Tetrathiafulvalene Hybridized with Indacenetetraone ...

Tetrathiafulvalene Hybridized with Indacenetetraone as Visible-light-harvesting Electron Acceptor Applicable to Bulk-heterojunction Organic Photovoltaics Kouki Akaike,1 Hideo Enozawa,*1 Takashi Kajitani,1 Mari Koizumi,1 Atsuko Kosaka,1 Daisuke Hashizume,1 Yoshiko Koizumi,1,2 Akinori Saeki,2 Shu Seki,2 and Takanori Fukushima*1,3 1 RIKEN Advanced Science Institute, 2-1 Hirosawa, Wako, Saitama 351-0198 2 Department of Applied Chemistry, Osaka University, 2-1 Yamadaoka, Suita, Osaka 565-0013 3 Chemical Resources Laboratory, Tokyo Institute of Technology, R1-1 4259 Nagatsuta, Midori-ku, Yokohama, Kanagawa 226-8503 (Received July 29, 2013; CL-130702; E-mail: [email protected])

REPRINTED FROM

Vol.42 No.11

2013 p.1417–1419 CMLTAG November 5, 2013

The Chemical Society of Japan Published on the web August 24, 2013; doi:10.1246/cl.130702

doi:10.1246/cl.130702 Published on the web August 24, 2013

1417

Tetrathiafulvalene Hybridized with Indacenetetraone as Visible-light-harvesting Electron Acceptor Applicable to Bulk-heterojunction Organic Photovoltaics Kouki Akaike,1,³ Hideo Enozawa,*1,³ Takashi Kajitani,1 Mari Koizumi,1 Atsuko Kosaka,1 Daisuke Hashizume,1,³ Yoshiko Koizumi,1,2 Akinori Saeki,2 Shu Seki,2 and Takanori Fukushima*1,3 1 RIKEN Advanced Science Institute, 2-1 Hirosawa, Wako, Saitama 351-0198 2 Department of Applied Chemistry, Osaka University, 2-1 Yamadaoka, Suita, Osaka 565-0013 3 Chemical Resources Laboratory, Tokyo Institute of Technology, R1-1 4259 Nagatsuta, Midori-ku, Yokohama, Kanagawa 226-8503 (Received July 29, 2013; CL-130702; E-mail: [email protected]) π-Extension of tetrathiafulvalene by hybridization with an indacenetetraone leads to an electron acceptor that can harvest visible light efficiently and serve as an electron-transport material for poly(3-hexylthiophene)-based bulk-heterojunction photovoltaics, where brief thermal annealing of the active layer results in better carrier transport pathways to enhance the device performance. Tetrathiafulvalene (TTF) is a typical electron donor, which forms charge-transfer (CT) complexes with a wide range of acceptors.1a Modifications of the TTF skeleton have been useful for the development of metallic and superconducting organic materials.1b Furthermore, a family of TTFs have recently been thought to serve as components in organic electronics,2a for which proper tuning of electronic properties depending on the type of device is essential. To this end, π-extension of TTFs by incorporating electron-accepting unit into the central double bond is promising,2b because the resulting donor/acceptor hybrid molecules are expected to adopt versatile electronic structures via intramolecular interaction between spatially separated TTF and acceptor units. Herein, we report that the hybridization of TTF with an electron-accepting indacenetetraone unit (1)3 leads to a good electron acceptor 3a (Scheme 1), which displays a large absorption coefficient due to the push pull structure in π-conjugation, as often observed for dye molecules.4 By virtue of its electronic and absorption properties, the hybrid molecule, when combined with poly(3-hexylthiophene) (P3HT) as an electron donor, provided an active layer for organic photovoltaic (OPV) devices.5 Indacenetetraone3 bears a highly electron-deficient aromatic core arising from four carbonyl substituents. Considering the nature of π-stacking interaction6 such arene compounds could exhibit a strong π-stacking to form an ordered columnar assembly. To construct the hybrid structure of 3, we initially attempted to prepare 3c by the condensation of 1 and 2c (Scheme 1). However, the reaction gave a solid substance, which was insoluble in any organic solvents. In order to improve solubility, we employed a methoxycarbonyl group-appended 1,3-dithiole-2-thione (2b) as a counterpart of 1, but again insoluble solid resulted. Finally, we found that 3a bearing propoxycarbonyl groups can be obtained in an excellent yield (91%) by the condensation of 1 and the corresponding 1,3dithiole-2-thione (2a) in the presence of Et3N and AgNO3 (Scheme 1).7 Fortunately, single crystals of 3a suitable for X-ray crystallography were successfully obtained by recrystallization from a CH2Cl2/hexane mixture. As illustrated in Chem. Lett. 2013, 42, 14171419

