"Design and synthesis of novel materials for optimization of the active layer morphology and stability of bulk heterojunction organic photovoltaics."
Due to the tremendous amount of research performed during the last decade, the performance of organic solar cells has strongly been improved and power conversion efficiencies (PCE’s) over 10% are currently reported for single junction polymer-based solar cells. However, for organic photovoltaics (OPV) to evolve into an economically viable technology, three major requirements need to be fulfilled, i.e. a high efficiency, a sufficiently long lifetime and the possibility to produce the solar cells at low cost. In this PhD thesis, work was performed on two of these parameters, performance and stability, through the development of new low bandgap copolymers and small molecule analogues and evaluation of their solar cell characteristics.
At first instance, we have focused on the optimization of the synthesis protocol for alkylated 4H-cyclopenta[2,1-b:3,4-b’]dithiophenes (CPDT’s), one of the main building blocks used throughout this thesis. To circumvent the long and tedious classical synthesis route, in 2010, an alternative three step protocol was developed in our group, which additionally allows for the straightforward introduction of (asymmetrically) functionalized side chains. Since this procedure suffers from relatively low yields for the final ring-closing step, some optimization efforts were performed. By (mainly) variation of the extraction solvent, the yield for this step could be improved from 55 to 74%. Despite the improved reaction efficiency, issues were still observed for more voluminous side chain patterns. This prompted us to develop a complementary synthesis protocol based on a Wittig-type carbonyl olefination reaction and subsequent reductive alkylation. For both steps, the products can be isolated in reasonably high yields (>70%), hence rendering this Wittig route more versatile (for sure for the more complex side chains).
After optimization of the monomer synthesis route, CPDT was copolymerized with a particular thiophene-extended quinoxaline (Qx) derivative to yield a PCPDTQx-type low bandgap copolymer. Due to the electron rich nature of the CPDT unit, this polymer did not only exhibit a small bandgap (1.5–1.6 eV), but also a relatively high HOMO energy level, limiting the solar cell performance through the moderate open-circuit voltage (VOC). In an effort to overcome this, fluorine atoms (1 or 2) were introduced on the Qx unit and the influence on the physicochemical material features and device properties was investigated. While affecting the optical properties to only a minor extent, a significant influence on the positions of the HOMO and LUMO energy levels was observed. For every fluorine atom added, the HOMO level decreased by ~0.1 eV, which translated into an increase of the VOC for the corresponding polymer solar cells by ~0.1 V (up to a maximum VOC of 0.83 V for the PCPDTQx(2F):PC71BM device). However, fluorination also influenced the active layer morphology of the solar cells, leading to the formation of unfavorable large domains (~250 nm) for the PCPDTQx(2F):PC71BM blend, which could not be prevented by the use of high boiling additives, limiting the polymer solar cell performance to 5.26%.
In follow-up work, the amount of side chains on the backbone of the PCPDTQx(2F) polymer was then varied. It was found that the side chain density largely affects the active layer morphology of the solar cells. Whereas domains of approximately 250 nm were observed for the PCPDTQx(2F):PC71BM active layer, a finely intermixed blend composition was obtained by removing 50% of the side chains on the Qx unit, leading to a maximum device performance of 5.63%. Furthermore, besides the influence on the active layer morphology, a large effect on the glass transition temperature (Tg) of the polymers was observed as well. Since the Tg of the polymer:fullerene blend is a crucial factor determining the thermal stability of polymer solar cells, the impact of the side chain density on the thermal stability of the solar cells was also evaluated. Complete removal of the side chains on all Qx units resulted in the best solar cell lifetime. PCE values of >80% of the initial value could be recovered for solar cells based on this polymer after 120 h exposure to a continuous thermal stress of 85 °C, whereas only 50% of the initial efficiency was retained for the device based on the fully alkylated copolymer.
Within our group, it was already shown that the incorporation of functional moieties (ester, alcohol, cinnamoyl) on the side chains of poly(3-alkylthiophenes) can enhance the (thermal) stability of polymer solar cells. This improvement is related to a decreased tendency for phase separation of the polymer and fullerene component (and fullerene crystallization) in the photoactive layer. In this thesis, an effort was done to translate this approach to low bandgap copolymers, affording higher solar cell efficiencies. To this extent, PCPDTBT-type copolymers (BT = 2,1,3-benzothiadiazole) containing alcohol or ester functional groups in the side chains were prepared. For both functionalized copolymers, the resulting polymer:fullerene solar cells showed an improved thermal stability during accelerated aging tests at 85 °C (in comparison to the non-functionalized analogue). However, in contrast to the polythiophene results, the enhanced lifetime could not be attributed to delayed phase separation because of the high Tg values (> 160 °C) for all of the materials (and the resulting blends), even the pristine PCPDTBT. From preliminary deconvolution of the various degradation pathways it seems that the incorporation of alcohol or ester functional groups leads (amongst others) to a higher resistivity toward reduced interface (active layer–cathode) quality.
Finally, we also studied a small series of solution processable small molecule donor materials for OPV applications. The effect of the central donor unit in D-A-D-A-D type (D = donor, A = acceptor) small molecules related to the well-known high-performance DTS(FBTTh2)2 material was investigated. The central dithieno[3,2-b:2’,3’-d]silole (DTS) moiety was replaced by CPDT, BDT (benzo[1,2-b:4,5-b’]dithiophene), DTP (dithieno[3,2-b:2’,3’-d]pyrrole) and TT (thieno[3,2-b]thiophene). The best solar cell performance was obtained for CPDT and TT as central donor units (PCE ~3%). These results remain, however, far below the efficiency obtained for DTS(FBTTh2)2. For BDT(FBTTh2)2 and (especially) DTP(FBTTh2)2, the presence of small amounts of homo-coupled side products was observed by MALDI-TOF analysis, which might explain the unexpected low VOC value obtained for the DTP-based devices. Despite extensive purification efforts, these side products could not be removed.