1. Final report. 1.1 Project details. 1.2 Executive summary

Index - Final Report 1. Final report 2 1.1 Project details 2 1.2 Executive summary 2 1.3 Project results 7 1.4 Utilization of project results 33 1.5 Project conclusion and perspective 34 1.6 Annual export of electricity (only ForskVE) 35 1.7 Updating Appendix P and submitting the final report 35 1

1. Final report The final report must be prepared in English. Please fill in the following sections of the template. 1.1 Project details Project title PolyStaR Project identification 2009-1-0143 Name of the programme which has funded the project ForskEL (ForskVE, ForskNG or ForskEL) Name and address of the enterprises/institution responsible for the project Technical University of Denmark, Anker Engelundsvej 1, 2800 Kgs. Lyngby DTU Energy Conversion, Frederiksborgvej 399, 400 Roskilde CVR (central business register) 30 06 09 46 Date for submission 17/12-2012 1.2 Executive summary PolyStaR aimed at demonstrating large-area processing of polymeric solar cells with improved stability and reliability. The specifications for the photovoltaic modules available by the end of this project were anticipated as a module efficiency of ~1%, a dark stability of 1 year and operational stability of 1000 hours, fabricated by large area roll-to-roll processing techniques. Full roll-to-roll production of 13.500 modules on 3x100 m foil was demonstrated with average initial efficiency of 2% and up to 2.8% efficiency and less than 20% degradation of performance under 1000 h illumination. The dark stability was less than a year, but it did not harm the use of the cells in the application and the efficiency did not degrade below 1%. The modules were used to demonstrate a prototype product a polymer solar cell powered flashlight - in collaboration with a Danish industry partner. Although the overall aim has been fulfilled this result is obtained through an effort extending beyond the PolyStar project. By far the most significant result of the PolyStar project itself is the knowledge and understanding of degradation of polymer solar cells and the international collaboration built in this field during PolyStar. Here the international consortium showed an added effect and a major part of the Danish effort has been directed here. Project progress 2

The last half year period of the project has been productive and successful. One publication have accepted for publication and a total of 4 prepared since March 2012. The PhD report by from Morten V. Madsen due on August 1 th, is currently on schedule. WP1 (management): The completion of the final report (Mi3.1) fulfils the WP1 milestones. WP2 (synthesis): Polymer materials with low band gap and thermo-cleavable polymers with enhanced photochemical stability was synthesized and new materials with crosslinkable sidechains were been prepared. Inks have been made and lab scale solar cells have been prepared. The morphological stability is improved. Thermocleavable polymers have shown significant increase in lifetime once the side chain had been cleaved. Fabrication of polymer solar cells using aqueous processing has in addition been demonstrated for all layers including the metal back electrode. Mi2.5 Formulated and characterized environmentally friendly ink system available (IMEC) (M36) FINAL REPORT WP3 (Application technologies): Processing of the active layer and of the back electrodes has been carried out with slot-die coating and screen printing. In both cases water has been used as an environmentally friendly solvent (the most environmentally friendly solvent). Mi3.4 Complete processing of polymer solar cells modules by R2R processing (M36) FINAL REPORT WP4 (degradation study): Mi4.5 Unification of process and stability in a R2R processed module offering 1% power conversion efficiency and >1000 hours of operational stability (M36) FINAL REPORT WP5 (Lifetime enhancement and encapsulation):lifetime tests on solar cells have been carried out for seven distinct sets of state of the art organic photovoltaic devices were prepared by leading laboritories. Mi5.7 Lifetime experiments on thin film encapsulated solar cells (IMEC, Cytec Surface Specialties) (M30) Mi5.8 Model predicting lifetime (IMEC) (M36) FINAL REPORT WP6 (Industrial Processing): The active layers, the electrodes and the encapsulation using barrier technology has been evaluated. The best approach has been identified and is being continued in activities towards milestones 6.4, 6.5 and 6.6. Mi6.4 Definition of production process (Riso-DTU, IMEC, Cytec Surface Specialties NV,SA) (M30) FINAL REPORT Mi6.5 Production process simulation (Riso-DTU, IMEC) (M33) Not fully met. All process steps were defined but not simulated - FINAL REPORT. Mi6.6 Calculation of investment costs for production start-up (Cytec Surface Specialties NV,SA) (M36) not met. The Flemish industry partner s results are not available. WP7 (Innovation and technology transfer): Final report 3

Milestone and timeplan: 9. Status of time schedule Activities/ milestones/payment Mi1.1 Mi1.2 Mi1.3 Year 2009 Year 2010 Year 2011 Year 2012 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 OK OK Mi2.1 Mi2.2 Mi2.3 Mi2.4 Mi2.5 OK OK OK, IMEC OK, IMEC&CYTEC Mi3.1 Mi3.2 Mi3.3 Mi3.4 Mi4.1 Mi4.2 Mi4.3 Mi4.4 Mi4.5 Publication 1 Publication 9 Publication 13 Publication 18 Publication 2,28,29 Publication 7,8,11,12 Publication 6,10,14,27,30 Publication 22 22NN Publication 22 18 Mi5.1 Mi5.2 Mi5.3 Mi5.4 Mi5.5 Mi5.6 Mi5.7 Mi5.8 OK OK OK OK, CYTEC OK, CYTEC OK, IMEC OK, Pub. 20, 25, 26 Mi6.1 Mi6.2 Mi6.3 Mi6.4 Mi6.5 Mi6.6 OK, CYTEC Publication 9, 14 OK, IMEC Mi7.1 Mi7.2 Mi7.3 Publication 3 and 4 Publication 5 4

Publication and dissemination (List the publications, articles, etc that have been published.) Publications 1-5 are from the 1 st year given briefly below: Publication 1: Nanoscale (2010), 2, 873-886. Publication 2: Sol. Energy Mater. Sol. Cells (2010), 94, 2018-2031. Publication 3: Energy & Environmental Science (2010), 3, 512-525. Publication 4: Sol. Energy Mater. Sol. Cells (2010), 94, 1553-1571. Publication 5: J. Mater. Chem (2010), 20, 8994-9001. Publications 6-16 are from second year given briefly below: Publication 6 J. Am. Chem. Soc. (2010), 132, 1688-16892. Publication 7 Polymer Degradation and Stability (2010), 95, 2666-2669 Publication 8 Sol. Energy Mater. Sol. Cells (2010), 95, 1308-1314 Publication 9 Sol. Energy Mater. Sol. Cells (2010), 95, 1348-1353 Publication 10 Sol. Energy Mater. Sol. Cells (2011), 95, 1268-1277. Publication 11 Sol. Energy Mater. Sol. Cells (2011), 95, 1389-1397. Publication 12 Sol. Energy Mater. Sol. Cells (2011), 95, 1398-1416. Publication 13 Adv. Energy Mater. (2011), 1, 68-71. Publication 14 J. Phys. Chem C (2011) 115, 10817-10822 Publication 15 35th PVSC. IEEE, (2010) 1068-1072 Publication 16 Adv. Funct. Mater. (2011) 21, 64-72. Publications 17-25 are from the last year until March given briefly below: Publication 17 Energy Environ. Sci. (2011), 4, 3741 Publication 18 Energy Environ. Sci. (2011), 4, 4116-4123 Publication 19 Synth. Met. (2011),doi:10.1016/j.synthmet.2011.10.008 Publication 20 RSC Advances (2011), online. DOI: 10.1039/c1ra00686j Publication 21 Sol. Energy Mater. Sol. Cells (2012), 97, 157-163. Publication 22 Materials Today (2012), 15, 36-49. Publication 23 Energy Environ. Sci. (2012), 5, 5117-5132. Publication 24 Advanced Materials (2012), 24, 580-612. Publication 25 Energy Environ. Sci. (2012). doi: 10.1039/C2EE03508A Publication 25 J. Mater. Chem. (2012). doi:10.1039/c2jm16340c Publication 26 Title: On the stability of a variety of organic photovoltaic devices by IPCE and in-situ IPCE analyses - The ISOS-3 inter-laboratory collaboration. Authors: Gerardo Teran-Escobar, David M. Tanenbaum, Eszter Voroshazi, Martin Hermenau, Kion Norrman, Matthew T. Lloyd, Yulia Galagan, Birger Zimmermann, Markus Hösel, Henrik F. Dam, Mikkel Jørgensen, Suren Gevorgyan, Suleyman Kudret, Wouter Maes, Laurence Lutsen, Dirk Vanderzandej, Uli Wurfel, Ronn Andriessen, Roland Rosch, Harald Hoppe, Agnes Rivaton, Gulsah Y. Uzunoğlu, David Germack, Birgitta Andreasen, Morten V. Madsen, Eva Bundgaard and Frederik C. Krebs, Monica Lira-Cantu Journal: Phys. Chem. Chem. Phys.(2012) doi: 10.1039/b000000x Publication 27 Title: TOF-SIMS investigation of degradation pathways occurring in a variety of organic photovoltaic devices the ISOS-3 inter-laboratory collaboration. Authors: Birgitta Andreasen, David M. Tanenbaum, Martin Hermenau, Eszter Voroshazi, Matthew T. Lloyd, Yulia Galagan, Birger Zimmernann, Suleyman Kudret, Wouter Maes, Laurence Lutsen, Dirk Vanderzande, Uli Würfel, Ronn Andriessen, Roland Rösch, Harald Hoppe, Gerardo Teran-Escobar, Monica Lira-Cantu, Agnès Rivaton, Gülşah Y. Uzunoğlu, David S. 5

