Efficient semitransparent organic photovoltaic (OPV) cells are presented in an inverted geometry employing ZnS/ Ag/ WO3 (ZAW) as a top anode and ITO/ Cs2CO3 as a bottom cathode. Upon identification of the light absorption that differs depending on the illumination direction, the degree of the absorption asymmetry is tuned by varying the ZAW structure to maximize the efficiency for one direction or to balance it for both directions. Power conversion efficiency close to that of conventional opaque OPV cells is demonstrated in semitransparent cells for the ITO side illumination by taking advantage of the internal reflection occurring at the organic/ZAW interface. Cells with efficiencies that are reduced but balanced for both illumination directions are also demonstrated.
© 2010 OSA
Organic photovoltaic (OPV) cells based on polymer:fullerene blends have been receiving growing attention as a potential solution as a low-cost renewable energy source. With continued efforts toward development of photoactive polymers and control over the nanoscale morphology of polymer:fullerene mixtures, power conversion efficiencies higher than 7% have recently been demonstrated . Such achievement as well as recent demonstrations of roll-to-roll based large-area OPV fabrication [2,3] is regarded as a key step in realizing commercially viable OPV technologies. Along with the improved efficiencies and development of scalable low-cost processing technologies, useful characteristics that can differentiate OPV technologies from other PV technologies also have to be further developed. Semi-transparency, which can be utilized in applications such as energy-harvesting windows, is one of the characteristics for which OPV technologies are better positioned than existing PV technologies. A relatively narrow absorption band and thin-film fabrication of organic materials allows semitransparent OPV (ST-OPV) cells to be made simply and seamlessly [4–10], as opposed to devices made of inorganic counterparts in which the numerous small cells need to be arranged with a given aperture ratio to define the average transmittance [11,12]. However, the power conversion efficiencies of most ST-OPV cells reported to date have been significantly lower than that of conventional cells with only a few exceptions . Although use of a relatively thick photoactive layer may provide a solution in some cases, as in the work by Huang et al. , such method is not applicable unless active layers have a sufficient level of transport properties. In this work, we demonstrate a simple but highly effective strategy to enhance the power conversion efficiency of ST-OPV cells by utilizing the partial internal reflection from a multilayer top anode consisting of ZnS, Ag, and WO3.
2.1 Overview of the device structure
The semitransparent OPV cells under study are based on inverted geometry for improved air stability [13,14], and their overall device structure is presented in Fig. 1(a) . The bottom ITO electrode was coated with a very thin Cs2CO3 layer (1 nm) as a buffer that makes the ITO/ Cs2CO3 system work effectively as a cathode . As a top transparent electrode, a multilayer system of ZnS (x nm)/ Ag (15 nm)/ WO3 (13 nm) (ZAW) was employed. The ZAW electrode belongs to the category of a multilayer, metal-based transparent electrode referred to as a dielectric-metal-dielectric (DMD) electrode [16,17] or oxide-metal-oxide (OMO) electrode , which is being studied as an alternative to ITO electrodes. The ZAW electrode has been demonstrated to work effectively as a bottom transparent anode in ITO-free OPV cells  and organic light-emitting diodes (OLEDs) , as well as a damage-free top transparent anode in inverted OLEDs . In the ZAW electrode, WO3 layer, positioned next to organic layer, is used as a buffer layer making Ag/ WO3 layers work effectively as an anode. In contrast, ZnS layer virtually has no electrical role and are used primarily for tuning the optical properties of the electrode as well as the overall device. WO3 layer can also have a similar optical role if the ZAW electrode is to be used as a stand-alone sheet conductor, but its optical role in a typical OLED or OPV device configuration is related more to the cavity effect or optical spacer effect than to the electrode transmittance due to the relatively small index contrast with adjacent organic layers [16,18].
2.2 Fabrication and characterization
A thin film of Cs2CO3 (Alfa Aesar, 99.994%) was thermally evaporated in vacuum (2 × 10−6 Torr) onto precleaned ITO-coated glass substrates. The substrates were treated with air plasma using a plasma cleaner (PDC-32G, Harrick Plasma) before being taken into the evaporation chamber (HS-1100, Digital Optics & Vacuum). After deposition of Cs2CO3, the samples were transferred to a nitrogen-filled glove box without exposure to ambient air. For the active layer, a mixture of poly(3-hexylthiophene)(P3HT) and [6,6]-phenyl C71 butyric acid methyl ester (PCBM70) dissolved in dichlorobenzene (20mg/ml, 1:0.7 by weight) was spun at 700 rpm for 60 s. The samples were then dried for 30 min at room temperature and subsequently annealed at 110 °C on a hotplate for 10 min. Finally, the samples were reloaded into the chamber for a successive deposition of WO3 (Alfa Aesar 99.99%), Ag (Alfa Aesar, 99.999%), and ZnS (Alfa Aesar, 99.99%). The active area of the fabricated devices was typically in the range of 0.07-0.13 cm2.
