Date Published: February 02, 2018
Publisher: John Wiley and Sons Inc.
Author(s): Wai‐Yu Sit, Flurin D. Eisner, Yen‐Hung Lin, Yuliar Firdaus, Akmaral Seitkhan, Ahmed H. Balawi, Frédéric Laquai, Claire H. Burgess, Martyn A. McLachlan, George Volonakis, Feliciano Giustino, Thomas D. Anthopoulos.
Fullerenes and their derivatives are widely used as electron acceptors in bulk‐heterojunction organic solar cells as they combine high electron mobility with good solubility and miscibility with relevant semiconducting polymers. However, studies on the use of fullerenes as the sole photogeneration and charge‐carrier material are scarce. Here, a new type of solution‐processed small‐molecule solar cell based on the two most commonly used methanofullerenes, namely [6,6]‐phenyl‐C61‐butyric acid methyl ester (PC60BM) and [6,6]‐phenyl‐C71‐butyric acid methyl ester (PC70BM), as the light absorbing materials, is reported. First, it is shown that both fullerene derivatives exhibit excellent ambipolar charge transport with balanced hole and electron mobilities. When the two derivatives are spin‐coated over the wide bandgap p‐type semiconductor copper (I) thiocyanate (CuSCN), cells with power conversion efficiency (PCE) of ≈1%, are obtained. Blending the CuSCN with PC70BM is shown to increase the performance further yielding cells with an open‐circuit voltage of ≈0.93 V and a PCE of 5.4%. Microstructural analysis reveals that the key to this success is the spontaneous formation of a unique mesostructured p–n‐like heterointerface between CuSCN and PC70BM. The findings pave the way to an exciting new class of single photoactive material based solar cells.
The performance of the organic bulk heterojunction (BHJ) solar cells has been increasing steadily over the last few years, and has reached 13% for the single‐junction solar cells.1, 2, 3 However, these high efficiencies have come at the cost of an increase in the complexity of the cells, where often finely tuned nanomorphologies are required, which renders them both less stable and reproducible, along with difficult‐to‐synthesize polymers and small molecules with high associated production costs.4, 5, 6 In addition, the inherent trade‐off between the short‐circuit current (Jsc) and the open‐circuit voltage (VOC), due the requirement for the lowest unoccupied molecular orbital offset between the donor and the acceptor to be more than 0.3 eV, is thought to be a limiting factor in pushing cell efficiencies to above 15%.7 Owing to the inherent disadvantages associated with the BHJ cell architecture, there has been a recent drive to move toward cell architectures which make use of only a single active layer, with the active material fulfilling the simultaneous roles of light absorption, exciton dissociation, and charge transport, in a manner similar to the emerging hybrid metal halide perovskite cells. A move toward such an architecture should reduce the complexity of the cell fabrication, increasing stability and potentially lowering the production cost.
An essential property of the photoactive layer of a solar cell is for it to be able to transport both holes and electrons. In order to study the charge transport properties of PC60BM and PC70BM, top‐gate, bottom‐contact (TG‐BC) thin‐film transistors (TFTs) were fabricated (Figure1a, inset) using previously reported procedures (see the Experimental Section). Figure 1 displays representative transfer characteristics measured for a PC60BM (1a) and a PC70BM (1b) transistor with channel length (L) and width (W) of 30 and 1000 µm, respectively. Both transistors exhibit balanced ambipolar characteristics as evident by the comparable current levels measured in the p‐channel and n‐channel operating regimes. Table1 summarizes the hole (µh(s)) and electron (µe(s)) mobility values measured in saturation for both transistors as well as values reported in the literature.35, 36, 37, 38 For PC70BM, the calculated µh(s) is the highest reported to date and surpasses the only value found in the literature (2 × 10−5 cm2 V−1 s−1)39 by more than three orders of magnitude. Similarly, the µh(s) value extracted for PC60BM is also the highest reported to date and exceeds that reported by Anthopoulos and co‐workers35 by more than one order of magnitude. The slightly higher electron mobility measured for both methanofullerenes is not believed to be an intrinsic property of the two molecules but is most likely attributed to the existent of larger injection barrier for holes as compared to electrons.
In conclusion, the efficient solar cells based on methanofullerenes, namely PC60BM and PC70BM, as the sole light absorbing material and CuSCN as the transparent hole‐extracting material, have been demonstrated. Although, the bilayer CuSCN/PC70BM devices were found to exhibit moderate performance with PCE of ≈1%, physical blending of the two components resulted in the solar cells with PCE of 5.4% and VOC in excess of 0.9 V. Cells with PCE values of >6% were also fabricated, but proved to be difficult to reproduce reliably due to complex processing protocols employed. Analysis of the individual materials and device microstructures revealed that there are several factors that contribute to the high performance achieved, including: (i) the balanced and high mobility ambipolar charge transporting nature of the PC60BM and PC70BM derivatives, (ii) the superb hole‐transporting/electron‐blocking character of CuSCN, which helps to facilitate exciton dissociation and hole extraction at the critical CuSCN:fullerene interface, and (iii) the presence of a spontaneously formed mesostructured CuSCN‐nanowire:fullerene heterointerface—a truly hybrid organic/inorganic p–n interface unlike any other reported to date. First‐principle calculations confirm the beneficial nature of the CuSCN/fullerene interface for the charge separation. In fact, CuSCN not only acts as an efficient hole extracting layer but it also blocks the electron flow to the anode, resulting to large VOC values. The significance of the present work is threefold: first, it conclusively shows that PC60BM and PC70BM are excellent ambipolar semiconductors, with highly balanced electron and hole mobilities; second, the solar cell data represent the highest reported PCE value for an organic solar cell based on a single light absorbing material; and third, it lays the foundation for an improved understanding of the charge photogeneration in the fullerene‐based solar cells where the role of the fullerene has been largely ignored. The latter may also underpins the widely reported beneficial role that fullerene interlayers play in the operating characteristics of the metal halide perovskite solar cells.
Transistor Fabrication and Characterization: 2 cm × 2 cm glass substrates were used for TG‐BC transistor fabrication. Before material depositions, the cleaning procedure for the substrates was carried out by sonication in detergent solution (DECON90), deionized water, acetone, and 2‐propanol (IPA) for 10 min each prior to use. 40 nm thick gold source and drain contacts were then thermally evaporated onto the cleaned substrates through shadow masks, defining the transistor channel length of 30 µm and the channel width of 1000 µm. The metal contact was modified with pentafluorothiophenol (PFBT) by immersing the samples into a 1:100 v/v% diluted solution of PFBT in IPA for 1 h at room temperature in air. The excess of PFBT was rinsed off with large quantity of IPA after removing the substrates from the diluted PFBT solution. The semiconductor layers using PC70BM (99%, Solenne) or PC60BM (99%, Solenne) based solution (both in chlorobenzene at concentration of 20 mg mL−1) were then spin‐cast at spin speed of 1200 rpm for 30 s onto the glass substrates, followed by a thermal‐annealing process at 100 °C for 10 min. The semiconductor deposition was taken place in a nitrogen filled glove box. The soluble fluoropolymer CYTOP (Asahi Glass) was used as the gate dielectric layer and spin‐cast on top of the PC70BM layer or the PC60BM layer at 2000 rpm for 60 s, followed by a drying process in a high vacuum chamber (≈10−7 bar) for 12 h. The device fabrication was completed with thermal evaporation of 40 nm Al gate electrode through shadow masks onto the gate dielectric. The electrical characterization of the transistors was carried out at room temperature in a nitrogen‐filled glove box using an Agilent B2902A parameter analyzer.
The authors declare no conflict of interest.