DOI:10.1051/epjap/2012120372
P HYSICAL J OURNAL A
PPLIEDP
HYSICS Regular ArticleEffect of the nature of the anode buffer layer – MoO 3 , CuI or MoO 3 /CuI – on the performances of organic solar cells based on oligothiophene thin films deposited by sublimation
Mohammed Makha1,2, Linda Cattin3, Sanoussi Ouro Djobo1,b, Nicolas Stephant3, Nicole Langlois3, Benoit Angleraud3, Mustapha Morsli4, Mohammed Addou2, and Jean-Christian Bern`ede1,a
1 Universit´e de Nantes, Moltech Anjou, UMR 6200, 2 rue de la Houssini`ere, BP 92208, 44000 Nantes, France
2 Laboratoire Opto´electronique et Physico-chimie des Mat´eriaux, Universit´e Ibn Tofail, Facult´e des Sciences, BP 133, Kenitra 14000, Morocco
3 Universit´e de Nantes, Institut des Mat´eriaux Jean Rouxel (IMN), UMR 6502, 2 rue de la Houssini`ere, BP 92208, 44000 Nantes, France
4 Universit´e de Nantes, Facult´e des Sciences et des Techniques, 2 rue de la Houssini`ere, BP 92208, 44000 Nantes, France
Received: 13 September 2012 / Received in final form: 18 October 2012 / Accepted: 31 October 2012 Published online: 3 December 2012 – cEDP Sciences 2012
Abstract. An oligothiophene having a donor-acceptor-donor chromophore with hydrogen bonding groups is used as electron donor in planar heterojunction organic photovoltaic cells. We focus on the contact between the anode and the oligothiophene. Different anode buffer layers (ABLs) have been used, MoO3 and CuI, alone or coupled with MoO3. The thicknesses were 4 nm and 3 nm for MoO3and CuI respectively.
It is shown that the ABL improves the cells performances. The best results are achieved with the couple MoO3/CuI through an increase of the open circuit voltage and short circuit current. The optical absorption, the surface roughness and the organic film conductivity depend on the ABL. The conductivity of the oligothiophene film is one order of magnitude higher when the ABL is a CuI film. The influence of the ABL can be explained partly by the fact that it raises the anode work function. Nevertheless, the study of the structures ITO/ABL/oligothiophene shows that each ABL exhibits specific advantages and disadvantages.
Therefore the couple MoO3/CuI allows summing up the advantages of both ABLs, MoO3 allows a very good band matching and avoids too high leakage current, while CuI allows achieving high Jsc thanks to its effect on the TTB conductivity.
1 Introduction
Today, organic photovoltaic cells (OPV cells) arouse a very big interest due to their potential owing to their flexibility, lightness and their possible moderate cost. The significant and continuous progresses of OPV cells per- formances demonstrate that OPV cells are a potential avenue to low-cost next-generation solar cells. Two tech- nological routes are mainly used to grow OPV cells, the deposition of polymers and nanoparticles from solution by spin-coating technique and the vacuum sublimation of small molecules [1]. The former route has given the concept of bulk heterojunction solar cells (BHJ), while the latter gives the multi-heterojunction solar cells fam- ily. Each of them has their advantages and disadvantages.
The BHJ concept allows an extended interface between the phases of the organic electron donor and the acceptor
a e-mail:jean-christian.bernede@univ-nantes.fr
b Permanent address: Universit´e de Lom´e, Facult´e des Sci- ences, Laboratoire sur l’´energie solaire, BP 1515, Lom´e, Togo.
which leads to an enhanced generation of free charge car- riers and therefore high solar cells efficiency. However, if solution processing is a relatively fast and low price tech- nique, the stacking of several layers is a great challenge, while vacuum deposition allows for simple fabrication of multilayer structures by successive depositions. Moreover, vacuum process does not need solvent, it allows achieving simply pure films with reproducible properties. A classi- cal multi-heterojunction solar cell is based on an organic junction between two electrodes. One of these electrodes must be transparent, while the other must be highly re- flexive. The organic junction is composed of, at least, one absorbing layer, which is usually an electron donor (ED), its band gap being between 1.5 and 2 eV and an electron acceptor (EA), often the fullerene (C60). For improving the (OPV cells) energy conversion efficiency, new organic photoactive layer can be used. For instance, boarding the quite narrow absorbance domain of usual organic mate- rial can induce short-circuit current improvement. It is also well known that carriers exchange at the interface
Fig. 1.Chemical scheme of TTB.