Germack, Markus Hösel, Henrik F. Dam, Mikkel Jørgensen, Suren A. Gevorgyan, Morten V. Madsen, Eva Bundgaard, Frederik C. Krebs and Kion Norrman Journal: Phys. Chem. Chem. Phys.(2012) doi: 10.1039/C2CP41787A Publication 28 Title: Influence of processing and intrinsic polymer parameters on photochemical stability of polythiophene thin films. Authors: Morten V. Madsen, Thomas Tromholt, Arvid Böttiger, Jens W. Andreasen, Kion Norrman and Frederik C. Krebs Journal: Polymn Degrad Stabil (2012) doi: 10.1016/j.polymdegradstab.2012.07.021 Publication 29 Title: Concentrated light for accelerated photo degradation of polymer materials. Authors: Morten V. Madsen, Thomas Tromholt, Kion Norrman, and Frederik C. Krebs Journal: Adv. Energy Mater. (2012) doi: 10.1002/aenm.201200663 Publication 30 Title: Comparative studies of photochemical cross-linking methods for stabilizing the bulk hetero-junction morphology in polymer solar cells. Authors: Morten Jon Eggert Carlé, Birgitta Andreasen, Thomas Tromholt, Morten V. Madsen, Kion Norrman, Mikkel Jørgensen, and Frederik C. Krebs Journal: J. Mater. Chem. (2012) doi: 10.1039/C2JM34284G 6

1.3 Project results Problem statement: PolyStaR aims at demonstrating large-area processing of polymeric solar cells with improved stability and reliability. The project thereby addresses the problem of unifying power conversion efficiency, operational stability and low cost processing techniques for polymer and molecular photovoltaic cells. To realize this, the project proposes dedicated efforts to every important step in the entire fabrication chain from development of novel materials through device optimization with advanced testing methodologies to industrial processing with cutting edge technologies. This concerted effort is expected to result in: 1/ new low bandgap materials with high processing versatility through side-chain engineering, enabling environmentally friendly large area processing of flexible polymeric PV modules 2/ increased operational device lifetime by introduction of new crosslinkable compounds to stabilize the bulk heterojunction nano-morphology and by development of flexible thin-film device encapsulation 3/ definition of the requirements for reliable module fabrication under industrial conditions 4/ demonstration of the developed technology by prototyping products powered by our flexible PV modules The specifications for the photovoltaic modules that will be a direct result of this project are anticipated to have a module efficiency of ~1%, a dark stability of 1 year and operational stability of 1000 hours, fabricated by large area roll-to-roll processing techniques. During the project period those objectives were met: modules with up to 2.8% and in average >1% efficiencies were demonstrated in production of more than 10.000 solar cell modules. An operational stability of >1000 hours was demonstrated for those cells. A dark shell life of >4 years were also demonstrated in the period from 2008-2012 (unpublished result). After the project ending, modules with 4.8% efficiency were demonstrated. The PolyStar project was conducted within the framework of ERA-NET together with partners from Flanders, Belgium: IMEC, Cytec Surface Specialties NV,SA. The Danish project and the ERA-PV was funded and managed under Energinet.dk, FORSKEL programme, whereas the Flemish project entitled OPVLife was funded under IWT (Agentschap voor Innovatie door Wetenschap en Technologie). OPVLife was started on March 2009. The PolyStar project period was 1/4 2009-30/6, 2012. The international collaboration was successful in the area of polymer solar cell stability. This work conducted by both DTU and IMEC formed the basis for increasing operational life of polymer solar cells. The third International Summit on Organic Photovoltaic Stability (ISOS- 3) took place at Risø DTU in 2010. ISOS is a transnational network and the PolyStar collaboration was a corner stone for a large part of the work presented at ISOS-3. 1/Objective: new low bandgap materials with high processing versatility through side-chain engineering, enabling environmentally friendly large area processing of flexible polymeric PV modules 1/Results summary: This objective was covered by the work package 2 efforts headed by IMEC/IMOMEC. Functionalized conjugated polymers were synthesized by IMEC/IMOMEC and delivered to RISO DTU for determining the best blend composition [Polymer/PCBM] towards the higher efficiency. An optimum efficiency of 1.3% was obtained. 1/Results description: 7

300 mg of the following sample were dispatched to RISO DTU. Also GPC has been performed by Krebs group in order to compare it with IMEC results. Results, obtained from RISO DTU were comparable to the ones in our lab, except for SKC-002-85/15. Batch code SKT-001 P3HT rieke, Homemade Mw= 31137 Mn= 21519 PDI= 1.44 Batch code SKC-001 O O 85/15 Mn = 14000 Mw= 24400 PDI= 1.7 90/10 Mn= 15700 Mw=27700 PDI= 1.7 95/05 Mn=17200 Mw=30200 PDI= 1.7 X= 85, 90, 95 y= 15, 10, 05 S x S y Batch Code SKC-002 OH 85/15 Mn = 14100 Mw= 25500 PDI= 1.8 90/10 Mn= 16500 Mw=29700 PDI= 1.9 95/05 Mn=1700 Mw=31700 PDI= 1.8 X= 85, 90, 95 y= 15, 10, 05 S x S y Batch Code SKC-003 85/15 Mn = 14800 Mw= 25700 PDI= 1.7 90/10 Mn= 16100 Mw=27800 PDI= 1.7 95/05 Mn=17400 Mw=29200 PDI= 1.7 X= 85, 90, 95 y= 15, 10, 05 8

Table:best efficiency obtained for blends: Sample reference % PCBM SKC-001 (85/15) 48 SKC-001 (90/10) 50 SKC-001 (95/5) 42 SKC-002 (85/15) Coating problem SKC-002 (90/10) 42 SKC-002 (95/5) 45 SKC-003 (85/15) 40 SKC-003 (90/10) Coating problem SKC-003 (95/5) 60 9