Scheme 1. Synthesis of a TTF/indacenetetraone hybrid molecule 3.

Figure 1. Crystal structure of 3a at 90 K. (a) ORTEP drawing of molecular structure of 3a showing 50% probability thermal ellipsoids. (b) Crystal-packing diagrams of 3a viewed along b axis. (c) HOMO and LUMO of 3c calculated at the B3LYP/6-31+G(d,p)//B3LYP/ 6-31G(d) level. (d) Electronic absorption spectra of 3a in CH2Cl2 at 25 °C.

Figure 1a, the π-framework of 3a adopts a planar geometry, where short intramolecular S£O atomic contacts less than the sum of van der Waals radii (3.32 ¡) were observed.8 The interaction between sulfur and oxygen atoms most likely operates as an attractive force, so that the 1,3-dithiole and

© 2013 The Chemical Society of Japan

www.csj.jp/journals/chem-lett/

Chem. Lett. 2013, 42, 14171419

Abs./arb unit

indacenetetraone units can adopt a coplanar conformation. In the crystal, 3a molecules π-stack on top of each other with a planeto-plane separation of 3.49 ¡ (Figure 1b). As a result, onedimensional columns are formed along the c axis (Figure 1b), which would provide pathways for charge transport. Interestingly, although TTF is a strong electron donor,9 its indacenetetraone hybrid 3a exhibited a good electron-accepting property. Thus, cyclic voltammetry (CV) of 3a showed a quasireversible one-electron reduction peak at ¹1.42 V (vs. Fc/Fc+, Figure S1),7 while electrochemical oxidation did not take place on anodic sweep up to +0.80 V. In electronic absorption spectroscopy, 3a displayed an intense absorption band (¾ = 156000 M¹1 cm¹1) at = 454 nm (Figure 1d), owing to its pushpull nature in π-conjugation.4 Time-dependent (TD) density functional theory (DFT) calculations of the parent molecule (3c, Scheme 1) at the B3LYP/6-31+G(d,p)//B3LYP/ 6-31G(d) level suggested that both the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) spread entirely over the π-framework (Figure 1c), and the HOMOLUMO transition is mainly responsible for the strong absorption (Figures S2S3a and Tables S1S3).7 According to the TD DFT calculations, the transition dipole moment along the longer axis of 3c is as large as 9.37 D (Figure S3b and Table S4).7 Considering that electronic transition probability is proportional to the square of transition dipole moment,10 the strong absorption of 3a is reasonable. Importantly, the absorption band of 3a has an effective overlap with solar spectrum in the range of 400 500 nm. Furthermore, based on the optical band gap (2.7 eV) estimated from the absorption maximum (Figure 1d), together with the reduction potential of 3a (Figure S1), the HOMO and LUMO energies of 3a were determined as 6.1 and 3.4 eV, respectively (Figure 2a). Hence, 3a is expected to serve as an electron acceptor for bulk-heterojunction (BHJ) OPVs with various electron donors including poly(3-hexylthiophene) (P3HT). With the above observations in mind, we fabricated an OPV device using a blend film of P3HT:3a (1:2, w/w)13 as an active layer with a structure of [ITO/PEDOT:PSS/P3HT:3a/Ca/Al] (Figure 2a) and found that the device indeed displays a photovoltaic response. Figure 2b shows the current density voltage (JV) curve of the P3HT:3a-based OPV device (blue), where short-circuit current density (Jsc), open-circuit voltage (Voc), fill factor (FF), and power-conversion efficiency (©) were determined as 0.51 mA cm¹2, 0.87 V, 0.39, and 0.18%, respectively (Table 1). Compared with Voc values reported for typical P3HT:[6,6]-phenyl-C61-butyric acid methyl ester (PC61BM) OPVs (ca. 0.6 V),14 the observed Voc is notably high. This is consistent with the fact that the LUMO level of 3a is 0.26 eV higher than that of PC61BM (Figure S1).7 We noticed that the device performance is greatly improved, when the active layer is thermally annealed at 150 °C for 10 min before attaching the cathode (Table 1). The best device showed Jsc and FF values of 1.67 mA cm¹2 and 0.55, respectively, while Voc value (0.90 V) remained unchanged (Figure 2b, red). As a consequence, a much higher efficiency (© = 0.83%) was achieved (Table 1). Figure 2c shows the external quantum efficiency (EQE) of the devices, together with the absorption spectrum of an annealed P3HT:3a (1:2 w/w) film. Again, thermal annealing resulted in the increase of EQE values in the entire wavelength region, even

