Naphthodithiophene (NDT) is composed of one naphthalene ring and two thiophene rings. With the different combination, it could afford many kinds of geometric isomers . As an emerging donor building block, it has previously been suggested that NDT seems to be a very promising candidate unit for the design and construction of novel organic semiconductors, but the lack of practical methods for the effective synthesis of NDTs severely limits their further investigation . Until 2010, Takimiya and coworkers successfully established a practical synthetic route of NDT  and reported NDT-based organic semiconductors with extremely high field-effect mobilities . Since then many other organic semiconductors based on NDT unit were synthesized for high performance organic field-effect transistors (OFETs) . Even so, when compared with the widespread usage of benzodithiophene (BDT, one benzene ring and two thiophene rings) in both OFETs and organic solar cells (OSCs) , it can be found that the deep study of NDT about the application in photovoltaic still needs to be strengthened. Marks and co-workers reported the first application of NDT-based donor material for OSCs in 2011 , which had raised the curtain on the research of NDT in the organic solar cells. Later, a few more research groups began to focus on this respect, which had promoted the development of NDT-based photovoltaic materials. In this short review, we will summarize the relevant conclusions and advances focusing on the most common linear and zigzag NDT (Fig. 1) based small molecules and polymers reported by several excellent research groups.
2. NDT-based small molecules for organic solar cells
lNDT is the first type of NDT unit to be used for the design and construction of organic photovoltaic donor materials. In order to improve the solubility of the small molecule and ensure the film uniformity, 2-ethylhexyloxy groups were introduced into the lNDT core at the 5, 10 positions by Marks and co-workers in 2011 . Furthermore, to broaden the absorption spectrum and lower the optical bandgap, thiophene-capped diketopyrrolopyrrole (DPP) units were attached to the each side of lNDT. When donor material 1 (Fig. 2) was combined with the electron acceptor [6, 6]-phenyl- C61-butyric acid methyl ester (PC61BM), the power conversion efficiency (PCE) of 4.06% (Table 1) was achieved—a record for a PC61BM-based small-molecule OSCs at that time. Later, Lee and coworkers first introduced zNDT as the central building block into the small molecule system and reported a zNDT-based donor material 2 (Fig. 2) using triphenylamine capped benzothiadiazole (BT) as arms . Unlike lNDT type materials, there are no side chains attached on the central zNDT unit, so the solubility wasensured only by the long alkyl side chains of adjacent thiophene p bridge. Since the optical bandgap of small molecule 2 was very large and the charge carrier transport ability was very poor, the optimized PCE of the corresponding device was only 2.2% (Table 1). After that, they replaced the acceptor unit from BT to bithiazole, the PCE of the derived small molecule 3 (Fig. 2) actually fell to 1.62% (Table 1), mainly due to the increased HOMO energy level and optical bandgap leading to a much lower Voc and Jsc . In order to further improve the photovoltaic performance of the zNDT-based small molecules, Marks and co-workers first introduced the alkoxy groups to the zNDT unit at 4, 9-positions. Compared with the linear molecule 1, molecule 4 (Fig. 2) exhibited enhanced Jsc, FF and a slightly higher PCE of 4.7% (Table 1) mainly due to the greater hole mobility . Recently, Wei and co-workers reported two novel A-D-A type small molecules 5 and 6 (Fig. 2), based on the zNDT unit containing alkylthienyl or alkylphenyl as side chains, respectively . Compared with its counterpart using BDT as the donor unit (PCE = 5.26%) , small molecule 5 exhibited a slightly smaller PCE (4.71%), which was mainly due to the blue-shifted absorption spectrum. When changing the conjugated side chains from alkylthienyl to alkylphenyl, both charge transport property and film formation capability were improved, leading to the higher fill factor (FF, 0.71 vs. 0.69), higher current density (Jsc, 10.77 vs. 7.4 mA/cm2), and higher PCE (7.2% vs. 4.71%) even with much thicker active layers (Table 1). In particular, the optimal PCE for small molecule 6 was achieved with an active layer thickness beyond 300 nm, which is the best result reported for NDT-based small molecule solar cells up to date.