2/ Objective: increased operational device lifetime (by introduction of new crosslinkable compounds to stabilize the bulk heterojunction nano-morphology and by development of flexible thin-film device encapsulation) 2/Results summary: Work from both IMEC and DTU is reported. The fulfillment of the overall project objectives are described in 3/. The work packages 4 and 5 were central to the project. The first set of results from DTU describes how detailed degradation mechanisms for polymer solar cells were identified. The second set of results from IMEC identifies proper strategies and setup for testing and interpreting data of stability for polymer solar cells. 2/Results description, DTU: Figure 1. TOF-SIMS mass spectra of a ZnO surface before and after 22 h of photo-annealing. The figure is reproduced with permission from The Journal of Solar Energy Materials and Solar Cells. 1 In the article entitled The effect of post-processing treatments on inflection points in current voltage curves of roll-to-roll processed polymer photovoltaics 1 (not included in the thesis). In this an investigation of an observed inflection point in roll-to-roll coated cells was studied. The inflection point was shown to be removed after continuous current-voltage sweeps during illumination (1000 W m 2 ) at 80 C for 15 30 minutes. In addition to classical IV testing under various conditions, devices were analyzed using X-ray photoelectron spectroscopy (XPS) and TOF-SIMS in order to ascertain a possible relationship between photoannealing and chemical changes in the devices. The chemical composition of the ZnO layer was observed to change significantly as a result of photo-annealing, see Fejl! Henvisningskilde ikke fundet.. A possible mechanism based on ZnO photo-conductivity, photooxidation and redistribution of oxygen inside the device as proposed by Verbakel was used to explain the observed inflection point. Re-distribution of oxygen within the cell was thus responsible for the reversible inflection point behavior. The oxygen was present as a result of photo-desorption from ZnO and/or decreased oxygen solubility in encapsulation layers (at elevated temperatures). It was concluded that devices employing ZnO will likely require some pre-treatments and/or chemical doping in order to optimize performance. In the article it was demonstrated that photo-annealing removes the remains of methoxyethoxy-acetate 10

used to make the ZnO nano-particles soluble, see the peak marked 1 in Fejl! Henvisningskilde ikke fundet.. The observation of characteristic fragment ions from the ionization process confirms the identity of the methoxyethoxy-acetate. The presence of an O 2 H peak that increases in intensity after photo-annealing was also observed. This is an indicator ion for the superoxide ion, known to form during photo-excitation of ZnO in the presence of oxygen, and contributes to the degradation of organics including P3HT. The article entitled Oxygen- and water-induced degradation of an inverted polymer solar cell: the barrier effect (Appendix 2.2) 2 describes the effect the solar cell itself plays as a barrier towards oxygen and water. In this work the difference between the stability of normal and inverted geometry devises was investigated. While cells of normal geometry have demonstrate high stability when subjected to oxygen and low stability when subjected to a water atmosphere the opposite is true for inverted device geometry cells, see Fejl! Henvisningskilde ikke fundet.. It was observed that both atmospheres lead to fast degradation of the initial response for the unencapsulated devises (black). The oxygen atmosphere led to complete degradation of the device in roughly 20 hours with all parameters showing fast decay. The comparable cell exposed to a humid atmosphere remained functional after the 480 hour time frame of the experiment. The encapsulated devises (grey) generally showed little degradation. Normalized PCE H 2 18 Oatmosphere 18 O 2 atmosphere Time (hours) Time (hours) Figure 2. Normalized PCE describing the degradation in performance of encapsulated (gray) and nonencapsulated (black) devices under continuous illumination (330 W m 2, AM1.5G, 65 ± 2 C). The mechanism for the diffusion of water into the normal geometry device is fairly well described by primarily diffusion of oxygen through pinholes in the metal electrode. 3 As a possible explanation of the difference in the behavior for normal and inverted cells is was hypothesized that the different layers act as a barrier toward both water and oxygen. If this barrier effect is different for oxygen and water the hypothesis can explain the observed behavior. The aim of the work presented was to determine the effect of each layer in the inverted geometry stack as a barrier material. A series of four part solar cells were prepared for both, see Figure 3. This allows the barrier effect of the layers from the active layer and up to be 11

tested. The experiment was based on the uptake of isotopically labeled oxygen ( 18 O 2 ) and water (H 18 2 O). The influence of the atmosphere was established by illumination of the samples at 330 Wm 2 at 65 C in a chamber with controlled atmosphere. Prior to the experiment a pressure of 10 4 mbar was established inside the chamber and the entire system was purged with nitrogen (99.9%) and pumped back down to 10 4 mbar. For the water atmosphere condition the chamber was then injected via a septum with H2 18 O (97%, 5 ml, 20 mmol). The entire system had a volume of 2.5 L resulting in a saturated isotopical labeled atmosphere. For the oxygen atmosphere the chamber was filled with 1 atm of 18 O 2 and N 2 in a ratio of 20 to 80. For both atmosphere conditions the samples were exposed for a period of 14 days. Encapsulation P3HT:PCBM Ag Ag Ag PEDOT:PSS P3HT:PCBM Ag Ag Ag PEDOT:PSS P3HT:PCBM ZnO ZnO ZnO ZnO ITO ITO ITO ITO A B C D Figure 3. Schematic illustration of partial (a) (c) and complete (d) solar cell devices. Information on where and to what extent oxygen uptake took place was investigated by analyzing the ZnO surfaces by TOF-SIMS. In order to obtain access to the ZnO surface delamination was used in the case of the encapsulated device. The delamination was shown to take place at the P3HT:PCBM / PEDOT:PSS interface. For the remaining sample layers the PEDOT:PSS layer was removed by gently swiping the surface with a cotton stick soaked in pure water. The underlying P3HT:PCBM layer was removed using the same procedure by substituting water with chloroform. Having exposed the entire ZnO interface for all part devices TOF-SIMS mass spectra analysis was carried out. Figure 5 shows the incorporation of 18 O at the ZnO surface in each of the given cases. In the oxygen atmosphere a clear barrier effect is seen for all layers (blue bars). It is seen that each layer has a distinctive effect as a barrier. In the humid atmosphere (red bars) it is seen that the active layer has a profound effect on the oxygen uptake. In fact the barrier effect of the active layer effectively shields the effect of the preceding layers as the difference between B and C lies within the error bars. The increase in incorporation of oxygen seen for the encapsulated devise (D) can seem puzzling. The explanation given in the article is that the binder used for the Alcan encapsulation is hygroscopic and acts as a reservoir for water. The observations that the uptake of oxygen is more pronounced in an dry oxygen atmosphere as compared to a humid atmosphere is in good correlation with the lifetime study demonstrating superior lifetime for cells in a humid atmosphere for inverted geometry devices. 12

Normalized 18 O intensity 100 75 50 H2O H 18 2 O atmosphere Atmosphere Dry 18 Ooxygen 2 Atmosphere atmosphere (dry) 25 0 A B C D Figure 4. Figure 5. Normalized 18O intensities for partial (a) (c) and complete (d) solar cells. The values was normalized to the largest degree of oxygen exchange seen in the oxygen-free humid atmosphere. (c) The functional cell without encapsulation, (d) the same cell with encapsulation, and (a) and (b) partial device. Figure 6. Contrast image (left) of the PEDOT phase marker (red) and the PSS markers (blue). The markers are PEDOT: S, 34 S, C 2 S, C 2 HS, C 6 H 5 O 2 S, C 6 H 6 O 2 S and PSS: SO 2, SO 3, C 8 H 7 SO 3. (right) 18 O - marker shown for the same data set. The sample was exposed to an 18 O 2 rich atmosphere. The article entitled Degradation patterns in water and oxygen of an inverted polymer solar cell 4 (not included in this thesis). In this study the spatial distribution of water and oxygen atmospheres induces reaction products in multilayer polymer solar cells was mapped. The geometry studied was an inverted geometry device with a layer sequence starting with an ITO coated glass slide. On top of this a zinc oxide layer followed by a P3HT:PCBM active layer and a PEDOT:PSS acting as a hole transporting layers. The top electrode was a printed silver electrode. By using labeled atmospheres detailed information on where and to what extent uptake took place was obtained. The labeling was employed using atmospheres with H2 18 O and 18 O2 respectively. X-ray photoelectron spectroscopy (XPS) and TOF-SIMS then enabled degradation patterns and failure mechanisms to be elucidated. It was concluded that the reactions taking place at the interface between the active layer and the PEDOT:PSS were the major cause of device failure. Phase separation in the PEDOT:PSS was observed, with the PEDOT-rich phase being responsible for most of the interface degradation in oxygen atmos- 13

