Influence of the exciton blocking layer on the stability of layered organic solar cells

(1)

Inﬂuence of the exciton blocking layer on the stability of layered organic solar cells

Y. Lare

^a,1

, B. Kouskoussa

^b

, K. Benchouk

^b

, S. Ouro Djobo

^a,1

, L. Cattin

^c

, M. Morsli

^a

, F.R. Diaz

^d

, M. Gacitua

^d

, T. Abachi

^e

, M.A. del Valle

^d

, F. Armijo

^d

, Gaston A. East

^d

, J.C. Bern ede

^a,ⁿ

aLAMP, Universite´ de Nantes, Nantes Atlantique Universite´ de Nantes, Faculte´ des Sciences et des Techniques, 2 rue de la Houssiniere, BP 92208, Nantes F-44000, France

bLPCMME, Faculte´ des Sciences, Universite´ d’Oran Es-Se´nia, 31100 Oran, Alge´rie

cIMN-CNRS, Universite´ de Nantes, Nantes Atlantique Universite´s, 2 rue de la Houssiniere, BP 92208, Nantes F-44000, France

dP. Universidad Cato´lica de Chile, Facultad de Quı´mica, Av. Vicun˜a Mackenna 4860, Santiago, Chile

eEcole Normale, Supe´rieure E.N.S. 16000 Kouba, Alge´rie

a r t i c l e i n f o

Article history:

Received 21 May 2010 Received in revised form 23 August 2010

Accepted 16 November 2010

Keywords:

A. Thin ﬁlms A. Interfaces A. Fullerenes B. Vapour deposition D. Electrical properties

a b s t r a c t

The life-time of multi-layer organic solar cells based on the couple donor acceptor copper phthalocyanine/

fullerene is studied as a function of the nature of the exciton blocking layer (EBL). It is shown that organic EBL are more efficient than are the inorganic In2S3EBLs. Moreover among the organic EBL, Alq3is the most efficient EBL protecting layer. An organic solar cell’s lifetime depends on oxygen- and water- contamination of the organic materials. The solar cell’s degradation may correspond to bulk or interface phenomena. Using equivalent electrical schemes of solar cell diodes, we show that the structure degradation is mainly related to bulk modification. It is proposed that oxygen- and water-diffusion into the C60induce a large increase in its resistivity and, therefore an increase in the series resistance, which decreases the solar cell efficiency. In the case of In2S3EBLs, the degradation law predicts that with time two different phenomena will be present. The classical oxygen- and water-diffusion into the organic material, during the first hour of air exposure, leads to a modification in the In2S3EBL/organic interface properties.

1. Introduction

Organic solar cells have the potential advantages of light weight and low cost processing[1,2]. Recently an integrated roll-to-roll process has been presented[3]. However, it is not only the power conversion efficiency (5–6%) [4,5] of these cells that should be improved, but also their lifetime that is far from satisfactory. More efforts are dedicated to increase their efficiency[6,7], while the ageing of the cells is rather rarely studied. As far as the stability is concerned, very interesting results have been reported by Bund- gaard and Krebs[8], Krebs[9]and Jorgensen et al.[10]. The most important degradation causes are photodegradation and environmental parameters. Small molecules, such as metal-phthalocya- nines used in layered organic solar cells, exhibit good photostability and therefore the lifetime of these cells is mainly dependent on their environment. The most important environmental parameters that influence the lifetime of organic solar cells are the diffusion of molecular oxygen and water into the active layer of the devices through the electrodes, mainly the upper

electrode. For these reasons, the permeability of the electrodes towards molecules such as oxygen and water is an important parameter. A possible solution to the problem of oxygen and water diffusion into the active organic layers is the use of barrier layers with low oxygen and water permeability. These layers can be either passive encapsulating layers or they may comprise of a layer present in the solar cell itself, such as the exciton blocking layer, or both. As a matter of fact, signiﬁcant efﬁciency improvements of layered cells, based on an electron donor/electron acceptor junction, have been achieved through the introduction of a buffer layer at the interface electron acceptor/cathode. This buffer layer has been called the exciton blocking layer (EBL) [11]. It has been proposed that this layer blocks exciton transport towards the cathode, which avoids exciton quenching at the contact organic/

cathode and protects the electron acceptor layer during the cathode deposition by evaporation[11]. The most effective exciton blocking layer is the bathocuproine (BCP). However, the lifetime of solar cells with BCP is only a few hours or even less when exposed to air without encapsulation[12]. This short lifetime has been attributed to the instability of BCP, as it is readily crystallized in a moist environment[13]. Some studies have shown that the aluminium tris(8-hydroxyquinoline) (Alq3) is appropriate to replace BCP [12,13]. They have shown that it allows an improvement in the lifetime of solar cells. It is suggested that this lifetime improvement Contents lists available atScienceDirect

journal homepage:www.elsevier.com/locate/jpcs

Journal of Physics and Chemistry of Solids

doi:10.1016/j.jpcs.2010.11.006

nCorresponding author.