Current density-voltage (J-V) characteristics were recorded using a source-measure unit (Keithley 2400) in a 4-wire sensing mode. Simulated AM 1.5G illumination was done using a solar simulator (ABET technologies) with AM1.5G filters. Irradiance was measured each time using a Si photodiode the response of which was calibrated against the reference Si solar cell. External quantum efficiency (EQE) spectra were also measured using a monochromator coupled to a Xe arc lamp and a calibrated Si photodiode. During the photovoltaic testing, samples were kept in an N2-environment without exposure to an ambient atmosphere. Transmittance measurement was done in ambient air using a UV-VIS spectrometer (SV2100, K-MAC).
Optical analysis was done either with commercially available thin-film optic software (Essential MacleodTM) or with the custom MATLABTM code. Thick substrate effect was also taken into account in all the calculation done in this work . Optical constants of participating layers were taken either from the literature or from the software. When not available, they were measured using the spectroscopic ellipsometry[Woollam, M2000D (RCT)].
3. The optical properties of the proposed ST-OPVs and their optimization strategies
Organic photovoltaic cells in general may be considered as an assembly of thin films, and their optical properties are known to be well described within the framework of thin-film optics using the transfer-matrix or characteristic-matrix formalism  (Also refer to the work by Koeppe et al.  and Ameri et al.  for similar approach previously applied in ST-OPVs). It is emphasized that the analysis of DMD-based OPVs should be done on the whole device structure rather than on the electrodes themselves for a correct and quantitative account of their optical properties, as was discussed in Refs. 16 and 18. It is also noted that validity in optical analysis depends largely on the precise determination of the optical constants of participating layers. The optical constants of P3HT:PCBM70 films obtained by the ellipsometric measurement are presented in Fig. 2(a) . It can be easily seen in Fig. 2(b) that the transmittance calculated with the measured optical constants shows a reasonably good agreement with the experimental values.
Note that ST-OPV cells may be configured in a symmetric geometry [6,7] with the same set of materials used for both the cathode and anode, or can be configured in asymmetric geometry [4,5] with the different set of materials as presented in this work. In thin-film optics, it is well known that the reflectance (R t) from an assembly of asymmetric thin films can depend on the direction of incidence [substrate side or top (film) side] in the presence of absorptive layers while transmittance (T t) does not . Since the absorption (A t) is related to R t and T t by A t = 1 − (R t + T t) due to energy conservation, this ‘asymmetry in R t and symmetry in T t’ leads to an asymmetry also in A t. Figure 3(a) and (b) shows the spectra of A t calculated in both ITO-side and ZAW-side illumination directions for ST-OPVs with ZnS layer of 20 nm and 50 nm and does confirm the asymmetric absorption that depends on the illumination direction in these cells.
In addition, A t of the proposed ST-OPV device can be varied by changing the thickness of ZnS layers (d ZnS) as presented in Fig. 3(c) showing A t calculated at the wavelength of 500 nm as a function of d ZnS. This is mainly because the net transmittance and reflectance (both external and internal) of the ZAW electrode is influenced sensitively by d ZnS . Similar trend has been reported by Tao et al. in ST-OPV cells with MoO3/Ag/MoO3 top anodes . It is noteworthy that A t for ITO-side illumination [≡ A t (ITO)] and that for ZAW-side illumination [≡ A t (ZAW)] vary differently as d ZnS changes. As can be seen in Fig. 3(c), A t (ITO) has its minimum near d ZnS of 20 nm and peaks near d ZnS of 50 nm - 80 nm at λ of 500nm (≡λ 0) while A t (ZAW) exhibits almost opposite trend. Hence, not only the overall values of A t but also the degree of absorption asymmetry or dependence of A t on the illumination direction can be tuned by d ZnS. It is also noted that A t (ITO) (λ 0) is always larger than A t (ZAW) (λ 0) regardless of d ZnS. This is consistent with the report made by Pandey and Samuel , and is regarded to come from the fact that the light enters into the active region with little reflection when the light is incident on the ITO side. The effective optical path length can also be larger in this case due to the internal reflection at the active (buffer)/ metal interface.