pheres. TOF-SIMS images displaying the distribution of 18 O - demonstrated that oxygen preferentially reacted with the PEDOT phase, see Fejl! Henvisningskilde ikke fundet.. This observed phase separation affects the barrier properties of the layers as a result. It was observed that the reaction pattern of 18 O - was persistent through the sublayers suggesting that oxygen diffuses more efficiently through the PEDOT as compared to the PSS phase. In the water atmospheres, little chemically induced degradation was observed as seen in Fejl! Henvisningskilde ikke fundet. where no contrast in the 18 O - image can be seen in relation to the PEDOT:PSS contrast image. Figure 7. The left image represents a contrast image of the PEDOT phase marker (red) and the PSS markers (blue). The markers are PEDOT: S, 34 S, C 2 S, C 2 HS, C 6 H 5 O 2 S, C 6 H 6 O 2 S and PSS: SO 2, SO 3, C 8 H 7 SO 3. The right represents the same data set, but shows the 18 O - marker. The sample was exposed to a humid H2 18 O atmosphere. The automation of the trivial task of performing photo-degradation has been an important step. By removing the element of an operator from the process the number of errors in an experiment can be reduced significantly. Secondly time intervals of measurements can be reduced and lag time almost completely removed. Lastly since an automated system can work day and night a much larger number of samples can be evaluated largely increasing the statistics. By operating a sample exchanger robot equipped with a spectrometer setup for transmission measurements this was achieved. This work has been described in three articles entitled; Photochemical stability of conjugated polymers, electron acceptors and blends for polymer solar cells resolved in terms of film thickness and absorbance (Appendix 2.3) 5, Influence of processing and intrinsic polymer parameters on photochemical stability of polythiophene thin films (Appendix 2.4) 6, and Comparative Studies of Photo Chemical Crosslinking Methods for Stabilizing the Bulk Hetero-Junction Morphology in Polymer Solar Cells (not included in the thesis) 7. The work has solely focused on the degradation of the active layer materials in the solar cells. The first scientific contribution made using the automated photo degradation setup was reported in the article entitled: Photochemical stability of conjugated polymers, electron acceptors and blends for polymer solar cells resolved in terms of film thickness and absorb- 14

ance (Appendix 2.2) 5 When making comparative studies of polymer stabilities, many different parameters influence the experimental conditions, which may be outside the control of the experimenter. The majority of the above mentioned parameters are normally approximately constant if not actively changed. Parameters such as the temperature, light spectrum, and light intensity are typically kept constant. The focus of the article was to expand the knowledge of the parameter room comprising the simple system of a thin polymer film on a substrate. The main question before the work of this article was what the influence of thickness on the stability had. Further since it is common to use absorption loss rate to compare polymer stability 8, it is important to known if this is indeed yields a fair comparison between stabilities or if the thickness directly is a better basis of comparison. The effect of varying optical density / thickness on material stability has not been studied systematically before the article and therefore the uncertainty introduced by thickness variation was unknown. By comparing stabilities without knowing the influence of the thickness of the film can in the worst case result in wrong conclusions to be drawn. In the article photochemical stabilities of six different polymers were studied, see Figure 3.8. A clear initial absorbance (thickness) dependence was visible for all polymers. By plotting the relative stabilities of the polymers to the stability of regio-regular P3HT revealed that reasonably flat lines were obtained when plotted against the initial absorbance, see Figure 3.9 (left). This indicated that the absorbance provides a relatively fair basis for comparison. However, as is evident from the figure, intersections between different polymers are present. Hereby comparing polymers at low initial absorption can yield the opposite conclusion when comparing polymers at high initial absorbance. This is extremely important to comparative studies where the absorbance has to be kept constant for all materials being studied to provide a basis for valid conclusions on relative stabilities. Still, however, the validity of estimating a material stability based on a single measurement at a single absorbance is considered doubtful. Only by studying a wide thickness range for all studied samples, a sound estimation of relative stabilities can be obtained. Consequently it was concluded that relative stabilities cannot be given in factors less than five if only a single degradation of each material has been performed. 10-1 Degradation rate (%/s) 10-2 10-3 10-4 10-5 Regular P3HT Random P3HT PT TQ1 PSBTBT MEH-PPV 0.0 0.5 1.0 1.5 2.0 Absorbance Figure 3.8. Absorbance resolved degradation rates for six different polymers While the precision relative stability estimation was not perfect in the initial absorbance basis, thickness as a basis can also be considered. It is far more cumbersome to use thickness as the basis of comparison as the thickness must be measured externally. Using AFM thickness / initial absorbance relations were established for all the polymers used. This allowed the degradation rates to be plotted in terms of thickness. This plot is different from the absorbance based plot since the materials have vastly different extinction coefficients. The relative stabilities with P3HT as a basis, is plotted in Figure 3.9 (right) with thickness as a basis. 15

It is clearly evident that this basis is inferior to the absorbance basis. It was therefore concluded that using initial absorbance as the basis of comparison was indeed the best choice. 10 1 Relative stability (P3HT stability) Relative stability (P3HT stability) 10 0 10-1 10-2 10 2 10 1 10 0 10-1 Regular P3HT Random P3HT PT TQ1 PSBTBT MEH-PPV 0.0 0.5 1.0 1.5 Absorbance Regular P3HT Random P3HT PT TQ1 PSBTBT MEH-PPV 10-2 0 100 200 300 Thickness (nm) Figure 3.9. Absorbance resolved stabilities (right and thickness resolved stabilities (left) in units of P3HT stability for the studied polymers. The effect of adding a fullerene derivative to the polymers was studied extensively within the article. For each of the studied polymers, their respective blends in a ratio of 1:1 with PCBM were studied and the blends of P3HT with 5 different electron acceptors were documented. It was shown that the absorbance basis remained the better choice as compared to the thickness basis (see Figure s9-s11 in Appendix 2.2 (supporting materials)) 5. The photochemical stability of blends of conjugated polymers and electron acceptors is a topic that has only been briefly discussed in the literature. Rivaton et al. evaluated the stabilities of regio-regular P3HT and P3HT:PC60BM (1:1 ratio) and reported a stabilization factor of 8. They used a thickness basis of comparison, where films of approximately 200 nm compared. 9 The observed degradation rate was in good correlation with the results observed at Risø for the same blend. The degradation rates were observed to vary with an order of magnitude between the most unstable blend, P3HT:ICBA, and the most stable blend, P3HT:C60. Significant variations in relative stabilities were observed for the different electron acceptors with C60 stabilizing by a factor of approximately 10 while ICBA is observed to destabilize the blend by a factor of 2. The magnitude of the stabilization of P3HT by the electron acceptor was observed to correlate clearly with the LUMO LUMO gap in the low absorbance range. A ranking of decreasing stabilization of C 60, PC 60 BM, PC 70 BM, bispcbm, and ICBA was found, which is in clear correspondence with a decreasing LUMO LUMO gap or increasing open circuit voltage of the corresponding solar cells. Overall, this result demonstrated the increasing thermodynamic tendency of increasing the population of excited states on the P3HT relative to the acceptor, thus implying a higher degradation rate. For this reason, the application of 16