E-mail address: jean-christian.bernede@univ-nantes.fr (J.C. Bernede).

1Present address: Laboratoire d’Energie Photovoltaı¨que, Universite´ de Lome´, Lome´, Togo.

(2)

results probably from the reduction in oxygen and water diffusion and also from an efﬁcient blocking of the diffusion of the cathode into the active organic layers.

In the present manuscript we compare the effect of different EBLs, either organic or inorganic, on the solar cell’s lifetime. We focus on the degradation of organic solar cells exposed to air. The samples were only illuminated when tested in duration between the tests, they were stored in daylight. The ageing process of the cells is studied by examining the evolution with time of their current density–tension (J–V) characteristics. We use the theoretical interpretation of experimental results to discuss the origin of different behaviours.

2. Experimental techniques

The cells were based on the well known donor/acceptor couple:

copper phthalocyanine (CuPc)/fullerene (C60). The CuPc/C60junc- tions are fabricated on pre-cleaned glass substrate coated with transparent conductive indium tin oxide (ITO) anode. Before organic deposition under vacuum was performed, the ITO substrate was covered with an ultra-thin gold layer (0.5 nm)[14,15]. We have already shown that this layer increases signiﬁcantly the cells efﬁciency through a good matching of the band structure anode/

CuPc[16]. All different films have been deposited in the same run in a vacuum of 10⁴Pa. The thin film deposition rates and thickness were estimatedin situwith a quartz monitor. The deposition rate and final thickness were 0.05 nm/s and 35 nm in the case of CuPc, respectively, 0.05 nm/s and.40 nm in the case of C60. The material used as a cathode buffer layer should be transparent across the solar spectrum acting as a spacer between the photoactive films and the cathode. It must transport charge, allowing low series resistance. Its HOMO (highest occupied molecular orbital) and LOMO (lowest unoccupied molecular orbital) values should permit the buffer layer to act as an efficient exciton blocking layer.

Following these requirements, different EBLs have been probed:

BCP, Alq3, an original organic layer—(Z)-5-(4-chlorobenzylidene)- 3-(2-ethoxyphenyl)-2-thioxothiazolidin-4-one, that we have called (CBBTZ) and also an inorganic thin ﬁlm, In₂S₃. In₂S₃ has been chosen because it is known to be a n-type, large band gap semiconductor easily deposited in an amorphous form. All these ﬁlms have a large band gap (EgZ2.5 eV) and they also transport electrons. It has been shown already that the optimum organic EBL thickness is around 9 nm. As for the inorganic EBL different thicknesses (9–26 nm) have been probed in order to optimise the In2S3thickness.

After EBL deposition, the aluminium top electrodes were thermally evaporated, without breaking the vacuum, through a mask with 210 mm²active areas.

Without the protecting layer the instability of solar cells causes rapid degradation of all performances, the non-encapsulated devices are practically dead after about 8 h in air. In order to mitigate this instability, before breaking the vacuum, an encapsulating layer of amorphous selenium (Se-a) of approximately 50 nm, is evaporated thermally. The selenium protective coating layer (PSe) has been proven to be efﬁcient in protecting the under layers from oxygen and water vapour contamination[17], at least during the early hours of room air exposure[18]. Encapsulation impedes the process, but does not eliminate the degradation process. Therefore the protective layer, increasing the solar cells lifetime, prolongs the duration of the degradation process and thus improves the preci- sion of the study on the EBL effect on this process.

Finally, the structures that were used are: glass/ITO(100 nm)/

Au(0.5 nm)/CuPc(35 nm)/C60(40 nm)/EBL(9 nm)/Al(120 nm)/PSe. Electrical characterizations were performed with an automated I–Vtester, in the dark and in light—global AM1.5 simulated solar

illumination. Performances of photovoltaic cells were measured using a calibrated solar simulator (Oriel 300 W) at 100 mW/cm² light intensity adjusted with a PV reference cell (0.5 cm²CIGS solar cell, calibrated at NREL, USA). Measurements were performed at an ambient atmosphere. All devices were illuminated through TCO electrodes.