Intensity/arb unit

1418

Figure 2. (a) Schematic illustration of a P3HT:3a-based OPV device and energy levels of the components [indium tin oxide (ITO); poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS)]. HOMO and LUMO energy levels of P3HT, work functions of PEDOT:PSS/ITO anode and Ca/Al cathode were taken from refs 11d and 12. (b) JV curves of devices fabricated with as-prepared (blue squares: illumination under AM1.5 irradiation at 100 mA cm¹2, blue broken curve: dark) and annealed P3HT:3a blend films (red circles: illumination, red broken curve: dark). (c) EQE spectra of P3HT:3abased OPVs fabricated with as-prepared (blue) and annealed (red) P3HT:3a blend films, along with the absorption spectrum of an annealed P3HT:3a blend film (black). (d) Powder XRD patterns of an annealed spin-coated film of P3HT (purple) on PEDOT:PSS/ITO/ glass, a spin-coated film of 3a (green) on ITO/glass, and as-prepared (blue) and annealed (red) P3HT:3a (1:2, w/w) blend films on PEDOT:PSS/ITO/glass. Peak assignments are summarized in Table S5. An asterisk denotes reflection from ITO. Table 1. Summary of OPV performance for 47 P3HT:3a devices fabricated with as-prepared (spin-coated) and annealed active layersa

Active layer As-prepared Annealed (150 °C 10 min) a

Jsc/mA cm¹2 0.51 (0.51) 1.44 (1.67)

Voc/V 0.87 (0.89) 0.90 (0.90)

FF 0.39 (0.42) 0.52 (0.55)

©/% 0.18 (0.19) 0.68 (0.83)

Values in parentheses are those observed for the best device.

at > 500 nm, where the absorption of P3HT is predominant (Figure S4a).7 Noteworthy, the EQE spectra display the characteristic absorption band of 3a ( = 454 nm, Figure 1d), indicating that light absorption by 3a largely contributes to the photocurrent generation. This is in contrast to the case of typical C60-based OPVs, where absorption by fullerene in visible-light region is poorly responsible for photocurrent generation. We confirmed that the spectral features of a P3HT:3a blend film are changed before and after annealing. For instance, tailing of the absorption spectrum of the as-prepared film, caused by charge-transfer between P3HT and 3a, disappeared upon