organic material/electrode can greatly influence device performances. Electrode contacts play a critical role in determining the device efficiencies. Ideally, the donor’s HOMO should match the work function of anode. Rates of charge collection of the electrodes must be fast and elec- tron or hole, selective. Contact selectivity is often achieved using interlayers, called buffer layers (BL), interposed be- tween the electrodes and the organic materials. Usually the anode is the highly transparent indium tin oxide film (ITO) and the cathode is the strongly reflexive aluminum film. Between the EA and the Al films the in- troduction of an exciton blocking layer (EBL) improves significantly the OPV cells performances, while an an- ode buffer layer (ABL) is usually introduced between the ED and the ITO anode. Therefore a classical multi- heterojunction solar cell is as follows ITO/ABL/ED/EA/
EBL/Al.
In the present manuscript we use an original organic donor, while different ABLs, MoO3, CuI, MoO3/CuI, or no ABL are probed. We show that the OPV cell effi- ciency achieved with this original organic material de- pends strongly on the ABL used.
2 Experimental
2.1 Organic solar cells realization
The electron donor used was an oligothiophene, the 5-[2,6-bis(E-2-{3,4-di-n-hexy[2,2:5,2-terthiophen]-5-yl}
vinyl)-4H-pyran-4-ylidene]pyrimidine-2,4,6(1H,3H,5H)tri- one (TTB) (Fig.1), its synthesis has been described in a previous article [2]. This oligomer is based on the donor- acceptor-donor structure, which deserves special interest because of the electronic coupling between the donor and the acceptor through ap-conjugated bridge [3].
Different ABLs have been probed: MoO3, CuI and MoO3/CuI. It is known that simple vacuum deposition of a thin film of MoOx[4] allows achieving good ITO/electron donor interface properties. More recently, it has been shown that CuI also can be used as efficient ABL [5] and we have probed it in the present work. Because of the pre- vious studies, the thickness of MoO3layer was 4 nm while the thickness of CuI layer was 3 nm.
For OPV cells realization, the ITO-coated glass sub- strate was scrubbed with soap, rinsed with distilled wa- ter and next placed in the vacuum chamber (10−4 Pa).
The layers were deposited onto the substrate by subli- mation. The deposited layers were: ABL, TTB, fullerene (C60), bathocuproine (BCP), aluminum and selenium.
The fullerene was the electron acceptor and BCP was utilized as the EBL. The effective area of each cell was 0.10 cm2. The thin films thicknesses were estimated in situ using a quartz monitor after calibration. Finally, the cell arrangement was glass/ITO (100 nm)/ABL/
TTB (Xnm)/C60(40 nm)/BCP (9 nm)/Al (100 nm). For comparison some OPV cells without any ABL have also been studied.
The characteristics of the photovoltaic cells were mea- sured using a calibrated solar simulator (Oriel 300 W) at 100 mW/cm2light intensity adjusted with a reference cell (0.5 cm2 CIGS solar cell, calibrated at NREL, USA).
In order to determine the effect of the ABL on the TTB film conductivity, we have investigated the (J-V) char- acteristics of hole-only devices with MoO3 or CuI ABL.
These devices were grown using high work-function elec- trode buffer layers. The hole-only devices were fabricated by replacing the C60 and the BCP EBL with high work- function MoO3, which is well known to be a hole injector (collector) and a blocking electron layer. Hole-only devices have been made using the same ITO covered glass sub- strate as those used to grow OPV cells. After deposition of the ABL, a TTB film thick of 40 nm has been deposited.
Then the organic film has been covered with a MoO3film with a thickness of 7 nm. Finally aluminum has been used as the top electrode.
2.2 Thin films characterization
In order to understand the effect of the different ABL on the OPV cells performances, different characterization techniques have been used for studying the structures ITO/ABL and ITO/ABL/TTB.
The thin films structures are analyzed by X-ray dif- fraction (XRD) by a Siemens D5000 diffractometer using Kαradiation from Cu (λKα= 0.15406 nm).
The optical measurements were carried out at room temperature using a Cary spectrometer. The film trans- mittance was measured at wavelengths of 1.2–0.30μm.
X-ray photoelectron spectroscopy (XPS) measure- ments (Leybold LHS12, University of Nantes-CNRS) were performed to investigate the composition and chemical state of the thin TTB films deposited onto the anodes.