Figure 4 presents the calculated photocurrent density (J ph) of the present devices as a function of d ZnS. The calculation was done for each illumination direction using the following relationship of J ph to absorption in active layers (A active):
From the results shown above, one can now set the two distinctive optimization strategies according to the needs of a target application: (i) in applications where the orientation of an ST-OPV cell is fixed with respect to the position of a dominant illumination source, as in solar windows in building-integrated PV (BIPV) applications, d ZnS can be adjusted so that the asymmetry in A t is strengthened and that J ph is maximized for the illumination direction allowing for a higher A t ( = ITO-side in this work), at the expense of a reduced J ph in the opposite direction (≡ type-I optimization scheme). One can then let ST-OPV cells face the source in the illumination direction that allows for a higher A t and J ph; (ii) in applications where the orientation of an ST-OPV cell is not fixed with respect to a light source (≡ type-II application), d ZnS is chosen so that the asymmetry in A t is minimized and that J ph is balanced for both illumination directions (≡ type-II optimization scheme).
4. Results and discussions
Figure 5 presents the experimental current density (J) – voltage (V) characteristics of the ST-OPV cells with a d ZnS of 50 nm and 20 nm. Recall that they can be regarded as ST-OPV cells optimized under type I and type II schemes, respectively. (See Fig. 4). J-V of the best device among 13 different devices per each case is provided. The average PV parameters of the ST-OPV cells under study are also given in Table 1 .
For the cell with d ZnS of 50 nm, the power conversion efficiency (η) for the ITO-side illumination was 3.7% (3.1% considering the mismatch factor with respect to the true AM1.5G; Refer to the following paragraph for further discussion), which corresponds to > 80-85% of that of our typical inverted cells with the opaque metal electrode. Note that this level of efficiency is among the highest that has been reported to date in ST-OPV cells based on P3HT:PCBM70 layers . On the other hand, η of the same cell measured for the ZAW-side illumination was only 1.4% in average, showing a severe asymmetry in J ph. For d ZnS of 20 nm, η was 2.8% and 1.8% for ITO-side and ZAW-side illumination, respectively, showing relatively good balance. Both of the results show a good agreement with the prediction made in Fig. 4, confirming the effectiveness of the proposed optimization strategies.
In order to check the spectrally resolved validity of the proposed analysis and to estimate J sc and η under the true AM1.5G illumination (≡ J sc(AM1.5G) and η (AM1.5G)), an additional batch of samples were prepared to measure their EQE spectra (Fig. 6 ). Comparison of J sc and η measured under illumination from the solar simulator with respect to those estimated from EQE data for the true AM1.5G spectrum indicates that the mismatch factors m of ST-OPVs and opaque reference cells for ITO side illumination are in the range of 0.74-0.85 while those of ST-OPVs for ZAW-side illumination are approximately 0.97. Recall that mismatch factor can differ and is dependent on the spectral response of OPV cells as well as the output spectrum of the solar simulator . It appears that the spectral output of the simulator used has a tendency to overestimate the response for ITO-side illumination in comparison to that for ZAW-side illumination. Mismatch factors as well as J sc(AM1.5G) and η (AM1.5G) expected for the respective cases are also summarized in Table 1.
It is noted that the measured EQE spectra are fitted reasonably well by the simulated spectra with η QE set as 65% and 80% for ITO-side and ZAW-side illumination, respectively, although a slightly large difference between the experimental and simulated data can be seen in devices with 20 nm-thick ZnS layers. Morphology of a ZnS layer may differ from that of ideal uniform thin films in case of 20 nm-thick layers. Nevertheless, highly asymmetric or balanced feature for each case shows a good agreement with the results obtained from the J-V characteristics under the broadband illumination. At this stage, it is not clear why η QE for ZAW-side illumination is higher than that for ITO-side illumination. Since η QE is related to the carrier collection efficiency, it may indicate that carriers are collected with a higher probability when excitons are generated near the anode side. This would be probable when the carrier transport is limited by the hole-transporting layers . Further study will be needed for clarification.