1419 annealing (see inset in Figure S4a).7 The absorption bands at 556 and 606 nm, assignable to the vibration coupling of P3HT,14a,15 emerged upon annealing (Figure 2c, black). On the other hand, the spectral feature below = 500 nm remained virtually unchanged (Figure S4a).7 In fluorescence spectroscopy, although fluorescence from P3HT in the as-prepared film was mostly quenched (Figure S4b, blue),7 it turned partially fluorescent after annealing (Figure S4b, red).7,16 These observations suggest that while P3HT and 3a in the as-prepared film are associated with one another due to weak charge-transfer interaction, thermal annealing leads to the segregation of the donor and acceptor components. In fact, powder X-ray diffractometry (XRD) of a P3HT:3a (1:2, w/w) blend film, prepared on a PEDOT:PSS-covered ITO substrate, revealed the occurrence of nanoscopic segregation of P3HT and 3a. As shown in Figure 2d, the blend film before annealing only showed a diffraction peak with a d-spacing of 14.2 ¡ (Figure 2d, blue), presumably due to P3HT associated with 3a at the molecular level. In contrast, XRD pattern of the annealed film displayed three distinct peaks with d-spacings of 16.7 (A), 13.3 (B), and 8.0 ¡ (C) (Table S5),7 which agree with a superposition of those observed for spin-coated films of P3HT (Figure 2d, purple) and 3a (Figure 2d, green). Because the XRD peaks (Figure 2d, red and green) are represented by the simulated powder pattern of single crystalline 3a (Figure S5),7 the molecular packing arrangement of 3a in the annealed film would be analogous to that in the single crystal. The crystallite sizes of P3HT and 3a domains were calculated as 10.2 and 22.4 nm,17 respectively, which are in the range of an excitondiffusion limit.5 Such a nanoscopic segregation of donor and acceptor components is considered essential for BHJ OPV5 to acquire separate hole- and electron-transport pathways in the active layer for improving Jsc and FF. To investigate the carrier-transport properties of P3HT and 3a in the blend film, we performed time-of-flight measurements on a P3HT:3a (1:2, w/w) film sandwiched by ITO and Al electrodes (Figures S6 and S7).7 The as-prepared film exhibited a hole mobility (®h) of 2.9 © 10¹2 cm2 V¹1 s¹1 and an electron mobility (®e) of 3.5 © 10¹3 cm2 V¹1 s¹1, which in turn displayed ®h = 5.0 © 10¹3 and ®e = 9.6 © 10¹3 cm2 V¹1 s¹1 upon annealing.18 Importantly, the ratio of ®h/®e changed from 8.3 to 0.52 before and after annealing. It is known that, when ®h and ®e are balanced, photocurrent is not limited by the space-charge effects,20 and unfavorable charge recombination in the active layer can be suppressed to enhance FF.14a,21 Therefore, we now conclude that thermal annealing provides the P3HT:3a film with increased ®e, resulting in better balanced carrier mobilities, whereby both Jsc and FF are improved to enhance overall performance of the OPV device. Nevertheless, Jsc observed for the P3HT:3a OPV device is still limited. This is presumably due to insufficient carrier generation at P3HT/3a interface. In summary, we demonstrated that TTF/indacenetetraone hybrid 3a, featuring extended pushpull-type π-conjugation, can serve as an electron acceptor for BHJ OPVs, which possesses (1) adequate HOMO and LUMO energies for photoinduced electron transfer with P3HT, (2) a strong absorption in visible-light region, and (3) a good electron-transport property. Also noteworthy is the fact that 3a is capable of nanoscopic segregation from P3HT in their blend film, rather than CT complexation, so that it can acquire pathways for electron transport. Although the Chem. Lett. 2013, 42, 14171419