ICBA in PSCs to obtain 6.5% efficiency 10 introduces a significant decrease in photochemical stability that in turn will affect the operational device lifetime. Studying the different polymer blended with PCBM the general expectation was that a highly unstable material should benefit highly from being blended with PC 60 BM, since each excitation has a large possibility of leading to a degradation event. For a highly stable material this effect is less pronounced. This exact tendency was observed as the unstable MEH-PPV is highly stabilized by a factor of around 15, while the stable PT is only stabilized by a factor 3. Additionally, PSBTBT is found to destabilize slightly by a factor of 0.3. A destabilization is expected if the polymer is comparable or more photo-chemically stable than the electron acceptor. This is the case for PSBTBT, where for absorbance above 1, the polymer stability even exceeds the stability of PC 60 BM. For this material combination a charge transfer to PC 60 BM will induce a larger degradation rate than by keeping the excited electron on the pure polymer. The second contribution came with the article entitled: Influence of processing and intrinsic polymer parameters on photochemical stability of polythiophene thin films (Appendix 2.3) 6. This article expanded upon the work of the previous article by investigating the influence of processing and intrinsic parameters on the photo-degradation of P3HT. As in the previous article it is common to express stability in units of stability of a reference material of wellknown stability, typically P3HT. This assumes that P3HT presents an intrinsic, constant stability that is independent of synthesis routes, regio-regularity (RR), molecular weight, molecular weight distribution, crystallinity etc if the relative stabilities are compared across different experiments. The overall effect is that the material stabilities expressed in units of P3HT stability as reported in the literature may be associated with significant uncertainty and cannot be compared directly. Furthermore, until this article, development of stable conjugated polymers for PSCs has been focused on the stability of the different functional groups used for the synthesis. However, understanding the influence of the above described intrinsic polymer properties on the photo-chemical stability is highly appealing, since this will provide a new set of tools when designing novel materials for PSCs. In the article 18 different batches of P3HT from different manufacturers and batches made in house were tested and compared. By studying films of different thicknesses insight into oxygen availability in the film and effects of light shielding could be discussed. Assuming that oxygen diffusion is not limited and that light shielding from the top layer of the film is insignificant, the concentration of oxidized thiophene rings is independent of film thickness. Figure 3.10 shows a plot of degradation event interval against film thickness and a plot of total film lifetime. The existence of a constant lifetime region implies that the degradation takes place in parallel for the entire depth of the film. This means that for this region light shielding is negligible and oxygen is equally available for all depths in accordance with the findings of Hintz et al. 11 For films thicker than 175 nm, either light shielding or lack of oxygen sets the bottom part of the film apart from rest of the film with a lower degradation rate. The event interval is therefore observed to stabilize in this region. The conclusion is consistent with observations of blueshift kinetics. For films in the stable region of 125 175 nm, the blueshift occurs late near the last 20% of the degradation. For films thicker than 175 nm the blueshift appears earlier. This is consistent with the fact that parts of the film degrade later than the top part of the film, thereby extending the degradation. The fast blueshift for thin films (<75 nm) indicates that another mechanism is involved in this region. A candidate for the increase in reaction rate is the higher surface to volume ratio. If the reactions are more likely on the surface the rate may easily be different. The polymers in the top layer can be expected to have a higher density of kinks, introducing more attack points for the reaction and explaining the fast blueshift observed for thin films. 17

60 Degradation Event Interval [fs] 50 40 30 20 25 75 125 175 225 275 Thickness [nm] 70 55 Film Lifetime [h] 40 25 10 25 75 125 175 225 275 Thickness [nm] Figure 3.10. Degradation event interval (left) plotted against the thickness of a film of R1 polymer. The film lifetime (right) as calculated from the time between degradation events and the initial number of monomers. P3HT polymers with significantly different M w and RR are included in the study. The first observation was that while the molecular weight did not seem to play an important role while the regio-regularity did. This was consistent with work presented by Hintz et al. 11. and Dupuis et al. 12 A hypothesis was established that in accordance with observations by Hintz et al. 11 the polymer is attacked only at terminal thiophene units. Assuming that each breach of regularity introduces two new attack points, it was possible to model the degradation rate as a function of regio-regularity. The relative number of attack points was written as N ap ( RRx ) ( RR ) 2 1 = 2 1 R1, where N ap is the number of attack points relative to R1, RR X is the regio-regularity of the specific polymer, and RR R1 is the regio-regularity of polymer R1 used for normalization. Figure 3.11 shows a plot of the normalized degradation rate as a function of regio-regularity and relative number of attack points. The dotted line in the graph is the theoretical value of degradation rate, calculated from the degradation rate of R1. It is evident that the simple model is capable of explaining the behavior in a convincing manner, suggesting that each breach of regularity induces new attack points that weaken the system. The conjugation length is proportional to the regio-regularity since the conjugation breaks when the polymer is not planar and the π electrons are not in the same plane. 18

Regio Regularity [%] 100 95 90 85 80 75 70 Normalized Degradation Rate 3.0 2.0 1.0 0.0 0.0 1.0 2.0 3.0 Relative Number of Attack Points Figure 3.11. Normalized degradation rate plotted against the calculated relative conjugation length / regio-regularity. The dotted line represents the predicted degradation rate. In the article it was demonstrated that annealing the films of P3HT increased the stability, see Figure 3.12. While it was documented that the crystallinity of the films increased for regio-regular films, it was also shown that regio-random films increased in stability. It was therefore concluded that the crystallinity plays a minor role in the stability. The effect of the stabilization is instead ascribed to the relaxation of the polymer leading to fewer high energy kinks. Normalized Deg. Rate 1.2 1.0 0.8 0.6 0.4 1/Crystallinity rr-p3ht rra-p3ht Pristine 120 C 140 C 1.2 1.0 0.8 0.6 0.4 1/Crystallinity Figure 3.12. (Left scale) Degradation rate of (dark grey) regio-regular and (white) regio-random P3HT normalized to their respective pristine degradation rates. (Right scale) Reciprocal crystallinity as deduced from X-ray diffraction studies. 19

References 1. Lilliedal, M.R., Medford, A.J., Madsen, M.V., Norrman, K. & Krebs, F.C. The effect of post-processing treatments on inflection points in current voltage curves of roll-to-roll processed polymer photovoltaics. Sol. Energy Mater. Sol. Cells 94, 2018-2031 (2010). 2. Madsen, M.V., Norrman, K. & Krebs, F.C. Oxygen- and water-induced degradation of an inverted polymer solar cell: the barrier effect. Journal of Photonics for Energy 1, 011104 (2011). 3. Norrman, K., Larsen, N.B. & Krebs, F.C. Lifetimes of organic photovoltaics: Combining chemical and physical characterisation techniques to study degradation mechanisms. Sol. Energy Mater. Sol. Cells 90, 2793-2814 (2006). 4. Norrman, K., Madsen, M.V., Gevorgyan, S.A. & Krebs, F.C. Degradation patterns in water and oxygen of an inverted polymer solar cell. J. Am. Chem. Soc. 132, 16883-16892 (2010). 5. Tromholt, T., Madsen, M.V., Carlé, J.E., Helgesen, M. & Krebs, F.C. Photochemical stability of conjugated polymers, electron acceptors and blends for polymer solar cells resolved in terms of film thickness and absorbance. J. Mater. Chem. 22, 7592-7601 (2012). 6. Madsen, M.V. et al. Influence of processing and intrinsic polymer parameters on photochemical stability of polythiophene thin films. Submitted to: Polymer Degradation and Stability (2012).at <http://linkinghub.elsevier.com/retrieve/pii/s0026269286800167> 7. Carlé, J.E. et al. Comparative Studies of Photo Chemical Cross-linking Methods for Stabilizing the Bulk Hetero-Junction Morphology in Polymer Solar Cells. Submitted to Journal of Materials Chemistry (2012). 8. Manceau, M. et al. Photochemical stability of π-conjugated polymers for polymer solar cells a rule of thumb. J. Mater. Chem. 21, 4132-4141 (2011). 9. Rivaton, A. et al. Light-induced degradation of the active layer of polymer-based solar cells. Polym. Degrad. Stab. 95, 278-284 (2010). 10. Zhao, G., He, Y. & Li, Y. 6.5% Efficiency of polymer solar cells based on poly(3- hexylthiophene) and indene-c(60) bisadduct by device optimization. Advanced materials (Deerfield Beach, Fla.) 22, 4355-8 (2010). 11. Hintz, H. et al. Photodegradation of P3HT A Systematic Study of Environmental Factors. Chem. Mater. 23, 145-154 (2010). 12. Dupuis, A., Wong-Wah-Chung, P., Rivaton, A. & Gardette, J.-L. Influence of the microstructure on the photooxidative degradation of poly(3-hexylthiophene). Polym. Degrad. Stab. 97, 366-374 (2012). 20