The ageing effect in ambient air was studied as follows: multiple tests under AM1.5 light were performed at different duration from 5 min to 10 days. Devices were stored in the room daylight at open circuit voltage between measurements.

3. Experimental results

TheJ–Vcharacteristics of the different cells evaluated immea- diately after the latter’s realization are shown inFig. 1. The open circuit voltageV_ocbarely depends on the EBL, while the short circuit currentJscvaries signiﬁcantly with it. Organic buffer layers allow the achievement of better results than that can be obtained with the inorganic EBL. The current of the cells with inorganic EBL is small, with an optimum short circuit currentJs¼1.83 mA/cm², this value is obtained with an In2S3thickness of 15 nm.

Typical degradation curves of air-exposed devices are presented inFig. 2with the example of BCP. One can see that the open circuit voltage does not vary with time and remains constant during the entire experiment, while the variations with time ofJscand of

Z

follows the same evolution. A similar relationship is obtained irrespective of the EBL used. It should be noted that there is also a

-200 0 200 400 600

-7.5 -5.0 -2.5 0.0 2.5 5.0 7.5 10.0

J (mA/cm2 )

V (mV) In S Alq CTTBZEN BCP

Fig. 1.J–Vcharacteristics of solar cells under AM1.5 illumination with different EBLs:&—In2S3;^J—Alq3;D—CTTBZEN andr—BCP.

0 50 100 150 200

0 1 2 3 4 5 6 7

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

Jsc (mA/cm)

t (h)

Jsc Voc FF

η

Voc FF

Jsc η(%)

Voc (v);FF

0.0 0.5 1.0 1.5 2.0 2.5

η (%)

Fig. 2.Evolution with time of the different parameters of an OPV cell with a BCP buffer layer.

(3)

degradation of the ﬁll factor (57–51% in the case of Alq3and 56–47%

in the case of BCP).

Sometimes, a period of device improvement is observed during the ﬁrst part of the device life (Fig. 3). The ﬁrst possibility is that this initial increase can be due to annealing phenomena, which lead to a greater rate of device improvement than that of device degradation initially. Moreover some other hypotheses also can be proposed.

A similar increase during the first hour has already been described by Xi et al.[19]in similar cells based on the couple CuPc/C60. They propose that this effect is due to oxygen diffusion through the Al electrode and due to Al2O3 formation at the interface Al/C60. Their hypothesis is that, in the first few hours, the thin Al2O3layer blocks the excitons and reduces the recombination probability of excitons at the interface, while electrons can pass through this insulating layer by means of tunnelling. Analogously, we have already shown that during the first minutes of air exposure of bare cells, the formation of a thin Al203thin layer at the interface Al/C60allows the increase in the shunt resistance and therefore that in the cell efficiency[20]. When the air exposition time increases, the C60invariably becomes O2contaminated, which, as discussed below, decreases its conductivity and, therefore degrades the cell performances. The improvement effect discussed above is visible in Alq3since, as shown by the study below, it is more efficient in protecting the C60 film from O2 diffusion than are the other CBLs.

Therefore, when kept in room air, device performance system- atically decreases more or less rapidly depending on the blocking properties of the EBL whereas, when kept in vacuum to rule out the effect of oxygen and water, these devices with different organic EBLs show extremly long lifetime. In room air, oxygen and water diffuse through the electrode surface regardless of whether the device is illuminated or not.

Fig. 3 shows normalized energy conversion efﬁciencies as a function of time of cells kept in an ambient atmosphere. It can be seen that the solar cell efﬁciency degradation rates depend on the EBL.

In the case of organic EBLs only a minor degradation was observed in the first hours of air exposure, whereas a rapid degradation occurs when the inorganic In2S3EBL is used. The definition of operational lifetime is usually the period of time that elapses between the initial performance and the point where 50% of the initial performance has been reached[21]. It can be seen fromFig. 3that the operational lifetime is only 4 h with In2S3, while it is 65, 120 and 140 h with BCP, CBTTZEN and Alq3, respectively. The lifetime of the cells with Alq3EBL is more than twice long as that of the OPV cells with BCP EBL. Since the structures of the devices are the same, except for buffer layers used, the lifetime variation can be attributed to the specific properties of the different EBLs. With Alq3the lifetime improvement results probably from the fact that Alq₃ effectively blocks oxygen and water to permeation through the acceptor layer, while the shorter lifetime

associated with the BCP EBL is due to the instability of BCP, as it is readily crystallized in a moist environment[22]. Atom diffusion is faster at the grain boundaries than it is in the bulk of the grains. The fast BCP crystallization facilitates oxygen and water diffusion and therefore leads to cell’s performance degradation. Alq3is more stable after deposition, which avoids similar structure modiﬁcation and therefore delays the cell’s degradation.