XPS analyses were performed with a magnesium X-ray source (1253.6 eV) operating at 10 kV and 10 mA. Dur- ing the measurements the vacuum was 10−7Pa, the pass energy for high-resolution spectra was 50 eV. The samples were grounded with silver paste to decrease charge effect.
To measure this charge effect on the measured binding energies, the spectra were referred to the C1s line. The spectrometer was calibrated at 284.5 eV for this line.
The morphology of the different structures used as an- ode was observed through scanning electron microscopy (SEM) with a JEOL 7600F at the “Centre de Micro- caract´erisation de l’IMN, Universit´e de Nantes”. Images in secondary (SEI) and backscattering (BEI) mode have been
Table 1.Photovoltaic characteristics, under light AM1.5 (100 mW cm−2), for ITO/CuI/TTB/C60/BCP/Al cells using different TTB thicknesses.
Anode buffer layer Thickness TTB (nm) Voc (V) Jsc (mA/cm2) FF (%) η(%)
CuI 35 0.29 0.53 26 0.04
CuI 25 0.35 0.79 46 0.128
CuI 15 0.15 1.65 26 0.06
CuI/MoO3 25 0.44 3.08 36 0.49
MoO3 25 0.30 0.43 43 0.059
done. The composition of the films has been studied by electron probe microanalysis (EPMA), the scanning elec- tron microscope being equipped with a BRUKER Quan- tax X-ray microanalysis system; X-rays were detected by a silicone drift detector.
AFM images on different sites of the film were taken ex situ at atmospheric pressure and room temperature.
All measurements have been performed in tapping mode (Nanowizard III, JPK Instruments). Classical cantilevers were used (Type PPP-NCHR-50, Nanosensor). The av- erage force constant and resonance were approximately 14 N/m and 320 kHz, respectively. The cantilever was ex- cited at its resonance frequency.
We have measured the surface energy of ITO covered, or not, by an ABL, the ABL being MoO3or CuI. A Kruss G40 Contact Angle Measuring System G40 has been used to estimate the surface energy of TCO surfaces through the sessile drop method. A drop of liquid with known sur- face tension (water, formamide, ethylene glycol and glyc- erol) is put on the TCO surface and according to the Owens, Wendt, Rabel and Kaelble method, the surface tension, split up into a polar and a disperse fraction, can be deduced from the accurate measure of the angle be- tween the substrate surface and a line which is tangential to the circumference of the drop [6].
3 Experimental results
3.1 Organic photovoltaic cells study
Since we have already shown that the nature of the ABL can modify very significantly the performances of the de- vices using oligothiophenes as electron donor [7], in a first attempt, we have compared the performances of the cells using the different ABLs: MoO3, CuI and no ABL. The re- sults show that the power conversion efficiency of the cells with a CuI ABL is at least twice as that obtained with a MoO3 ABL, while it is more than one order of magni- tude higher than that achieved without ABL. It should be noted that the improvement induced by CuI relatively to MoO3 is mainly related to the open circuit voltage (Voc) and to the short circuit current density (Jsc), while there is no significant improvement of the fill factor (FF).
Further to this result we have deepened the study with the CuI as ABL.
The influence of the thickness of the film of TTB on the cells performances is presented in Table1.
Table 2. Series (Rs) and shun resistance (Rsh), under light AM1.5 (100 mW cm−2), for ITO/CuI/TTB/C60/BCP/Al cells using different TTB thicknesses.
TTB thickness (nm) Rs (ohm) Rsh (ohm)
35 5000 5000
25 6.5 960
15 4 95
The results presented show that the critical TTB thick- ness is 25 nm. Indeed, if we use the classical equivalent electrical scheme of OPV, the shunt resistance, Rsh, which is defined by the slope of the (J-V) curve at J = Jsc mA/cm2, varies significantly with the TTB film thickness, while the series resistance, Rs, which is defined by the slope of the (J-V) curve atJ = 0 mA/cm2, is less dependent on this thickness, except for the thick film.
For TTB films with a thickness of 35 nm, the values of Jsc, FF and also Voc are smaller than those of OPV cells using a TTB film with a thickness of 25 nm, this is in good agreement with the very high value of the series resistance, which is probably due to the low hole mobility of TTB [2]. For thinner films with a thickness of 15 nm, while Jsc is higher, Voc and FF are far smaller than with 25 nm. Here, the film being very thin, the higher Jsc value can be related to the low hole mobility, while the smaller FF and Voc values can be related to the shunt resistance value.