2/Results description, IMEC/IMOMEC: Lifetime and modeling Lifetime and modeling analysis was performed on P3-PCBM Cinnamoyl material with and without UV-treatment and on the P3HT-PCBM reference material. All samples were prepared on a glass plate and were topped with a glass encapsulation lid. The photovoltaic cells were tested @ 65 C and @ 80 C. The general approach is to establish the lifetime of the solar cells by performing statistical analysis on the measured data that was obtained on a statistically relevant population of samples at each individual ageing condition. Time zero is taken at the point where the efficiency is at its maximum and not the initial value. This is a little bit arbitrary but all samples behave in the same way: in the beginning an increase in efficiency occurs and then degradation sets in. As this stabilizes, degradation sets in. For this reason, the maximum of the curve is taken as time zero. The 20% efficiency decrease is referenced in this way. Figure 1: References for the 20% efficiency decrease This data was entered in FAILURE. This program was created by the company DESTIN, a spinoff of the university of Hasselt in collaboration with IMEC. The software makes it very easy to fit different experiments done at different temperatures in one run. The data is simultaneous fitted finding a unique β (scale parameter) for both experiments. The result of the FAILURE-analysis for non UV treated samples is shown in figures 2. A Weibull distribution was used to fit the two datasets. A prediction for 25 C is also included in the analysis. 21

Figure 2: Weibull distribution of the samples that were not UV-treated We also investigated the degradation behavior of UV-treated samples @ 65 C and @80 C. The next graph shows the results for the UV treated samples. For the test that was performed @ 65 C, only a few samples gave useful results (see later). Figure 3: Weibull distribution of the samples that were UV-treated From the data, it was clear that the UV cured samples proved to be more stable at the highest operating temperature as expected. The scale parameter of the distribution is clearly higher (165 hours vs. 127 hours) and the activation energy for both degradation processes is rather low with values in between 0.3 >> 0.4eV. This means that even higher operating 22

temperatures will not shorten the lifetime by a large amount. On the contrary this also means that the lifetime will not improve that much by lowering the operational temperature. Furthermore, a software tool was developed to determine the evolution of the series and shunt resistance and photocurrent as a function of time during the degradation experiment. An example is shown in the figure below where an increase of the series resistance and drop in shunt resistance can be observed. Figure 4: Series resistance during degradation Figure 5 Shunt resistance during degradation 23

Figure 6: Photocurrent during degradation Appendix: lessons to be learned from the project for future investigations A lot of degradation measurements have been performed and it was observed that there were some critical points to be tackled in the future. 1. Loss of electrical contact during ageing measurements. This has been a major hurdle during the project especially since a large amount of samples were evaluated. The consequences of intermittent electrical contact can be observed in figure 7. 4.00 Eff [%] Glass_1 3.50 3.00 2.50 ] f [% 2.00 E 1.50 1.00 0.50 0.00 0 50 100 150 200 250 300 350 400 t [hr] Figure 7: Eff.vs. time for different samples in one testing lot. 2. Glass lid encapsulation not bullet proof After extensive testing, it was observed that the glass lid encapsulation proved to be insufficient. The sequential heating/cooling cycles proved to be problematic leading to penetration of air. A calcium deposited glass lid was used to prove the penetration under sequential heating /cooling cycles as it is the case during regular experiments. 24

Figure 8: Ca deposited on glass plate before ageing cycle (mirror finish) Figure 9: Ca deposited on glass plate after 500 hours of ageing cycle (transparent) In a statistical approach, all these failures are called extrinsic failures as these are not intrinsic but are production related. In a normal production environment, extrinsic failures are much lower in amount compared to the intrinsic failures. It is of great importance that the production related failures can be diminished substantially for future investigations. This will lead to a better determination of the lifetime of organic solar cells. 25

3/Objective: definition of the requirements for reliable module fabrication under industrial conditions 3/Results summary: During the projects duration, Risø DTU has developed a baseline process for fabricating polymer solar cells that has been implemented with Danish industry partners who did not participate in the PolyStar project. Nonetheless, this baseline process is the fundament on which many results in PolyStar rely. This baseline system is a bulk heterojunction polymer solar cell of inverted geometry produced according to DTU s already published process for roll-to-roll coating and printing of the solar cell, the so-called ProcessOne, (Krebs F. C., Gevorgyan, S. A., Alstrup J., A roll-to-roll process to flexible polymer solar cells: model studies, manufacture and operational stability studies. J. Mater. Chem. 2009, 5442-5451.). 3/Results summary: Complementary to the processing work developed at Riso-DTU, also at imec investigations have been done on alternative deposition techniques to spin coating, the most used technique in laboratories for the production of polymer based organic solar cells. The promise of these devices, indeed, lies in their low-cost high throughput manufacturability, but this low cost aspect can only be fully realized if the layers are deposited by inline compatible methods. In addition to Riso-DTU studies, ink-jet printing and spray coating are studied in depth at imec for the deposition of the active layer. A thorough investigation on the behavior of spray coating is reported (Mi6.1), together with the optimization of the deposition of polymer based layers. The development of two-solvent systems extended the uniformity of the sprayed films owing to the spreading properties induced by outward Marangoni flows in the liquid. The benefits of the enhanced uniformity are clear from the closing of the performance gap between spin coated devices and spray coated ones. The technique is particularly advantageous since the deposition is obtained as a superposition of micron-sized droplets, which either merge into a full wet layer or dry upon impact with substrate. In the latter case, the damage of the solvent on the previously deposited layers is reduced to minimum, so that spray coating can be used to deposit even a metal top contact from a metal nanoparticle based ink (Mi6.2). This allowed for the demonstration of fully spray coated devices, showing that spray coating is a good candidate for the replacement of spin coating, since it is a large-area, roll-to-roll scalable technique that can produce spray coated organic solar cells with efficiencies in the same range of spin coated/evaporated reference devices. These results are for TOWARD EFFICIENT AND PRINTABLE ORGANIC SOLAR CELLS a large part obtained through the Ph.D. study:, Claudio GIROTTO, January 2011, Katholieke Universiteit Leuven Faculty of Electrical Engineering Kasteelpark Arenberg 10, B-3001 Leuven (Belgium) Results description: The ProcessOne polymer solar cell is a structure comprising 5 layers of individual functionality; a transparent front electrode facing the sun, an electron-transporting layer, a photoactive layer, a hole-transporting layer and finally a metallic back electrode, see Figure 13. 26

Figure 13: The multilayered ProcessOne solar cell comprises the following layers; a transparent front electrode, a electron-transporting layer, and active layer, a holetransporting layer and a metal back electrode. The heart of the solar cells is the active layer which is an intimate mixture of an electron donor (P3HT) and an electron acceptor (PCBM) forming a bulk heterojunction. The heart of the ProcessOne solar cell is the active layer where the sunlight is absorbed and converted to an electrical current. The active layer consists of an intimate blend of an electron donor material and an electron acceptor material, respectively the light-absorbing P3HT (poly(3-hexylthiophene) and PCBM (phenyl-c61-butyric acid methyl ester). This blend is chosen, because it is well researched and serves as a standard blend for any work within polymer solar cells. When light shines on the active layer blend, electrons in the light-absorbing donor material P3HT will be photoexcited leaving behind positively charged holes. If the electrons are not physically removed from the site of excitations, they will sooner or later recombine with their counterpart, the positively charged hole. However, as the active layer is an intimate blend, the regions of the P3HT donor - and the PCBM acceptor material are separated only by some nanometers, the charge carrier can thus readily diffuse from the point of excitation to the boundary between the donor - and the acceptor material where charge separation takes place. The blend forms a three-dimensional junction between the donor and acceptor material, a so-called bulk heterojunction, which is the equivalent of the silicon solar cell s planar p-n junction. The interfacial area of the bulk heterojunction is, however, orders of magnitude larger than the planar heterojunction. At the junction where the negatively charged electrons are separated from the positively charged holes, a photocurrent is constituted that has to be extracted from structure. For this reason the active layer is sandwiched between two current-transporting layers separated by an engineered potential. The engineered potential ensures that the electrons move into the electron-transport layer and the holes into the hole-transport layer from which they are collected by the two electrodes. The materials of the two electrodes, respectively the front - and the back electrode, are chosen to match the potential of the two current transport layers. As the sunlight has to enter unhindered through several layers to reach the photoactive layer embedded in the middle of the cell, it is necessary for the layers in front of the active layer to be transparent for light within the spectrum absorbed by the photoactive layers. This fac- 27