The results achieved with CBBTZ indicates that it is nearly as efﬁcient as is Alq3for blocking oxygen and water permeation.

As stated above, the positive effect of the EBL is usually attributed to its ability to avoid exciton quenching at the interface C60/Al and to protect the C60layer from Al diffusion during cathode deposition, and it can also act as optical spacer. Here all the EBL thicknesses are equal and therefore the later effect should be seen in all EBLs that were probed. It is the same reason that the C60

protection from Al contamination should be similar irrespective of the EBL used. As for the exciton blocking efﬁciency: it depends on the EBL band gap. The Alq3band gap (3.2 eV) is smaller than that of BCP (3.5 eV) and it would be a less effective exciton blocker, which can justify the efﬁciency difference achieved with these EBL.

Therefore we have probed multi-layer EBLs. The ﬁrst experimental result is that the entire multi-layer EBLs thickness should be the same as that used (9 nm) of the single-layer EBL. Following this rule, the performance of the cells with multi-layer EBLs is within the range of values achieved with the single-layer EBL, the value being balanced by the thickness ratio of EBL constituents. Similarly, regarding the cells stability, it has already been shown that, when a multi-layer EBL is used, the lifetime of the cells is within the values achieved with the corresponding single-layer EBL[23].

Since all the devices degrade from one degree to the other, oxygen and water diffuse through the electrode surface regardless of whether the device is illuminated or is stored in darkness.

It was already shown that, according to the couple donor/

acceptor used, cells are more or less sensitive to the water or to the oxygen[24–27]. Therefore, to discriminate between the effect of the oxygen and that of the water we stored some samples in one desiccator. An example of BCP is shown inFig. 4. One can see that when the sample is stored in one desiccator the cell degradation is far slower. It means that in presence of water the devices degrade more rapidly. It has already been shown that water diffuses quickly from the top Al to the bottom ITO electrodes[25], which justiﬁes the high sensitivity of the solar cells performance to water.

4. Discussion

While the stability has been addressed, little attention has been devoted to the mechanisms that cause degradation. Knowledge of 0 20 40 60 80 100 120 140 160

0.0 0.2 0.4 0.6 0.8 1.0 1.2

η/η

t (h)

In S BCP CBBTZ

Alq

Fig. 3.Variation with time of the normalized efﬁciency of OPV cells with different EBLs.

0 20 40 60 80 100 120 140 160 180 0.0

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

η/η°

t (h)

air dissecator

Fig. 4.Variation with time, in room air and in a desiccator, of the normalized efﬁciency of OPV cells with BCP as EBL.

(4)

the degradation during storage is vital in the search for ways to eliminate, or at the least to diminish, degradation. Different hypotheses can be proposed:

Water and oxygen diffuse into the device and react with the materials in the OPV, resulting in the degradation of the photovoltaic performance. Upon exposure to water and oxygen, a decrease in conductivity of C60several orders of magnitude was observed[28]. Therefore it is assumed that the conductivity decrease of C60ﬁlm, due to water and oxygen permeating easily through thermally deposited aluminium cathode, might be one of the reasons for the degradation of performance and lifetime of OPV cells.

The transfers of charges at electrodes are sensitive to degradation phenomena. The reduced injection current may result from decreased injection efﬁciency of the contact, due either to the appearance of insulating domains, or to an increase in the energy barrier at the organic/inorganic interface. This behaviour is most readily explained by the hypothesis that, as air exposure progresses, an increasing fraction of organic/inorganic interface becomes an insulater. Such insulating domains could result from some reaction with water and/or oxygen. For instance, an ABL often used is the PEDOT:PSS, which is hygroscopic. It has been suggested that the water absorption by the PEDOT:PSS during air exposure increases the series resistance of the cells [29], which is excluded in the present work; the ABL in use being Au. Also, aluminium is a low work function metal (F_WAl¼4.3 eV) it could be subject to fast oxidation in air and some attempts have been made to use some other anode material[13]and a different cell geometry through inverted structures[26,30,31].

One way to identify the mechanisms that cause degradation is the study of the evolution, with time, of the cells parameters through the examination ofI–Vcharacteristics of the cells, using an equivalent electrical circuit model.

An equivalent circuit model could be helpful in understanding and in the optimisation of organic solar cells, as it provides some information on losses in the cells. The evolution of the I–V characteristics shape, and therefore of the electrical process, monitoring the variation in the cells behaviour with the room air exposure time could be helpful in understanding their ageing process.

