Fabrication of Polymer Solar Cells Using Aqueous Processing for All Layers Including the Metal Back Electrode

Abstract

The challenges of printing all layers in polymer solar cells from aqueous solution are met by design of inks for the electron-, hole-, active-, and metallic back electrode-layers. The conversion of each layer to an insoluble state after printing enables multilayer formation from the same solvent (water). The photograph here was taken just before screen printing of the aqueous silver ink. Utilization of sunlight as an energy source is one of the least exploited carbon-neutral methods available today. The potential is enormous, with a wide range of possible applications such as large scale energy production, small standalone energy production units in remote areas situated ‘off the power grid’ or tiny power production units aimed at recharging small electronic equipment that we surround ourselves with. Over recent years polymer and organic solar cells have been perfected at the laboratory level and the performance now approaches many of the inorganic thin film solar cells. It has been argued that the ∼8% power conversion efficiency recently reported for polymer solar cells might challenge polycrystalline silicon when projecting the steady increase in polymer solar cell performance a few years into the future. Many challenges remain that have to be addressed efficiently before the vision of large scale manufacture and widespread usage of low cost polymer solar cells can be anticipated. Ideally the polymer solar cell should be manufactured in a fast, large-area, environmentally friendly process. The methodologies employed in typical laboratory studies do not represent this well. The most commonly employed film forming technique is spin-coating which is incompatible with large areas, large volume, and low cost. Another troublesome aspect is the use of toxic organic solvents in the film-forming process. In the large-scale application where production volumes corresponding to several GWpeak are envisaged this is not a viable approach and alternative solvents will be a requirement. Until now, solvent-free or environmentally friendly solvent processing have not been studied to any significant level. An explanation for this can possibly be sought in the delicate interplay between the processing solvent and the performance of the solar cell. In many ways the state-of-the-art polymer solar cell has evolved around aromatic solvents such as chlorobenzene, dichlorobenzene, toluene, and xylene. Any change of solvent adversely affects the nanomorphology of the film and the device performance. The ambition to use more benign solvents would require a redesign of the mole­cular structures and a re-establishment of the interplay between nanomorphology and processing for the new material-solvent combination. Most of the top-performing polymer solar cells are furthermore very sensitive to water and oxygen which lead to the requirement that all processes have to be performed in a protected atmosphere that, while possible, makes everything much more complex. Finally, the state-of-the-art polymer solar cell employs evaporated metal electrodes such as aluminium as the back electrode – a process that does not at all correlate well with high throughput production because of the high vacuum required for the deposition. The general processing steps of laboratory solar cells are shown in Figure 1. We present an alternative environmentally friendly process that is also outlined in Figure 1 where all the processing steps are aqueous and vacuum deposition of metal electrodes has been replaced by printing of an aqueous metal ink. The geometry of the device is of the “inverted” type with a zinc oxide electron transport layer (ETL) on an ITO coated glass substrate. Comparison of the potential for high throughput production using the standard laboratory buildup of a solar cell, which involves both the use of toxic solvents and the slow metal deposition by evaporation (top) and the all-water-processable method where all steps are processed from aqueous solution, which makes it compatible with high throughput processing like roll-to-roll-coating. Furthermore the use of water based solutions guarantees a minimal impact on the environment (middle). The environmental compatibility of the processing steps is shown (lower left) along with an IV-curve for a device prepared using the aqueous processing of all layers described here (lower right). On top of this is a spin-coated active layer comprising a new developed water soluble polymer and a phenyl-C61-butyric acid methylester (PCBM) type acceptor, which will be discussed later. This is then followed by a hole transport layer (HTL) of poly­ethylenedioxythiophene: poly(styrene sulfonate) (PEDOT:PSS) and finally a printed counter electrode is applied in the form of a silver ink. Water soluble analogues to poly(3-hexylthiophene) (P3HT), which is probably the most abundantly used polymer used in solar cells until now, was first reported in the mid 1980s.1 Introduction of sulfonic acid salts at the end of the side chains provided the solubility, and polymers employing the same principle for water solubility using different ions such as sulfonium,2, 3 pyridinum4 or ammonium salts5, 6 and varying chain