tor limits the choice of materials for the front electrode and the electron-transporting layer drastically. ProcessOne s electron-transporting ZnO layer is by means of nano-particles formulated to be transparent. ITO serves presently as the standard material for the transparent front electrode, due to its availability. ITO is however expensive and requires energyintensive vacuum processing. Finding more cost- and energy-effective alternatives is accordingly a hot R&D topic, and the effort is most likely to pay off in the near future. ProcessOne applies, PEDOT-PSS as hole conductor, transparent ITO as front electrode and a silver back electrode. The silver back electrode might be either a grid or fully covering. Figure 14 shows an overview of the various steps in ProcessOne, starting from the purchased PET substrate coated by a fully covering layer of ITO. The first step in ProcessOne is a patterning of the ITO, in an etching process which removes the conducting ITO in thin strips that defines the boarder of each cells. This is done by screen printing an etch resist onto the ITO, Figure 14 a). The film is then taken through an etch bath with CuCl 2 acid which effectively etch away areas not covered with the etch resist, Figure 14 b). The resist is subsequently chemically stripped off, Figure 14 b) and the film is dried. The now patterned ITO is now run through the slot coating line three consecutive times, Figure 14 c-d). This is the essential part of the process as it is where the three OPV specific layers are processed, the electron-transporting layer, the active layer and the hole conducting layer. Subsequently and in a separate line the metal back electron is screenprinted onto the slot-die coated layers, Figure 14 f) and the foil is finally laminated in a moisture and oxygen protective barrier by means of a pressure-sensitive adhesive, Figure 14 g). The output from ProcessOne is a foil comprising individual solar modules laminated in a protective barrier, see Figure 15. The band is ready to be cut into individual cells or ready to be further roll-to-roll processed into more refined products. Figure 14: ProcessOne comprises 7 individual steps shown from a) to g). 28

Figure 15: The roll of laminated solar module coming out of ProcessOne (left) and closeup of one individual module consisting of 16 strip-shaped cells connected in series. 4/Objective: demonstration of the developed technology by prototyping products powered by our flexible PV modules 3/Results summary: DTU have successfully demonstrated the medium scale manufacture and product integration of OPV at the level of 10 000 units with a high technical yield. The demonstration that comprised an OPV flashlight was demonstrated and distributed freely at the LOPE-C conference 2011 and it became the official demonstrator of the Organic Electronic Association (OEA). At the same time, it was also a very successful demonstration of technology transfer to industry: The manufacture of the demonstrators started with full roll-to-roll manufacture of the solar cells using a polymer solar cell coater from Grafisk Maskinfabrik (at Risø DTU) followed by lamp assembly (at Mekoprint) in sheets of 15 lamps. The technical yield of the demonstrator is approximately 89% based on all steps. This technical yield is exceptionally high considering that the entire manufacturing process comprises more than 35 discrete steps that are only tested twice in the process (OPV function and lamp function). To achieve this all steps were performed in automated processes except for the mounting of the batteries and the final testing of the lamp function which were done manually. We demonstrate the possibility to miniaturize OPV and illustrate avenues for adaptability and easy product integration. This chapter describes how the polymer solar cells are turned into usable devices. This is a matter of shaping and sizing the modules for use in products, and a matter of protecting the modules by barriers. The chapter describes furthermore the polymer technology s learning curve and the environmental impact of the ProcessOne devices that are both of vital importance for the market acceptance. The analysis of the environmental impact provides furthermore valuable information for the further development of the devices into fully sustainable products. 4/Results details: Miniaturized devices Many potential applications for polymer solar modules require highly limited current, for example charging of small Li-polymer batteries. For such applications a small credit-card sized module was developed. The module comprises 16 serial connected solar cells spaced by 1 mm. Three sets of 16 stripes were prepared simultaneously on the standard 305 mm wide 29

web. Each printed motif (305 mm x 305 mm) presented 15 independent modules. The processing of the credit-card sized cell device is shown in Figure 16. Figure 16: R2R manufacture of the credit-card sized devices. A. The PET foil with ITO, ZnO, P3HT:PCBM. B: Slot-die coating of PEDOT:PSS. C. Screen printing of the electrical connection between the individual cells. D. Lamination, from (Krebs F. C. et al, 2011) The narrow stripe wide (3 mm) and the relatively high conductivity of PEDOT-PSS enabled the successful preparation of this devices without a metal back electrode. Silver was only printed in thin stripes to connect cells and thus not formally as a back electrode. The miniaturization brings the challenge of being able to keep the registration marks that are printed along the web for correct juxtaposition of subsequent layers with respect to the patterned ITO, see Section Fejl! Henvisningskilde ikke fundet.. This challenge is however slightly eased as the metal back electrode is left out. The use of PEDOT-PSS as back electrode implies that the initial performance of the device is significantly higher that what could be obtained with the standard geometry s printed silver back electrode, both full and grid. This implies also that the device performance presents an initial drop due to a drop in conductivity of the PEDOT-PSS back electrode, see Figure 17. Humidity and variations in humidity surrounding the device during operation have previously been shown to cause phase separation in the PEDOT-PSS layer (Krebs F. C. et al, The OE-A OPV demonstrator anno domimi 2011. Energy Environ. Sci. 2011, 4116-4123), which will cause a drop in conductivity and a corresponding drop in device performance. The initial power conversion efficiency (PCE) of the solar cells was found to be around 2 % when tested at 1.5 suns. Upon 3 months storage 1 on the roll the performance dropped significantly, see Figure 17. The main degradation happens to the generated current and not to the module voltage which is critical for the charging of a battery. This implies that even though the charging efficiency will decrease over time, it will not lead to complete failure as would be the case if the voltage dropped below the charging voltage of 4.7-5.2 V. It would have been possible to print full silver back electrodes and maintain a higher performance over time, but on the expense of a higher materials cost and lower technical yield for such miniaturized devices. The paper The OE-A OPV demonstrator anno domimi 2011 (Krebs F. C. et al, 2011) refers all further details of the credit-card sized device. 1 ) at room temperature (22+5 C) and ambient humidity (40+10 % rh) 30

Figure 17: The performance of 1000 modules as prepared (black) and the same modules after 3 month storage on the roll, from (Krebs F. C. et al, 2011) Barriers for low- and medium demanding applications Standard ProcessOne prescribes the use of a food packaging barrier from Amcor which is laminated on both sides of the modules in a roll-to-roll process by means of an optically clear pressure sensitive adhesive, 467 MPF from 3M. This encapsulation has been chosen, because its cost/performance ratio harmonizes with the polymer solar cell s cost-performance profile. The food packaging barrier is suited for applications where the operational life length of the device is not critical, and where the device is not exposed to mechanical stress. For more demanding applications, an additional outer encapsulation might be required in the form of a protection against the unavoidable repetitive bending and buckling that might damage flexible devices over time, or in the form of more sophisticated encapsulations designed for products that shall last for many years under outdoor conditions, see Section Fejl! Henvisningskilde ikke fundet.. The performance of the miniaturized devices which was protected with the Amcor foil and had a PEDOT-PSS back electrode has been observed to drop over time, see Section 0. This behaviour is not observed for the standard device, and it is explained by humidity variations in the PEDOT- PSS layer, see Section 0. In order to search for at better protection against humidity two alternatives to the Amcor barriers have been tested for the miniaturized device; a barrier from Fuji- Film and a self-made barrier based on a 100 µm thick PEN foil coated by 150 nm silicon nitride (Si x Ni y ). The stability of the miniaturized devices when protected by the three foils is shown in Figure 18. The two dashed lines mark the acceptance threshold for respectively stability, T80 2 and open circuit voltage, V oc. It is notable the Fuji barrier and the home-made barrier (PEN/Si x N y ) samples are bimodal in performance with half of the devices above and half below the T80 threshold of 100 hours. In comparison 7 out of 8 Amcor samples are above the T80 threshold, promising a more consistent performance. 2 ) the time at which the cell has degraded to 80 % of its peak efficiency 31