The equivalent circuit commonly used to interpret the I–V characteristics of solar cells consists of a photogenerator connected in parallel with a diode, which represents theI–Vcharacteristics in the dark. This corresponds to an ideal model in the absence of parasitic resistances. However, in real organic solar cells, it is necessary to introduce a series resistance,Rs, and a shunt resistance,Rsh(Fig. 5a). In such solar cells the mathematical description of this circuit is given by the following equation:

I¼I0 exp VIRs

nkT

1

þVIRs

Rsh

Iph ð1Þ

The Lambert W-function method has been used to determineRs, the series resistance;Rsh, the shunt resistance;n, the ideality factor of the diode; andIph, the photo-generated current. The Lambert W-function is defined as the solution to the equation W(x)exp[W(x)]¼x. The problem to be solved is the evaluation of a set of five parameters,Rs,Rsh, n,IphandIsin order for it to fit the given experimentalI–Vcharacteristics using a simple diode circuit.

It is necessary to distinguish between two types of series and shunt resistances—these under illuminated conditions and those in the dark[32]. In the illuminated devices, electrons and holes are generated homogeneously over the whole cell area, which induces higher carriers to ﬂow towards their respective electrode. The resistances values under illumination are the most interesting

values for solar cell characterisation and they will be determined carefully.

The presence of a high contact resistance at the interface electrode/organic material should be attributed to a poor band match[33]. This poor band match induces a speciﬁc shape of the I–Vcharacteristics with a visible kink effect. We have shown that it is necessary to introduce a pseudo-Schottky diode in the reverse direction, in the equivalent scheme, for achieving a good ﬁt between the experimental and theoretical schemes[34]. In such a case, the presence of a back-contact barrier at the interface C60/Al (Fig. 5b) should be assumed. With a thermoionic current at this interface, the hole current is

I_b¼ I_b0ðexpðqV_b=kTÞ1Þ ð2Þ whereIb0is the saturation current,Vbis voltage across the back contact,kis the Boltzmann constant, andTis temperature.

The current is negative because the polarity of the C60/Al junction is opposite to the main CuPc/C60junction.

Therefore the current-limiting effect, the ‘‘rollover’’, is due to the back-contact barrier height. It occurs because the total current saturates at the valueJb0[33]. The value ofJb0is the current value where theJ–Vcurve starts to indicate rollover.

Demtsu and Sites[33]have treated the main junction and the back-contact junction as independent circuit elements. When a forward biasVis applied to the circuit, the voltage is divided intoVm

across the main CuPc/C60 junction, Vb across the back-junction TCO/CuPc andIRsacross the series resistance:

V¼VmþVbþIRs

Under illumination the current across the main junction is Im¼Im0ðexpðqVm=nkTÞ1Þ-I_phþVm=R_sh ð3Þ

And through the back contact the current is:

Ib¼ Ib0ðexpðqVb=kTÞ1Þ þVb=R^b_sh ð4Þ Equating Eqs. (3) and (4):

Im0ðexpðqVm=nkTÞ1Þ-IphþVm=Rsh

þI_b0ðexpðqV_b=kTÞ1ÞV_b=R^b_sh¼0 ð5Þ

The parameters Rs, Im0, n, and Rsh of the main diode are calculated in a region far from the saturation currentIb. As stated above,Ib0is the current value, where theJ–Vcurve starts to show rollover. Then Eq. (5) can be solved.

R_S I

R_pl V R_pl

V_m I_ph

V_b Rs

Is Rsh V

Fig. 5.Equivalent circuit of: (a) solar cell with Ohmic contacts, and (b) solar cell with rectifying back-contact.

(5)

Therefore we have used these equivalent electrical schemes in order to achieve a good ﬁt between experimental and theoretical curves. When the classical electrical equivalent scheme ofFig. 5a is used, a good agreement between the experimental and theoretical ﬁtted curves is achieved in the case of an organic EBL (Fig. 6, Table 1), while the same result is obtained in the inorganic In2S3EBL only with fresh OPV cells.

In the case of organic EBLs the variation with time of the different parameters is given inTable 1. It can be seen that, more or less rapidly depending on the organic EBL, then and R_svalues increase, whileRsh,Jph andJsdecrease with elapsed time. These evolutions explain the degradation of OPV cells performance with time. The increase inRscan be attributed to the decrease in the conductivity of C60after oxygen exposure.

