezige transportwegen wordt onderbroken. In vergelijk-ing met [60]PCBM en [60]ThCBM heeft [70]PCBM, dat een mengsel van iso-meren is, veel minder de neiging om kristallen te vormen, zodat deze degra-datieroute als gevolg van fasescheiding in het geval van actieve lagen met ditmethanofullereen veel minder aanwezig is. Verdere resultaten in dit hoofdstuktonen aan dat de topelektroden geen belangrijke invloed hebben op de stabiliteitvan de P3HT/methanofullereen zonnecellen. Na 360 uur continue belichtingbleken de diverse parameters van de zonnecellen in het geval van een Sm/Alelektrode dezelfde daling te geven als bij een Al elektrode. Dit is opnieuw eenaanwijzing dat de verslechtering van de eigenschappen van de zonnecellen,zoals gevonden in deze studie, voornamelijk het gevolg is van veranderingenin de actieve laag.

In Hoofdstuk 6 worden experimenten beschreven aan bulkheterojunctiezon-necellen waarvan de actieve laag naast het donormateriaal P3HT twee verschil-lende soorten n-type halfgeleiders bevat, namelijk [60]PCBM en [70]PCBM (waar-bij de laatste uiteraard als mengsel van isomeren aanwezig is). De positievan zowel de HOMO als de LUMO van deze twee verbindingen is nagenoeggelijk aan elkaar. Deze twee moleculaire halfgeleiders hebben ieder afzonderlijkechter behoorlijk verschillende oplosbaarheidseigenschappen en een verschil-lende tendens tot precipiteren. Opmerkelijk genoeg blijkt at het gebruik vanmengsels van beide stoffen in de actieve laag niet leidt tot veranderingen inde morfologie noch problemen geeft bij het fabriceren van de films. Het een-voudigweg vervangen van de zuivere n-type halfgeleider door een mengsel van[60]PCBM en [70]PCBM zonder verdere wijziging van de procescondities of an-dere parameters levert cellen op die binnen de experimentele foutenmarge eengelijk rendement geven. Verdere experimenten tonen aan dat dit principe alge-

fullerene based organic solar cells

Documents