Figure 7(a) presents the internal reflectance R ZAW (int) at the organic/ ZAW interface vs. d ZnS calculated for the light that is incident from the medium of P3HT:PCBM70 to the ZAW-electrode at the wavelengths of 450 nm, 500 nm, and 550 nm, which represent the major absorption band of P3HT:PCBM70 films. One can note that R ZAW (int) becomes largest near a d ZnS of 50-80 nm and becomes lowest near a d ZnS of 20 nm at all those wavelengths in the ST-OPV cells under study, indicating that R ZAW (int) is one of the key factors in controlling the J ph in the present cells. Figure 7(b) shows the distribution of the squared magnitude of the optical electric field [≡ |E(z)|2] inside the present cells calculated at the wavelength (λ) of 500 nm along the symmetry axis (≡ z-axis) for both ITO-side and ZAW-side illumination. Numerical integration of |E(z)|2 over the P3HT:PCBM70 layer is directly proportional to the absorption within the active layer and thus to the available photocurrent at a given wavelength . In case of ZAW-side illumination, |E(z)|2 drops almost exponentially as in Beer-Lambert’s law, whether d ZnS is 20 nm or 50 nm, due to the lack of internal reflection from ITO electrodes. In case of ITO-side illumination, however, |E(z)|2 within the active layer varies differently depending on d ZnS: for d ZnS of 20 nm, a profile similar to the ZAW-side illumination is observed because R ZAW (int) is small; for d ZnS of 50 nm, |E(z)|2 is enhanced in the middle of the active layer, improving the absorption of photons by the active layer. This improvement is consistent with the enhanced R ZAW (int) at d ZnS near 50 nm because the reflected field can add up to yield the net improvement in |E(z)|2.
It would be of concern in some applications if the proposed optimization scheme involves a significant drop in T t of ST-OPV cells. As can be seen in Fig. 8 , T t at a specific wavelength indeed shows some variation when d ZnS changes from 20 nm to 50 nm, and therefore, the apparent color looks different [see the inset of Fig. 8 for photographs of the actual samples]; however, T t averaged for the visible spectral range of 400-700 nm (≡ T avg) of the devices with a d ZnS of 20 nm and 50nm were 28.3% and 27.3%, respectively, showing only a small difference. It is also noted that these values are comparable to the typical average transmittance of tinting films used in applications such as automobiles .
In this study, we have presented semitransparent organic photovoltaic (ST-OPV) cells in inverted geometry in which a top anode is based on a multilayer transparent electrode consisting of ZnS, Ag, and WO3 (ZAW) layers. Upon identification of the asymmetric absorption that is characterized as a higher absorption in the ITO-side illumination, the ST-OPV devices were optimized by changing the thickness of the ZnS layer (d ZnS), for two distinctive objectives: (i) to let the asymmetry be strengthened so that the efficiency for the ITO-side illumination can be further enhanced to its maximum (≡ type-I optimization scheme); or, (ii) to let the asymmetry be lessened so that the efficiencies are balanced for both illumination directions (≡ type-II optimization scheme). The former method led to a P3HT:PCBM70-based ST-OPV cell with efficiency as large as 3.7% (3.1% estimated for the true AM1.5G, 1Sun condition) for the ITO-side illumination, which corresponds to ~80-85% of that of conventional opaque cells. Since the orientation of ST-OPV cells with respect to the major illumination source is fixed in many applications (e.g. building-integrated solar windows), the type I optimization method opens up the opportunity to obtain the most power out of a given ST-OPV cell with its ITO-side facing the major light source.
This work was supported in part by the Korea Energy Management Corporation (KEMCO) under the New and Renewable Energy R&D Grant (2008-N-PV08-02), by Korea Institute of Energy Technology Evaluation and Planning (KETEP) in Ministry of Knowledge Economy (MKE) under the New and Renewable Energy R&D Grant (N02090016), by EEWS program of KAIST, and by Korea Iron Steel corporation (KISCO). Authors are also grateful to W. S. Soun for spectroscopic ellipsometry measurement and to Dr. W. S. Shin and J.-U. Park for help in EQE measurement.