Also the contact resistances with electrodes can increase. In organic solar cells, the process of carrier collection is one of the main factors which control the electrical characteristics and the efficiency of the devices. Therefore electrode modification can lead to performance degradation. The electron collection depends in the barrier height at the interface, which, in a first approximation, is equal to the energy difference between the aluminium work functionFAland the lowest unoccupied molecular orbital of the EA (LUMO_EA).F_Alis around 4.3 eV[16]while the LUMO_C60is 4.4 eV, which means that normally there is no barrier at the interface.

In the present work, in organic EBLs, only one diode is necessary for the ﬁtting of theJ–V characteristics, even after ageing and degradation of the cells (Fig. 6b,Table 1). Therefore the ageing process of these organic solar cells is not related to the interface but to bulk effect.

As for In2S3EBL, the poor performances achieved with In2S3EBL can be explained by its band structure. It has been shown earlier that the energy levels BV, of the valance band (LUMO), and BC, for the conduction band (HOMO) are 4.25 and 7.15 eV, respectively [35]. It can be seen inFig. 8that this LUMO (conduction band) is below that of CuPc, which can decrease its efﬁciency as EBL.

During the ageing process of In2S3EBL, we found that the decay curve can be ﬁtted with two-terms function corresponding to an initial fast decay and a second slower decay.

0 100 200 300 400 500 600 700

-5.0x10^-4 0.0 5.0x10^-4 1.0x10^-3 1.5x10^-3 2.0x10^-3 2.5x10^-3 3.0x10^-3

Iph)A(

V (mv)

Fe ITO/Au/CuPc/C60/BCP120/Al t = 0 h.

---- Theoretical curve Experimental curve

0 100 200 300 400 500 600 700 0.0

1.0x10^-4 2.0x10^-4 3.0x10^-4 4.0x10^-4 5.0x10^-4 6.0x10^-4 7.0x10^-4

---- Theoretical curve Experimental curve Iph (A)

v (mv) Fe ITO/Au/CuPc/C60/BCP120/Al t = 168 h

Fig. 6.I–Vcharacteristics under AM1.5 illumination of a solar cell using a BCP EBL, (a) 0 and 168 h old (b), (____) experimental and (----) theoretical curves.

Table 1

Variation with time of the efﬁciency and of the calculated parameters for the cells using different EBL.

t(h) Z/Z1 n Rs(O) Rsh(kO) Js Jph

Fitting with one diode CBBTZ (Z1(%)¼2.04)

0 1 2.4 27 230 1.710⁷ 7.6110⁴

4 0.99 2.4 27 230 1.710⁷ 7.0010⁴

96 0.60 4.0 30 20 510⁶ 5.6610⁴

120 0.49 3.5 40 10 110⁶ 4.8810⁴

BCP (Z1(%)¼2.06)

0 1 2 35 25 1.6510⁷ 3.010⁴

24 0.87 2 45 10 1.210⁷ 3.1710⁴

72 0.42 2 75 10 6.410⁸ 1.4510⁴

96 0.30 4 80 25 1.010⁷ 5.5810⁵

Alq3(Z1(%)¼1.46)

0 1 2.9 35 25 6.010⁷ 4.9710⁴

96 0.76 3.4 95 25 1.010⁶ 4.2410⁴

168 0.40 4 100 5 2.010⁶ 3.5110⁴

In2S3(Z1(%)¼0.43)

0 1 3.6 410 60 810⁷ 1.8410⁴

Fitting with two diodes

t(h) Z/Z(deg) n Rs(O) Rpb Rpm Jsb Jsm Jph

In2S3

5 h 30 min 0.45 3.9 450 12 15 3.010¹⁰ 7.010⁷ 1.410⁴

(6)

Immediately after room air exposure the classical one diode equivalent electrical scheme is sufficient to achieve a good agreement between experimental and theoretical curves (Fig. 7a). After the fast decay of the cell efficiency, it is necessary to use the equivalent electrical scheme with two diodes as illustrated inFig. 7b to achieve a good agreement between experimental and theoretical curves. It means that there is some modification of the interface properties with, diode formation at the interface C60/In2S3. The presence of a thin In2S3layer at the interface C60/Al induces, with time, formation of a barrier. Then, the slow decay, similar to that observed with organic EBL, can be attributed to the bulk effect related to progressive C₆₀ contamination by water and oxygen

Some hypothesis can be proposed to explain the variation with time of the contact C60/In2S3. In2S3has been deposited by classical thermal evaporation. That means that there is some compound decomposition during its heating with preferential deposition of the more volatile element, the S, at the beginning of the deposition.