performance of the prototype OPV device is not satisfactory at present, in terms of power-conversion efficiency,11 we expect that it could be improved by structural modifications of 3a, including the choice of different types of substituents on the 1,3dithiole ring. Further studies along this line are in progress. This work was supported by RIKEN Incentive Research Grant (K.A. and H.E.), KAKENHI (Nos. 24750188 for K.A. and 24350055 for T.F.), and a grant from the Nagase Science and Technology Foundation (T.F.). We thank Dr. K. Tajima (RIKEN) for EQE measurements. References and Notes ³ Present address: RIKEN Center for Emergent Matter Science, 2-1 Hirosawa, Wako, Saitama 351-0198 1 a) TTF Chemistry: Fundamentals and Applications of Tetrathiafulvalene, ed. by J.-i. Yamada, T. Sugimoto, Kodansha & Springer, Tokyo, 2004. b) Special issue on Molecular Conductors ed. by P. Batail: Chem. Rev. 2004, 104, Issue 11. 2 a) F. G. Brunetti, J. L. López, C. Atienza, N. Martín, J. Mater. Chem. 2012, 22, 4188. b) Y. Morioka, J.-i. Nishida, E. Fujiwara, H. Tada, Y. Yamashita, Chem. Lett. 2004, 33, 1632. 3 a) H. Esener, T. Uyar, Dyes Pigm. 2007, 72, 109. b) C. Niebel, V. Lokshin, V. Khodorkovsky, Tetrahedron Lett. 2008, 49, 7276. 4 J. Nakayama, M. Ishihara, M. Hoshino, Chem. Lett. 1977, 77. 5 a) K. M. Coakley, M. D. McGehee, Chem. Mater. 2004, 16, 4533. b) B. Kippelen, J.-L. Brédas, Energy Environ. Sci. 2009, 2, 251. c) A. Mishra, P. Bäuerle, Angew. Chem., Int. Ed. 2012, 51, 2020. 6 a) C. A. Hunter, J. K. M. Sanders, J. Am. Chem. Soc. 1990, 112, 5525. b) F. Cozzi, M. Cinquini, R. Annunziata, T. Dwyer, J. S. Siegel, J. Am. Chem. Soc. 1992, 114, 5729. c) F. Cozzi, M. Cinquini, R. Annuziata, J. S. Siegel, J. Am. Chem. Soc. 1993, 115, 5330. 7 Supporting Information is available electronically on the CSJ-Journal Web site, http://www.csj.jp/journals/chem-lett/index.html. 8 M. Iwaoka, N. Isozumi, Molecules 2012, 17, 7266. 9 J.-i. Nishida, D. Kumaki, S. Tokito, Y. Yamashita, J. Am. Chem. Soc. 2006, 128, 9598. 10 H. Bürckstümmer, E. V. Tulyakova, M. Deppisch, M. R. Lenze, N. M. Kronenberg, M. Gsänger, M. Stolte, K. Meerholz, F. Würthner, Angew. Chem., Int. Ed. 2011, 50, 11628. 11 a) F. G. Brunetti, X. Gong, M. Tong, A. J. Heeger, F. Wudl, Angew. Chem., Int. Ed. 2010, 49, 532. b) P. Sonar, J. P. F. Lim, K. L. Chan, Energy Environ. Sci. 2011, 4, 1558. c) J. E. Anthony, Chem. Mater. 2011, 23, 583. d) E. Zhou, J. Cong, Q. Wei, K. Tajima, C. Yang, K. Hashimoto, Angew. Chem., Int. Ed. 2011, 50, 2799. 12 L.-M. Chen, Z. Xu, Z. Hong, Y. Yang, J. Mater. Chem. 2010, 20, 2575. 13 The ratio of P3HT/3a = 1/2 was found to be optimal in terms of overall device performance. 14 a) G. Li, V. Shrotriya, J. Huang, Y. Yao, T. Moriarty, K. Emery, Y. Yang, Nat. Mater. 2005, 4, 864. b) T. Erb, U. Zhokhavets, G. Gobsch, S. Raleva, B. Stühn, P. Schilinsky, C. Waldauf, C. J. Brabec, Adv. Funct. Mater. 2005, 15, 1193. 15 M. Sundberg, O. Inganäs, S. Stafström, G. Gustafsson, B. Sjögren, Solid State Commun. 1989, 71, 435. 16 Even in the annealed film, 95% of fluorescence from P3HT was still quenched (Figure S4b).7 17 The crystallite sizes (²) were obtained by ² = 0.89/½1/2 cos ª, where is the wavelength of incident X-ray beam (1.5415 ¡), ½1/2 is full width at half-maximum of reflection, and diffraction angle ª is the maximum of reflection, see: P. H. J. Kouwer, W. F. Jager, W. J. Mijs, S. J. Picken, Macromolecules 2002, 35, 4322. 18 As for ®h, as-prepared film showed a higher value than the annealed one. Although the origin is not clear, similar behavior has been reported previously.19 19 A. C. Mayer, M. F. Toney, S. R. Scully, J. Rivnay, C. J. Brabec, M. Scharber, M. Koppe, M. Heeney, I. McCulloch, M. D. McGehee, Adv. Funct. Mater. 2009, 19, 1173. 20 C. Melzer, E. J. Koop, V. D. Mihailetchi, P. W. M. Blom, Adv. Funct. Mater. 2004, 14, 865. 21 A. M. Goodman, A. Rose, J. Appl. Phys. 1971, 42, 2823.