It can be seen in Table 2, that, when the TTB film thickness decreases from 25 nm to 15 nm, Rsh is divided by one order of magnitude, which can justify the small FF and Voc values due to current leakage.
In a last attempt we associated both the ABLs so that the OPV cells were: ITO/MoO3(4 nm)/CuI (3 nm)/
TTB (25 nm)/C60 (40 nm)/BCP (9 nm)/Al. Typical (J-V) characteristics are shown in Figure2.
It can be seen that the double buffer layer (DABL) im- proves significantly the energy conversion efficiency. The best result obtained is Voc = 0.44 V, Jsc = 3.08 mA/cm2 and FF = 36%, which gives an efficiency of 0.49%, while only 0.13% has been obtained with CuI alone (Table 1).
Here Voc and Jsc are improved, while FF is slightly smaller. The Rs and Rsh values are 5 ohms and 300 ohms respectively. The quite limited value of Rsh can justify the small FF value. Finally, in order to check the posi- tive effect of CuI with regard to MoO3 alone, we realized OPV cells with the optimal thickness of TTB (25 nm) and a MoO3 buffer layer. The results are shown in Table 1.
-5,0 0 200 400 600 -2,5
0,0 2,5 5,0 7,5 10,0
-200 J (mA/cm2)
V (mV)
Under light In the darkness Voc = 0.44 V Jsc = 3.08 mA/cm2 FF = 36 %
η = 0.49 %
ITO 4.7 eV CuI 5.4 eV
MoO3
5.9 eV 5.9 eV 3.91 eV
TTB LUMO
HOMO
Fig. 2. (Color online) Typical (J-V) characteristics in the dark (full symbols) and under AM 1.5 illumination (open sym- bols) of ITO/MoO3/CuI/TTB /C60/BCP/Al OPV cell. Inset:
band scheme of the contact anode TTB.
0,00 400 500 600 700
0,02 0,04 0,06 0,08 0,10
O.D.
λ (nm) ITO/MoO3/TTB ITO/MoO3/CuI/TTB ITO/CuI/TTB
Fig. 3. (Color online) Optical density spectrum of TTB thin films (35 nm) deposited onto different anodes.
It can be seen that, even after optimization of the thick- ness of the TTB film, the efficiency of the OPV cells stays half less high than that of those which use CuI as buffer layer.
3.2 Influence of the different ABL configurations on the TTB properties
In order to try to understand the influence of CuI and MoO3 on the performances of the OPV cells, we have submitted the structures ITO/CuI/TTB and ITO/MoO3/ TTB to some specific measurements such as absor- bance, X-ray diffraction, X-ray photoelectron spectro- scopy (XPS), scanning electron microscopy and AFM.
The absorption of the TTB films (35 nm) deposited onto ITO covered with an ABL layer is reported in Figure 3. One can see that when the TTB layer is de- posited onto a CuI layer, the optical density of films under-
goes a bathochromic shift by comparison with the films de- posited onto MoO3. More precisely, when deposited onto MoO3 the maximum of the absorption band is situated at 505 nm, while it is situated at 523 nm when deposited onto CuI. The second band situated at around 390 nm is not visible here due to the substrate absorption.
It can be said from the X-ray diffraction study (not shown) that the TTB films deposited by sublimation un- der vacuum are nearly amorphous, a small peak is visible only when they are deposited onto MoO3.
In the case of MoO3 ABL, the small peak is situated at 2θ = 4.9◦, which corresponds to adspacing of 19.6 ˚A and it has been indexed with a (1 0 0) plane of TTB, indicating that there is some thin film growth along the a direction [2].
The XPS survey spectrum (not shown here) allows see- ing that all the expected elements, C, S, N and O, are present at the surface of the TTB film. Due to surface contamination there is some excess of oxygen, however we did not proceed to an etching of the surface due to the fact that it induces organic bonds’ destruction with preferential etching of atoms other than carbon. In Figure4, XPS spectra of the surface of a TTB thin film are presented. We have proceeded to a decomposition of the measured peaks into different components. The curve fit- ting programs, which include background removal, peak integration and area measurement, quantification, curve synthesis and peak fitting, permit the variation of parame- ters such as the Gaussian/Lorentzian ratio, the full width at half maximum (FWHM), and the position and the in- tensity of the contributions. These parameters were op- timized by the curve fitting programs to obtain the best fit.