It means that there is formation of a thin S layer between the C60

and the In₂S₃. It has been shown that S can be used as n-dopant of C60[36]. S introduces an impurity level in the band gap of C60. In the present work, the ultra-thin S layer present at the interface with C60

can progressively reacts with it, since the respective oxidation potentials allow the C60oxidation in the presence of S[37]:

C60→ C60+n + ne 1°= 1,760 V*

S + 2e→ S^2- E =-0,508V C60 + nS → C60+n

+ nS^-2 ΔE°= 1,26 V E

Therefore the possible S-doping of C60 can modify the band diagram of the structure, justifying a reverse diode in the electrical equivalent circuit.

So, while oxygen and water contamination explains the performance degradation of cells with organic EBL, it is necessary to propose two degradation processes in the case of In2S3: ﬁrst, modiﬁcation of the interface properties followed by, oxygen and water contamination.

5. Conclusion

The purpose of measuring the lifetime of a device through a decay curve is to establish qualitatively whether the device is stable and to ascertain how it degrades, also it allows for a comparison with devices prepared from different materials and thus ideally provides a method for improving solar cells stability through design of materials, devices and fabrication methods. As expected, the degradation of air exposed devices can be attributed to oxygen and water induced, light independent, mechanism.

By comparing the results at the different devices it is evident that the Alq3introduction in the solar cell is beneficial to lifetime performance. The most important environmental parameters that influence the lifetime of organic photovoltaic is diffusion of molecular oxygen and water into the active layer of the photovoltaic device In the case of inorganic In2S3 EBL there is also reaction between the C60and the sulphur. For this reason, the permeability of the electrodes towards small molecules like oxygen and water is an important parameter. All the degradation mechanisms result in irreversible changes in the device that impairs the performance and lead to dysfunction of the device. The use of materials that does not readily allow diffusion phenomena to take place has demonstrated significant improvements of the lifetime.

Some of the possible solutions to the problem of oxygen and water diffusion into the cell include the use of barrier layers with low oxygen and water permeability.

Acknowledgement

This work has been ﬁnancially supported by the France-Chile contract ECOS-CONICYT No. C09E02.

References

[1] F.C. Krebs, Sol. Energy Mater. Sol. Cells 93 (2009) 394–412.

[2] F.C. Krebs, T. Tromholt, M. Jorgensen, Nanoscale 2 (2010) 873–886.

[3] F.C. Krebs, T.D. Nielsen, J. Fyenbo, M. Wadstrom, M.S. Pedersen, Energy Environ.

Sci. 3 (2010) 512–525.

[4] W. Cai, X. Gong, Y. Cao, Sol. Energy Mater. Sol. Cells 94 (2010) 114–127.

-400 -200 0 200 400 600 800 1000 -0.0002

-0.0001 0.0000 0.0001 0.0002 0.0003 0.0004 0.0005 0.0006

Iph (A)

ITO/Au/CuPc/C60/In S Al t = 0 h

0 200 400 600 800

-0.00016 -0.00014 -0.00012 -0.00010 -0.00008 -0.00006 -0.00004 -0.00002 0.00000 0.00002 0.00004

% (mX) ITO/Au/CuPc/C60/In₂S_3/Al t = 5 h 30

Fig. 7.I–Vcharacteristics under AM1.5 illumination of a solar cell using a In2S3EBL, (a) 0 and 5 h 30 old (b), (____) experimental and (----) theoretical curves.

BCP CTTBZEN In₂S₃

CuPc C₆₀

Buffer layer Vacuum level

Alq₃

Fig. 8. Energy diagram.

(7)

[5] M.Y. Chan, S.L. Lai, M.K. Fung, C.S. Lee, S.T. Lee, Appl. Phys. Lett. 90 (2007) 023504.

[6] B. Kippelen, J.L. Bredas, Energy Environ. Sci. 2 (2009) 251–261.

[7] M. Helgersen, R. Sendergaard, F.C. Krebs, J. Mater. Chem. 20 (2010) 36–60.

[8] E. Bundgaard, F.C. Krebs, Sol. Energy Mater. Sol. Cells 91 (2007) 954–985.

[9] F.C. Krebs, Organic Electron. 10 (2009) 761–768.

[10] M. Jorgensen, K. Norrman, F.C. Krebs, Sol. Energy Mater. Sol. Cells 92 (2008) 686–714.

[11] B.P. Rand, J. Li, J. Xue, R.J. Holmes, M.E. Thompson, S.R. Forrest, Adv. Mater. 17 (2005) 2714–2718.