lengths have been reported and/or are commercially available. Unfortunately these polymers are not compatible with processing of multilayer films from the same processing solvent (i.e. water) since each new layer can dissolve/interact with previous ones. One way to circumvent this problem is to use thermo-cleavable materials in order to switch off the solubility of the polymer after film processing by removal of side chains. This led us to develop a new thiophene based polymer incorporating this concept shown in Scheme 1. The use of tertiary esters as thermo-cleavable side chains on polymers for solubility switching is a well known process, which, when the tertiary ester is attached to a thiophene unit, can proceed in two steps with increasing temperature by initial removal of the tertiary substituent followed by decarboxylation.7, 8 Both products are insoluble, but have different electronic properties of which the final decarboxylated product has been found to perform best in solar cells.8 Besides the obvious processing advantages of solubility switching, removal of the side chains from the bulk of the active layer furthermore means removal of non-absorbing material, and several studies show enhanced stability of the active layer towards general degradation.9-11 A tertiary ester side chain containing polyether units as well as free alcohol groups was synthesized in order to promote solubility towards aqueous media and simultaneously allowing for evaporative removal of the cleaved side chains from the active layer after cleavage – a feat that would not have been possible if ionic containing side chains had been used. Top: Chemical structure of polymer 1 and the reaction products forming during the thermocleavage process as the temperature rises. Initially poly-3-carboxydithiophene (P3CT) is formed followed by decarboxylation into polythiophene (PT) at higher temperatures. Both compounds are insoluble but have different electronic properties. Bottom: Chemical structure of the modified PCBM-analogue used for aqueous processing. A tertiary alcolhol with two additional tert-butyldimethylsilyl (TBDMS) protected alcohols was prepared and coupled with 2,5-dibromothiophene-3-carboxyllic acid to give the corresponding tertiary ester. This was then reacted with 2,5-bis(trimethylstannyl)thiophene in a Stille coupling afforded the initial polymer which, because of the TBDMS-protected alcohols, could undergo normal workup procedures. Removal of the TBDMS-groups was performed with TBAF in THF/MeOH solution yielding the final polymer 1. The polymer showed to be soluble only in DMF and DMSO when using pure solvents, and was also soluble in a mixture of equal amounts water and isopropanol when a little THF was present (down to 4 vol%). In addition to the thermocleavable polymer a new acceptor material had to be prepared, as the generally used PCBM-acceptor is too insoluble to be processed when water is present. After several failed attempts to prepare a thermocleavable fullerene that could be rendered insoluble by the same processes as the polymer, a different approach was chosen. Previous work by Deguchi et al. have shown it possible to create monodisperse clusters of C60 (60 nm) in water through forced precipitation by addition of a THF-solution into water.12 The thought was that a PCBM-analogue, that was not soluble in water, which would allow for subsequent processing, but still containing coordination properties towards aqueous media, might behave in a similar way if mixed with an aqueous solution from THF. The fullerene 2 (see Scheme 1) containing an ester substituent of triethyleneglycol mono methyl ether to promote coordination to water was therefore prepared from commercial C60-PCBA and triethyleneglycol mono methyl ether by coupling with dicyclohexylcarbodiimide. The purified fullerene was soluble in polar aprotic solvents dimethylformamide, dimethylsulfoxide and tetrahydrofurane (DMF, DMSO, and THF). Inverted structure polymer solar cells (substrate | ITO | ZnO | active layer | PEDOT | Ag, (ITO: indium tin oxide)) were prepared on ITO-substrates. The electron transport layer (ETL) was zinc oxide doped with aluminium obtained from aqueous solution (see Supporting Information). The advantage of the aqueous ink formulation is that the inflection point that ZnO based devices exhibit as a consequence of photodoping is alleviated. The ink could be spin coated or roll-to-roll slot-die coated. An important step is that the initially dry film must be heated for 5–40 min at 140 °C to yield electron-transporting films. After this treatment the ZnO layer is insoluble and subsequent processing is then possible. The active layer was prepared by mixing equal amounts of solutions of polymer 1 (15 mg ml−1 in 47.5% water, 47.5% isopropanol and 5% THF) and the fullerene 2 (15 mg ml−1 in THF) just prior to spin-coating. A thermal treatment (310 °C for approximately 10 s) removed the solubilizing side chains of the polymer leaving an insoluble film. The duration of the thermal treatment is quite critical, as both shorter and longer periods led to a decrease in efficiency of the final devices. This is probably due to either incomplete decarboxylation or some degradation. The hole transport layer (HTL), comprising an aqueous dispersion of PEDOT:PSS