It is well known that the binding energy of the peaks increases with the resistivity of the analyzed sample, which is called ΔE the charge effect. With the spectrometer used, the carbon-carbon binding energy reference is 284.5 eV, while the value deduced from Figure 4a is 284.9 eV which shows that ΔE = 0.4 eV. In order to compare the binding energies of the different components to values given in the handbooks, ΔE should be system- atically deduced from the values estimated from Figure4.
One can see in Figure 4a that the carbon peak C1s can be decomposed into four components. As said above, the peak situated at 284.5 eV corresponds to C-C bonds. The peak situated at 286 eV can be attributed to C-N. The third peak (287 eV) can correspond to C-O-C and also to some C-OH surface contamination. Then the fourth peak situated at 288 corresponds to C=O. For XPS measure- ments, the samples are transferred in the air from the deposition chamber to the XPS apparatus, which justi- fies the superficial contamination. After charge effect sub- traction, the binding energies of S2p and N1s correspond to the covalent bonds of the TTB. The O1s peak corre- sponds to three components. The contribution situated at 533.1 eV corresponds to C-O-C, but also to the C-OH sur- face contamination. The peak situated at 532 eV can be assigned the C=O bonds of TTB, while the last peak, at 534.6 eV, can be due to some H2O surface contamination.
(a)
(c) (d)
(b)
Fig. 4.(Color online) C1s (a), S2p (b), N1s (c) and O1s (d) spectra of TTB thin films deposited by sublimation.
Similar results were obtained whatever the ABL used. It can be concluded that, after sublimation in vacuum the TTB molecule seems preserved.
The cross-sections of the TTB film with a thickness of 40 nm deposited onto the different ABLs are visualized by SEM in Figure 5. The different films are clearly vis- ible in the cross-sections. The ABLs are continuous and homogeneous. The surfaces of the films have also been visualized. Figure6shows the example of TTB deposited onto ITO covered with MoO3. Similar results are obtained when the ABL is CuI. For large enhancements, grains are clearly visible in these images of the TTB films (Fig.6).
For smaller enhancement, some features randomly distrib- uted are visible. In order to check if they correspond to the polymer or some other compound, we have visualized the same surface in the backscattering mode. In that mode, the heavier atoms appear brighter on the photographs.
If there is no large gray contrast when a small acceler- ating voltage is used, the features appear brighter than the film itself, when a higher voltage is used. That shows not only, the atoms constituting these features are heav-
ier than those of the rest of the film, but also that they are not situated on the surface of the layer, but in depth, at the interface organic film ABL. However as shown by Figure6c, they induce some surface roughness.
Similar features are visible whatever the ABL is, MoO3 or CuI (Fig. 7). The SEM pictures of Figure 7have been exploited by using the open-source software ImageJ (http:/rsb.info.nih.gov/ij/). Such software allows improving the exploitation of the images themselves. SEM backscattered images were first filtered to reduce the noise then converted to binary image by threshold operation to see the area corresponding to the features (in black) and that of the substrate (in white). Then ImageJ can mea- sure this covering and the averaged particles diameter [8].
When the images of Figure 7 are treated with the soft- ware image J, it can be calculated that, the averaged size of the features is more or less the same, 44 nm with MoO3 and 42 nm with CuI. However, the surface coverage ratio is 2% with MoO3, while it is only 1.2% with CuI. In or- der to investigate the origin of these different features, we have proceeded to microprobe mapping of the visualized
Fig. 5. Cross-section visualization of ITO/CuI/TTB (a) and ITO/MoO3/TTB (b) structures.
surfaces. Unfortunately, Cu, Mo, C and O maps are homo- geneous without any features similar to those visualized by the SEM.
In order to check the surface morphology of the dif- ferent structures, they have been studied with an AFM apparatus.
Figure 8 shows the AFM image in three dimensions of 1 μm × 1 μm TTB surface area deposited onto ITO/MoO3(a) and ITO/CuI (b) structures. Small islands can be observed whatever the ABL is. In order to improve the precision of the averaged measured roughness, we have measured these values on larger surfaces. Figure9 shows the two dimensions images 10 μm × 10 μm. The root means square (rms) roughness is 2.23 nm with MoO3and 3.39 nm with CuI, while the peak to valley roughness is 23 nm with CuI and 17 with MoO3. The profiles shown below the 10μm ×10 μm images correspond to the col- ored drawn on the image. As regards the rate of cover- ing of the features, these profiles are in good agreement with those deduced by SEM. Actually, it can be seen in Figures 8 and 9 that the feature density is smaller with CuI. However, the rms is higher in the case of CuI ABL, this is due to the fact that the peak to valley roughness is significantly smaller with MoO3. Therefore the averaged value of the roughness is smaller with MoO3.