[12] Q.L. Song, F.Y. Li, H. Yang, H.R. Wu, X.Z. Wang, W. Zhou, J.M. Zhao, X.M. Ding, C.H. Huang, X.Y. Hou, Chem. Phys. Lett. 416 (2005) 42–46.

[13] P. Vivo, J. Jukola, M. Ojala, V. Chukharev, H. Lemmetyinen, Sol. Energy Mater.

Sol. Cells 92 (2008) 1416–1420.

[14] J.C. Bernede, Y. Berredjem, L. Cattin, M. Morsli, Appl. Phys. Lett. 92 (2008) 083304.

[15] J.C. Bernede, L. Cattin, M. Morsli, Y. Berredjem, Sol. Energy Mater. Sol. Cells 92 (2008) 1508–1515.

[16] A. Godoy, L. Cattin, L. Toumi, F.R. Diaz, M.A. del Valle, G.M. Soto, B. Kouskoussa, M. Morsli, K. Benchouk, A. Khelil, J.C. Bernede, Sol. Energy Mater. Sol. Cells 94 (2010) 648–654.

[17] A. Latef, J.C. Bernede, Phys. Status Solidi (a) 124 (1991) 243–252.

[18] Y. Berredjem, N. Karst, A. Boulmokh, A.H. Gheid, A. Drici, J.C. Bernede, The European Phys. J.: Appl. Phys. 40 (2007) 163–167.

[19] X. Xi, F. Li, Q. Meng, Y. Ding, J. Ji, Z. Shi, G. Li, Sol. Energy Mater. Sol. Cells 94 (2008) 924–929.

[20] N. Karst, J.C. Bernede, Phys. Status Solidi (a) 203 (2006) R70–R72.

[21] G. Dennler, C. Lungenschmied, H. Neugebauer, N.S. Sariciftci, M. Latreche, G. Czeremuszkin, M.R. Wertheimer, Thin Solid Films 511-512 (2006) 349–353.

[22] P. Pumans, A. Yakimov, S.R. Forrest, J. Appl. Phys. 93 (2003) 3693–3695.

[23] Q.L. Song, F.Y. Li, H. Yang, H.R. Wu, X.Z. Wang, W. Zhou, J.M. Zhao, X.M. Ding, C.H. Huang, X.Y. Hou, Chem. Phys. Lett. 416 (2005) 42–46.

[24] M.H. Peterson, S.A. Gevorgyan, F.C. Krebs, Macromolecules 41 (2008) 8986–8994.

[25] K. Norrman, S.A. Gevorgyan, F.C. Krebs, A.C.S. Appl., Mater. Interfaces 1 (2009) 102–112.

[26] F.C. Krebs, Sol. Energy Mater. Sol. Cells 92 (2008) 715–726.

[27] F.C. Krebs, S.A. Gevorgyan, J. Alstrup, J. Mater. Chem. 19 (2009) 5442–5451.

[28] A. Hamed, Y.Y. Sun, Y.K. Tao, R.L. Meng, P.H. Hor, Phys. Rev. B 47 (1993) 10873–10880.

[29] K. Kawano, R. Pacios, D. Poplavskyy, J. Nelson, D.D.C. Bradley, J.R. Durrant, Sol. Energy Mater. Sol. Cells 90 (2006) 3520–3530.

[30] Y. Long, Sol. Energy Mater. Sol. Cells 94 (2010) 744–749.

[31] C.Y. Jiang, X.W. Sun, D.W. Zho, A.K.K. Kyaw, Y.N. Li, Sol. Energy Mater. Sol. Cells 94 (2010) 1618–1621.

[32] D. Pysch, A. Mette, S.W. Glunz, Sol. Energy Mater. and Sol. Cells 91 (2007) 1698–1706.

[33] S.H. Demtsu, J.R. Sites, Thin Solid Films 510 (2006) 320–324.

[34] B. Kouskoussa, M. Morsli, K. Benchouk, G. Louarn, L. Cattin, A. Khelil, J.C. Bernede, Phys. Status Solidi (a) 206 (2009) 311–315.

[35] J.C. Bernede, N. Barreau, S. Marsillac, L. Asman, Appl. Surf, Science 195 (2002) 222–228.

[36] H. Yan, Y. Zou, X. Zhang, Y. Li, G. Chen, J. Mater. Sci. Lett. 20 (2001) 449–451.

[37] S.M. Peregudova, L.I. Dnisovich, E.V. Martynova, M.V. Tsikalova, N. Nvikov Yu, Russian J. Electrochem. 44 (2008) 249–253.