References and links
1. H.-Y. Chen, J. Hou, S. Zhang, Y. Liang, G. Yang, Y. Yang, L. Yu, Y. Wu, and G. Li, “Polymer solar cells with enhanced open-circuit voltage and efficiency,” Nat. Photonics 3(11), 649–653 (2009). [CrossRef]
2. A. J. Medford, M. R. Lilliedal, M. Jørgensen, D. Aarø, H. Pakalski, J. Fyenbo, and F. C. Krebs, “Grid-connected polymer solar panels: initial considerations of cost, lifetime, and practicality,” Opt. Express 18(S3), A272–A292 (2010). [CrossRef] [PubMed]
3. F. C. Krebs, T. D. Nielsen, J. Fyenbo, M. Wadstrøm, and M. S. Pedersen, ““Manufacture, integration and demonstration of polymer solar cells in a lamp for the “Lighting Africa” initiative,” Energy Environ. Sci. 3(5), 512–525 (2010). [CrossRef]
4. R. Koeppe, D. Hoeglinger, P. A. Troshin, R. N. Lyubovskaya, V. F. Razumov, and N. S. Sariciftci, “Organic solar cells with semitransparent metal back contacts for power window applications,” ChemSusChem 2(4), 309–313 (2009). [CrossRef] [PubMed]
5. C. Tao, G. Xie, C. Liu, X. Zhang, W. Dong, F. Meng, X. Kong, L. Shen, S. Ruan, and W. Chen, “Semitransparent inverted polymer solar cells with MoO3/Ag/MoO3 as transparent electrode,” Appl. Phys. Lett. 95(5), 053303 (2009). [CrossRef]
6. H. Schmidt, H. Flugge, T. Winkler, T. Bulow, T. Riedl, and W. Kowalsky, “Efficient semitransparent inverted organic solar cells with indium tin oxide top electrode,” Appl. Phys. Lett. 94(24), 243302 (2009). [CrossRef]
7. J. Huang, G. Li, and Y. Yang, “A semi-transparent plastic solar cell fabricated by a lamination process,” Adv. Mater. 20(3), 415–419 (2008). [CrossRef]
8. F. C. Krebs, S. A. Gevorgyan, and J. Alstrup, “A roll-to-roll process to flexible polymer solar cells: model studies, manufacture and operational stability studies,” J. Mater. Chem. 19(30), 5442–5451 (2009). [CrossRef]
10. T. Ameri, G. Dennler, C. Waldauf, H. Azimi, A. Seemann, K. Forberich, J. Hauch, M. Scharber, K. Hingerl, and C. J. Brabec, “Fabrication, optical modeling, and color characterization of semitransparent bulk-heterojunction organic solar cells in an inverted structure,” Adv. Funct. Mater. 20(10), 1592–1598 (2010). [CrossRef]
11. P. Nath, and C. Vogeli, “Translucent photovoltaic sheet material and panels.” US patent, No.5,176,758 (1993).
12. A. Ricaud, J. Schmitt, and J.-M. Siefert, “Semi-transparent solar module panel.” US Patent, No. D 353,129 (1994).
13. S. K. Hau, H.-L. Yip, N. S. Baek, J. Zou, K. O’Malley, and A. K.-Y. Jen, “Air-stable inverted flexible polymer solar cells using zinc oxide nanoparticles as an electron selective layer,” Appl. Phys. Lett. 92(25), 253301 (2008). [CrossRef]
14. M. Jørgensen, K. Norrman, and F. C. Krebs, “Stability/degradation of polymer solar cells,” Sol. Energy Mater. Sol. Cells 92(7), 686–714 (2008). [CrossRef]
15. G. Li, C.-W. Chu, V. Shrotriya, J. Huang, and Y. Yang, “Efficient inverted polymer solar cells,” Appl. Phys. Lett. 88(25), 253503 (2006). [CrossRef]
16. S. Han, S. Lim, H. Kim, H. Cho, and S. Yoo, “Versatile Multilayer Transparent Electrodes for ITO-Free and Flexible Organic Solar Cells,” IEEE J. Sel. Top. Quant. Electron. (available online. doi: ).
17. H. Pang, Y. Yuan, Y. Zhou, J. Lian, L. Cao, J. Zhang, and X. Zhou, “ZnS/Ag/ZnS coating as transparent anode for organic light emitting diodes,” J. Lumin. 122–123, 587–589 (2007). [CrossRef]
19. C. Yun, H. Cho, H. Kang, Y. M. Lee, Y. Park, and S. Yoo, “Electron injection via pentacene thin films for efficient inverted organic light-emitting diodes,” Appl. Phys. Lett. 95(5), 053301 (2009). [CrossRef]
20. H. A. Macleod, Thin-Film Optics, 3rd Ed. pp.53–72 (Taylor & Francis, New York, 2001).
21. L. A. A. Pettersson, L. S. Roman, and O. Inganas, “Modeling photocurrent action spectra of photovoltaic devices based on organic thin films,” J. Appl. Phys. 86(1), 487–496 (1999). [CrossRef]
22. A. K. Pandey and I. D. W. Samuel, “Photophysics of solution-processed transparent solar cells under top and bottom illumination.” IEEE J. Sel. Top. Quant. Electron. (available online. doi: ).
23. S. E. Shaheen, C. J. Brabec, N. S. Sariciftci, F. Padinger, T. Fromherz, and J. C. Hummelen, “2.5% efficient organic plastic solar cells,” Appl. Phys. Lett. 78(6), 841–843 (2001). [CrossRef]
24. P. Schilinsky, C. Waldauf, J. Hauch, and C. J. Brabec, “Simulation of light intensity dependent current characteristics of polymer solar cells,” J. Appl. Phys. 95(5), 2816–2819 (2004). [CrossRef]
25. http://solutions.3m.com/wps/portal/3M/en_US/WF/3MWindowFilms/Products/TechnicalLibrary/ (accessed on July, 2010).