The (J-V) characteristics of the hole only devices are shown in Figure10. Figure10shows typical curves log(J) against log(V) plots obtained when injecting holes through the ITO contact. Electron injection from the aluminum electrode can be neglected, due to the large electron bar- rier caused by the MoO3interlayer. Thus charge transport in these devices is limited to holes. The devices using CuI ABL exhibit much higher current density than the ones using MoO3 ABL. Clearly the conductivity is enhanced by approximate one order of magnitude when a CuI ABL is used in the place of the MoO3 ABL. All the devices display near ohmic transport at low voltage. Actually, in the low voltage range, ohmic contacts can be seen by the
Fig. 6.Visualization of an ITO/MoO3/TBB in the secondary mode, (a) large and (b) small enhancement, and in the backscat- tering mode with different accelerating voltage, (c) 5 kV and (d) 10 kV.
Fig. 7. Image in the backscattering mode of ITO/CuI/
TBB (a) and ITO/MoO3/TBB (b).
linear relation between the current densityJand the volt- age V with a slope m = 1 in log-log plot (Fig. 10). At higher current density the slope starts increasing, which corresponds to a transition toward Space Charge Lim- ited Current (SCLC) often obtained in such devices [9].
However, the TTB film thickness is quite small and, the structure being in a simple sandwich configuration, it is submitted to very high electric field and, mainly in the case of CuI ABL, J becomes instable for higher voltage.
Therefore the (J-V) behavior at higher voltage, which cor- responds to the SCLC domain, could not be studied be- cause above 500 mV the devices undergo erratic variations (Fig. 10).
In order to try to understand the influence of the ABL on the growth of the TTB layers, we have measured the surface energy of the anodes: ITO/MoO3 and ITO/CuI.
Their surface energies are very different, when covered with MoO3 the anode surface is hydrophilic, while when it is covered with CuI it is hydrophobic. It means that the surface energy of ITO/MoO3 is far higher than that of ITO/CuI. More precisely it is 59.3 mJ/m2 with MoO3 ABL while it is only 26.3 mJ/m2 with CuI ABL. It has
fast
fast Height
Height (measure)
(a)
(b)
1μm
1μm μ
0 m 0 nm
0 nm 30 nm
30 nm
μ 0 m
Fig. 8. (Color online) AFM images in three dimensions of TTB films deposited onto ITO/MoO3 (a) and ITO/CuI (b) structures.
been already shown [10] that a fairly hydrophobic sur- face, such as that modified by CuI in the present work, could provide a better compatibility of the subsequent hy- drophobic organic material, such as TTB. It is also clearly established that the alignment of the molecules deposited onto the electrode depends on its surface energy, which can modify the carrier mobility and the trap density of the organic film and the contact resistance between the electrode and the organic material [11].
4 Discussion
The TTB thin films deposited by spin coating have been already characterized [2] and therefore the present study will be discussed in the light of the results already published.
When deposited by spin coating, the absorption spec- trum of TTB exhibits two absorption bands at 517 and 390 nm, the first peak being the strongest. The shorter wavelength band corresponds to aπ-π* transition and the longer wavelength corresponds to an intermolecular charge transfer (ICT) band of the D-A-D system. Figure3shows that the strongest peak of absorption of TTB is clearly vis- ible. The absorption spectra of the TTB deposited under vacuum onto CuI (Fig. 3) undergo a bathochromic shift relatively to those deposited onto MoO3, the absorption bands being at 523 and 505 nm respectively. Since, as dis- cussed by Siram [2], the planarization of the chromophore
Fig. 9.(Color online) Two dimension AFM images (10μm×10μm) and profiles of TTB films deposited onto ITO/MoO3 (a) and ITO/CuI (b) structures.
in the solid state induces a bathochromic effect of the ab- sorption bands, it can be said that this planarization is more effective when the films are deposited onto CuI.