diluted with isopropanol, was coated on top of the ZnO/polymer:fullerene stack. After heating for 5 min at 140 °C further processing was possible and the device was completed by printing a silver electrode from a paste comprising only silver flakes, water and an aqueous binder. The final printing step is finalized in 2 min and is finished off by protection of the electrodes by encapsulation using a simple food packaging barrier. All preparative steps were carried out in air. Pictures of the printing process and aqueous solutions are shown in Figure 2. Devices with areas of 0.5 cm2 gave efficiencies of 0.4–0.7% with excellent stability during storage and operation under ambient conditions when encapsulated using the simple food packaging barrier. The J–V characteristics of a representative solar cell are: open-circuit voltage Voc: 0.49 V, short-circuit current Isc: 4.46 mA cm−2, fill factor FF: 32%, and power conversion efficiency PCE: 0.70%. In comparison devices prepared with the well known P3HT:PCBM materials combination showed performances of 1.6–2% when processing the active layer (P3HT:PCBM) from chlorobenzene while using aqueous processing for all the other layers (ZnO, PEDOT:PSS and silver back electrode). These results are admittedly lower than what is commonly reported for P3HT/PCBM devices with evaporated electrodes. We would however like to stress that aqueous processing is a very significant challenge and our results show convincing feasibility for devices with relatively large active area and also show the potential of the printable aqueous silver as an alternative to evaporation. Top: picture of the inks involved in the ‘all water based processing’. From left to right: aqueous ZnO precursor solution, aqueous solution of polymer 1 (diluted in order to show color), aqueous PEDOT:PSS with isopropanol and aqueous silver paste. Middle: Picture of the screen printing machinery. Bottom: Close-up of the screen printing process. The silver paste has been applied and is ready to be pushed though the pattern of the mask by sweeping the squeegee across the steel mesh screen. Insert shows a picture of the final device after encapsulation. In summary we have successfully developed four new methods which allows for aqueous processing of all layers in polymer based solar cells and solves the problem of the inflection point in ZnO. Bulk heterojunction solar cells prepared on the basis of these methods exhibited efficiencies up to 0.70% when using specially prepared water processable polymer and fullerene and up to 2% when processing P3HT/PCBM from chlorobenzene while using aqueous processing for all other layers. This result must be considered an essential first step on the way to an environmentally friendly large scale production of polymer solar cells using only printing techniques and water as the processing solvent. Experimental details on the synthesis of the polymer 1 and of the modified PCBM-analogue, fullerene 2, can be found in the Supporting Information. Device Fabrication: ITO-substrates were first cleaned in isopropanol and water for 10 min respectively using an ultrasonic bath. The ZnO precursor solution was then applied by spin-coating at 1000 rpm and after drying this was followed by heating for at least 5 min at 140 °C to convert to ZnO. The active layer was prepared by mixing a solution of polymer 1 (15 mg/ml) in a mixture of water/isopropanol/THF (47.5:47.5:5) and the fullerene 2 (15 mg/ml) in THF just prior to spin coating at 400 rpm After drying the ITO:ZnO:active layer the substrate was heated on a hotplate at 310 °C for approximately 10 s in order to perform the thermocleavage of the side chains to yield native polythiophene. (For P3HT/PCBM devices (1.2:1) a 44 mg/ml solution in chlorobenzene was spin coated at 800 rpm, followed by annealing at 140 °C for 2 min). This was followed by application of PEDOT:PSS (Agfa EL-P 5010) diluted with isopropanol 2:1 (w/w) by spin-coating at 1000 rpm for 15 s and then drying on a hotplate at 140 °C for 5 min. The device was finished by screen printing of a silverpaste prepared from silver flakes (FS 16 from Johnson Matthey) mixed with an aqueous binder (Viacryl 175W40WAIP from Cytec ) and water, followed by heat treatment at 140 °C for 2 min. Device Testing: The conditions of the characterization under simulated sunlight were KHS 575 using a solar simulator from Steuernagel Lichttechnik operating at 1000 Wm−2, AM1.5G. The spectrum of the solar simulator was checked using an optical spectrum analyzer made for measuring irradiance, and its intensity was calibrated bolometrically using a precision spectral pyranometer from Eppley Laboratories. During measurements, the incident light intensity was monitored continuously using a CM4 high-temperature pyranometer from Kipp & Zonen. Supporting Information is available from the Wiley Online Library or from the author. This work was supported by the Danish Strategic Research Council (2104-07-0022), EUDP (j. nr. 64009-0050) and PV-ERA-NET (project acronym POLYSTAR). Detailed facts of importance to specialist readers are published as ”Supporting Information”. Such documents are peer-reviewed, but not copy-edited or typeset. They are made available as submitted by the authors. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.