3. NDT-based polymers for organic solar cells
Research about NDT-based photovoltaic polymers also started with lNDT. In 2012, Peng and co-workers reported a donor- acceptor (D-A) copolymer 7 (Fig. 3) utilizing the lNDT as the donor block and electron-deficient DPP as the acceptor unit. Donor material 7 exhibited a much better photovoltaic performance with a PCE of 5.37% (Table 2) in the conventional device structure when compared with its counterpart which used BDT as the donor block (PCE = 2.91%), which benefited from the higher Voc and FF of the NDT-based polymer . And when utilizing the inverted device structure, both the Jsc and FF had been much increased, and consequently an impressively high PCE of 6.92% (Table 1) wasobtained. More encouragingly, as a low bandgap copolymer, donor material 7 also exhibited potential application in tandem OSCs, especially when combined with a large bandgap copolymer, and a high PCE up to 9.40% was achieved by fabricating a typical doublejunction tandem polymer solar cell (PSC) . Since lNDT possesses several bonding sites, so the substituent group can be introduced to another site other than 5, 10 positions. Lee and coworkers also developed a synthetic route to alkoxy-substituted lNDT in the same period, but unlike the polymer 7, two alkoxy side chains were attached to the lNDT core of donor material 8 (Fig. 3) in the 4, 9 positions . However, compared with polymer 7, polymer 8 did not show any significant improvement and the optimized PCE was only 4.0% (Table 2) even using [6, 6]-phenyl- C71-butyric acid methyl ester (PC71BM) as the electron acceptor, which should ascribe to the higher highest occupied molecular orbital (HOMO) energy level and the larger optical bandgap had lagged open circuit voltage (Voc 0.69 V) and Jsc (11.54 mA/cm2) of the corresponding device. In order to achieve a better performance, they further modified the polymer structure by replacing the acceptor unit from thieno[3, 4-c]pyrrole-1, 4-dione (TPD) to fluorine- substituted thieno[3, 4-b]thiophene (TT) . As a result, polymer 9 (Fig. 3) really displayed a higher PCE of 5.16% (Table 2) which was attributed to the increased Voc (0.72 V) and Jsc (13.50 mA/cm2).
Then, the focus of research about NDT-based polymer in OSCs began to shift from lNDT to zNDT. For example, besides above mentioned zNDT-based small molecule 2 and 3, Lee and coworkers also developed a series of D-A copolymers, such aspolymer 10  and 11 (Fig. 4) , based on the same zNDT unit but different acceptors. Unfortunately, these polymers did not afford much better device performances (around 1.6% for both of them) (Table 2). Using the same zNDT and BT units, Takimiya and co-workers reported a similar D-A copolymer (polymer 12, Fig. 4) in 2012 . Compared with polymer 11, the most obvious change is the introduction of branched alkyl side chains into the thiophene bridge. As we know, altering the type and length of the p bridge has played an important role in improving the chemical properties, stacking modes of the donor materials and film morphologies of the active layers. With these effects, the active layer of polymer 12 based PSC exhibited better film morphology and relatively ordered structure facilitating the charge carrier transport. Consequently, the PCE of the optimal device had significantly raised up to 3.8% (Table 2). By introducing a more rigid and electron deficient block: naphthobisthiadiazole (NT) as the acceptor unit, the derived polymer 13 (Fig. 4) showed lower HOMO energy level and larger absorption coefficient compared to its BT counterpart, resulting in increased Voc and Jsc. Moreover, polymer 13 provided more straight-shaped backbone and narrower π-π stacking distance than 12, which could also be the reasons for the high mobility and improved PCE of 4.9% (Table 2) .