We have shown that, when the ITO is covered with CuI, the surface energy of the anode is far smaller than when it is covered with MoO3. It is clear that such a large surface energy difference of the anode will modify the growth of the TTB layer. At least, it can be said that the order of the TTB films deposited onto CuI is different from that present in the films deposited onto MoO3. This dependence of the TTB film structure is corroborated by the XRD study.
When the ABL is MoO3, a small diffraction peak is visible at 2θ= 5◦, which corresponds to the most intense diffraction peak of the films deposited by spin coating. It corresponds to the d-spacing of 19.6 ˚A and it has been indexed with a (1 0 0) plane. Which means that the crys- talline phase of the film growth is preferentially along the a direction. With CuI as ABL, the films are amorphous.
These differences in the order present in the TTB films influence significantly the morphology and the conductiv- ity of the films. However, whatever the ABL used, the XPS analysis shows that the organic molecule is preserved after sublimation.
About the morphology (Fig.5–9) the main differences are that the density of bumps present at the surface of the films is higher in the case of MoO3 ABL, while the mean roughness and the peak to valley roughness are higher with CuI ABL. About the conductivity, we have seen (Fig.10) that it is enhanced by one order of magni- tude when a CuI ABL is used in the place of MoO3ABL.
Although injection on the ITO/ABL/TTB side may still limit the current to a certain extent, we assume ohmic
1E-3 0,01 0,1
10 100
1 10 100
10 100 1
10
100 ITO/CuI/TTB m = 1.01
J (mA/cm2 )
V (mV)
ITO/MoO3/TTB/MoO3/Al ITO/CuI/TTB/MoO3/Al
Fig. 10. (Color online) Log-log plot of the current density-voltage curves of the hole only devices, open symbols ITO/CuI/TBB/MoO3/Al and full symbols ITO/
MoO3/TBB/MoO3/Al, inset slope of the curve logJ-logV for the ITO/CuI/TBB/MoO3/Al hole only device.
charge injection since the slope of the (J-V) curves in log- log plot is nearly equal to unity in the low field regime.
Moreover, the HOMO and LUMO values of TTB are 5.91 eV and 3.91 eV respectively [2] and we have shown that for such high HOMO value MoO3 is a very efficient ABL due to its high work function [7]. As a matter of fact, it has been shown that the value of the surface work func- tion of MoOx is 5.9 eV, which corresponds to the molyb- denum oxide deposition by thermal evaporation [12]. The work function of CuI being 5.4 eV [13], which is far higher than the work function of ITO (4.7 eV), it should provide an ohmic contact with high HOMO value
material (band scheme, inset Fig. 2). Therefore, the in- crease of the conductivity of the TTB films deduced from the hole-only technique should be mainly attributed to the different
orders of molecules induced by the ABLs.
The performances of the different OPV cells can be interpreted in the light of the above discussion. First of all, as expected, the small efficiency of the cells without ABL is mainly due to the bad band matching between the ITO anode and the TTB [7]. In the case of the differ- ent ABLs, while both induce good band matching, which allows achieving reasonable efficiency, the higher conduc- tivity of the films deposited onto CuI induces a high short circuit current and therefore a higher OPV cell efficiency.
Moreover, we have seen that the rms of the films de- posited onto CuI is higher than that of the TTB films deposited onto MoO3, which can contribute to the en- hancement of Jsc through the increase of the interface area of charge separation [14–16]. However, it is known that under electrode excrescences can reduce the parallel resistance, increasing the leakage current of the current- voltage (I-V) diode characteristics, which results in lim- ited short circuit current and open circuit voltage.
Therefore, since in return for an improvement of the conductivity, the CuI infers an increase of the roughness and thus leakage currents, while the MoO3 is known to decrease these [4,6], it is expected that the best results are achieved with the double ABL MoO3/CuI (Fig. 2).
Actually there is a strong enhancement of Jsc and Voc.
This is due to the fact that such DABL allows summing the advantages of both ABL, MoO3 allows a very good band matching and avoid too high leakage current, while CuI allows achieving high Jsc thanks to its effect on the TTB conductivity.
Finally, we compare the performances of the present OPV cells based on multilayer heterojunction deposited under vacuum with those achieved by cells based on TTB:
PCBM BHJ deposited by spin coating. It can be seen that the performances of the cells using CuI ABL are of the same order of magnitude as those of the cells using PEDOT:PSS ABL, with a better result for cells deposited by spin coating. When the double ABL, MoO3/CuI, is used, their efficiency is significantly better than those of all the other configurations. We assume that the reason for this may be due to the dual function of MoO3and CuI since both of them were reported to be able to reduce the hole injection barrier compared with bare ITO, while CuI improves the TTB films conductivity and MoO3 prevents the OPV cells from leakage path formation. It should be noted that the improvement of the TTB and cell perfor- mances can also be due, at least partly, to the fact that, after sublimation the purity of the TTB films should be very high.