The paper The OE-A OPV demonstrator anno domimi 2011 (Krebs F. C. et al, 2011) gives a comprehensive view of the work on the various barriers. Figure 18: Stability of credit card sized devices represented by T80. The devices were protected by three different barriers with two different barriers from respectively Alcan and Fuji, and one home-made barrier (PEB/SixNy). The dashed lines represent thresholds for satisfactory devices; T80 > 100 hours and V oc > 5V, from (Krebs F. C. et al, 2011) 32

1.4 Utilization of project results All results in the Danish part of the PolyStar project were published in a total of 30 publications excluding reports. The project is a part of a larger plan for R&D on polymer solar cells with the objective of developing polymer solar cells for power production. When the project ended the ability to produce polymer solar cells for powering consumer electronics had been successfully transferred to an industrial Danish partner to DTU. The manufacture of the demonstrators reported previously in the report was a very successful demonstration of the current applications of the technology. The industry partner continues to improve their ability to make OEM type applications for polymer solar cells; e.g. a OPV powered autonomous sensor integrated in an adhesive label. The key point here is to automate fabrication; to improve the technical yield; and to improve product lifetime a task where DTU and the partner continuous collaboration. The road towards power production is still challenging with respect to efficiency and new materials; stability; avoiding the use of energy intensive processes and improving production speed; reducing the environmental footprint i.e., avoiding the use of scarce materials, materials with large embedded energies, and harmful solvents and materials. It is generally believed that 10% efficiency, 10 years of lifetime and module production speeds superseding 10 m/min are needed to turn polymer solar cells into a competitive alternative to today s photovoltaic technologies. DTU continues to pursue these objectives in collaboration with both industry and academic partners all over the world. No patents were taken as result of the PolyStaR efforts. All results have been made public available and knowledge obtained on ProcessOne was transferred to industry - which has resulted in a Danish production of polymer solar cells for OEM applications. DTU holds patents in the field of polymer solar cells but up to now they have not been licensed to industry. Future inventions will be protected and industrial licensing will be considered. Major parts of the results will continue to become public knowledge through dissemination. The PhD student Morten V. Madsen (MVES) has completed a 3 year PhD program funded through the PolyStaR project. Dissemination was in the form of presentation of group activities as well as personal research both for visiting groups as well as outside Risø Campus. The external research stay consisted of 5 short visits to laboratories in Austria, Germany and Sweden with the objective of training the Ph.D. student in ellipsometry and to perform mutual research activities. MVES has attended 4 conferences and a workshop in USA, Greece, Denmark, Italy, and Germany. In addition, MVES participated in a panel discussion at a conference in Italy. The work that MVES was involved in has been published in various international journals. The Ph.D. student co-author on 14 peer reviewed papers. Short synopsis of the Ph.D. study objectives and results: The subject of MVES's PhD work has been focused on lengthening the lifetime of roll -to-roll coated polymer solar cells by studying degradation behavior of polymer solar cells and polymer solar cell materials. MVES has successfully contributed to 4 projects relevant to the subject of the PhD work. In the first part of the PhD MVES invested a lot of work on attempting to implement ellipsometry as a novel non-destructive and non-invasive analytical tool to describe degradation behavior in polymer solar cell materials. However, after the realization that this was not possible MVES reacted constructively and wisely by shifting his focus to other projects. MVES got affiliated to a project that started a couple of years before MVES started his PhD study. The project 33

employed the analytical tools TOF-SIMS and XPS in order to study degradation behavior of polymer solar cells. MVES successfully contributed to the project by acquiring a lot TOF-SIMS data. MVES has proven to be very efficient in handling large systematic data sets, which was manifested in the creation of software that could ease the data evaluation procedure. The contribution from MVES made it possible finalize the project that was successfully published in the high impact journal JACS. A very successful work from MVES is work that he did in collaboration with his fellow PhD student Thomas Tromholt, Professor Frederik Krebs, and two resourceful technicians that rebuilt existing equipment. Two very valuable projects was the outcome of this collaboration. In the first project a fully automated degradation setup was constructed where materials and polymer solar cells could be degraded under standard conditions. It was possible to degrade 20 samples in parallel, which provided a sound statistical basis for any conclusions on stability. A rigorous study of the validity of photochemical testing of polymer stability was carried out with this setup. The automated setup is a very valuable tool for the solar cell group at DTU, so the contribution from MVES to this project has been important. The highest impact from the aforementioned collaboration was work on concentrated light. Experimental setups were constructed that were cable of concentrating light originating from either the sun or from artificial light sources. The setup was used to study accelerated photochemical degradation of polymers. Degradation rates were accelerated by up to a factor of 1200. In addition, from the observed stability at accelerated conditions, the stability at standard conditions could be predicted with high precision. Concentrated light was thus shown to be a powerful tool for rapid, precise polymer stability evaluation. MVES has successfully contributed to this very important project that has significantly impacted the field of degradation and stability of polymer solar cell materials. 1.5 Project conclusion and perspective DTU have successfully demonstrated the medium scale manufacture and product integration of OPV at the level of 10 000 units with a high technical yield. The demonstration that comprised an OPV flashlight was demonstrated and distributed freely at the LOPE-C conference 2011 and it became the official demonstrator of the Organic Electronic Association (OEA). At the same time, it was also a very successful demonstration of technology transfer to industry: The manufacture of the demonstrators started with full roll-to-roll manufacture of the solar cells using DTU s polymer solar cell coater manufactured by a Danish company followed by lamp assembly (at a Danish industry partner) in sheets of 15 lamps. The technical yield of the demonstrator is approximately 89% based on all steps. This technical yield is exceptionally high considering that the entire manufacturing process comprises more than 35 discrete steps that are only tested twice in the process (OPV function and lamp function). All results in the Danish part of the PolyStar project were published in a total of 30 publications excluding reports. The project is a part of a larger plan for R&D on polymer solar cells with the objective of developing polymer solar cells for power production. When the project ended the ability to produce polymer solar cells for powering consumer electronics had been successfully transferred to an industrial Danish partner to DTU. The manufacture of the demonstrators reported previously in the report was a very successful demonstration of the current applications of the technology. The industry partner continues to improve their ability to make OEM type applications for polymer solar cells; e.g. a OPV powered autonomous sensor integrated in an adhesive label. The key point here is to automate fabrication; to improve the technical yield; and to improve product lifetime a task where DTU and the partner continuous collaboration. The PolyStar project was conducted within the framework of ERA-NET together with partners from Flanders, Belgium: IMEC, Cytec Surface Specialties NV,SA. The Danish project and the ERA-PV was funded and managed under Energinet.dk, FORSKEL programme, whereas the 34

Flemish project entitled OPVLife was funded under IWT (Agentschap voor Innovatie door Wetenschap en Technologie). OPVLife was started on March 2009. The PolyStar project period was 1/4 2009-30/6, 2012. The international collaboration was successful in the area of polymer solar cell stability. This work conducted by both DTU and IMEC formed the basis for increasing operational life of polymer solar cells. The third International Summit on Organic Photovoltaic Stability (ISOS- 3) took place at Risø DTU in 2010. ISOS is a transnational network and the PolyStar collaboration was a corner stone for a large part of the work presented at ISOS-3. One Ph.D. student and one postdoc was trained during the project period. 1.6 Annual export of electricity (only ForskVE) 1.7 Updating Appendix P and submitting the final report DONE! 35