5 Conclusion
The work function of the ITO anode is adjusted to the HOMO level of TTB thanks to the ABL layer, which in- duces ohmic behavior, at least for small voltage. Moreover,
besides a good band matching between the anode and the ED, there are other factors which affect the contact ITO/ED after inserting the double ABL: MoO3/CuI. One is the enhancement of the TTB film conductivity induced by the CuI film and another is the prevention of leakage current due to MoO3. Thus Voc and Jsc are simultane- ously improved leading to optimum OPV cells efficiency.
Therefore, the interfacial layer located between the active layer and the anode in OPV cells is important in terms of hole extraction as well as control of morphology of the photoactive layer.
The authors thank S. Patil and his colleagues of the Solid State and Structural Chemistry Unit of Indian Institute of Science Bengalore 560012 India to have supplied the TTB gracefully.
This work has been financially supported by the France-Maroc contract: PHC Volubilis No. MA/10/228 and the Hassan II Academy of Science and Technology (Morocco).
References
1. J.C. Bernede, J. Chil. Chem. Soc.53, 1549 (2008) 2. B.K. Siram, K. Tandy, M. Horecha, P. Formanek,
M. Stamm, S. Gevorgyan, F.C. Krebs, A. Kiriy, P.
Meredith, P.L. Burn, E.B. Namdas, S. Patil, J. Phys.
Chem. C115, 14369 (2011)
3. A.F. Painelli, F. Terenziani, J. Phys. Chem. A104, 11041 (2000)
4. L. Cattin, F. Dahou, Y. Lare, M. Morsli, R. Tricot, K.
Jondo, A. Khelil, K. Napo, J.C. Bern`ede, J. Appl. Phys.
105, 034507 (2009)
5. C.H. Cheng, J. Wang, G.T. Du, S.H. Shi, Z.J. Du, Z.Q.
Fan, J.M. Bian, M.S. Wang, Appl. Phys. Lett.97, 083305 (2010)
6. J. Garnier, A. Bouteville, J. Hamilton, M.E. Pemble, I.M.
Povey, Thin Solid Films518, 1129 (2009)
7. J.C. Bern`ede, L. Cattin, S. Ouro Djobo, M. Morsli, S.R.B.
Kanth, S. Patil, P. Leriche, J. Roncali, A. Godoy, F.R.
Diaz, M.A. Del Valle, Physica Status Solidi A208, 1989 (2011)
8. F. Papadopulos, M. Spinelli, S. Valente, L. Foroni, C.
Orrico, F. Alviano, G. Pasquinelli, Ultrastruct. Pathol.
31, 401 (2007)
9. D. Chirvase, Z. Chiguvare, M. Knipper, J. Parisi, V.
Diakonov, J.C. Hummelen, J. Appl. Phys.93, 3376 (2003) 10. A. Sharma, P.J. Hotchkiss, S.R. Marder, B. Kippelen,
J. Appl. Phys.105, 084507 (2009)
11. D.-J. Yun, D.-K. Lee, H.-K. Jeon, S.-W. Rhee, Org.
Electr.8, 690 (2007)
12. M.T. Grenier, M.G. Helander, Z.B. Wang, W.M. Tang, J. Qiu, Z.H. Lu, Appl. Phys. Lett.96, 213302 (2010) 13. J.-H. Lee, D.-S. Leem, H.-J. Kim, J.-J. Kim, Appl. Phys.
Lett.94, 123306 (2009)
14. A. Fujii, H. Mizukami, Y. Hashimoto, T. Umeda, Y.
Nishihara, M. Ozaki, K. Yoshino, Synth. Met.152, 121 (2005)
15. Z. Liu, Q. Liu, Y. Huang, Y. Ma, S. Yin, X. Zhang, W.
Sun, Y. Chen, Adv. Mater.20, 3924 (2008)
16. S.W. Tong, C.F. Jiang, Q.D. Ling, E.T. Kang, D.S.H.
Chan, C. Zhu, Appl. Phys. Lett.93, 043304 (2008)