|Figure 4. Chemical structures of zNDT (without side chains) based polymers 10-13.|
Since 2013, Lee and co-workers have tried to develop a series of D-A copolymers using the same alkoxyl substituted zNDT as donor block and various electron-deficient unit as the acceptor, such as compound 14  and 15  (Fig. 5). But unfortunately, both of them afforded a little poor PCEs (2.6% and 4.88%) (Table 2) less than those of their linear counterparts (4.0% for 8 and 5.16% for 9), probably due to the large optical bandgap resulting in lower Jsc. Li and co-workers also reported several similar D-A copolymers at that time, such as polymer 16  and 17  (Fig. 5). In contrast, they had gained much better results (5.26% for 16 and 5.07% for 17) (Table 2) by changing the type of alkyl side chains and the kind of acceptor unit. Among these zNDT type polymers, the most successful one should be polymer 18 (Fig. 5) reported by Cho and co-workers . The optimized conventional device showed a maximum PCE of 6.35% (Table 2), which should be attributed to much linear conformation of backbone, the faceon orientation of polymer 18, and small-scale phase separation of active layer.
|Figure 5. Chemical structures of zNDT (using alkoxy groups as side chains) based polymers 14-18|
In addition, alkyl side chains were also used to modify the central zNDT unit. Cheng and co-workers reported a D-A copolymer (polymer 19) (Fig. 6) which utilized the 4, 9-dialkylzNDT as the donor block and fluorine-substituted BT as the acceptor unit. Overall, each specific parameter of polymer 19 based photovoltaic device was almost the same as the polymer 18 . The difference was that a slightly higher PCE of 6.86% (Table 2) for polymer 19 was obtained with an inverted architecture. Then they introduced NT to the copolymer system and synthesized a new polymer 20 (Fig. 6) by co-operating with Takimiya and finally a further improvement of PCE (8.01%) (Table 2) was achieved , the trend of which was very similar to the description about polymer 12 and 13. Moreover, they also developed a facile synthetic strategy to synthesize a new type of NDT unit (β-aNDT) in 2013 . But unlike the extensive research in lNDT/zNDT based donor materials, β-aNDT was rarely used in OSCs. For the new D-A copolymer 21 (Fig. 6) utilizing the alkylated β-aNDT as the donor unit and NT as the acceptor unit , although the HOMO energy level and the optical bandgap were effectively reduced by the introduction of stronger acceptor (NT), the PCE of 21 based PSCs did not demonstrate a significant improvement (just 3.60%, Table 2). The results of the theoretical calculations have shown that the optimized backbone geometry of β-aNDT based D-A copolymer possesses a more wave-shaped backbone than zNDT based one, leading to weaker π-π stacking interaction and less ordered packing structure in the thin film, which may be the main reason accounting for the poor photovoltaic device performance. While Takimiya and co-workers also developed a different methodology for introducing various functional moieties at the 5, 10-positions of zNDT unit , which could derive a series of new type zNDT-based copolymers, such as polymer 22 (Fig. 6) . Surprisingly, introducing alkyl side chains to the 5, 10-positions brought significant change in molecule orientation into the face-on mode, which was very conducive to the charge carrier transport in the organic solar cells. An extremely high PCE up to 8.2% (Table 2) was achieved even with an active layer thickness of 300 nm, which have so far remained the best result for NDT-based polymer solar cells. All of these results indicated that change in molecular structure and choice of acceptor unit in D-A copolymer play vital roles in the optimizing the materials performances.
|Figure 6. Chemical structures of NDT3 and NDT4 (using alkyl groups as side chains) based polymers 19-22.|
Two dimensional (2D) conjugated concept , introducing conjugated side groups to the central donor block, has been widely used in the BDT system, which could exhibit red-shifted absorption spectra, lower HOMO energy levels, higher hole mobility, and greatly improved photovoltaic properties . Inspired by this, Wang and co-workers developed a 2D conjugated zNDT unit attached by two alkylphenyl groups also at 5, 10-positions in 2014. Compared with the widely used alkylthienyl side group, alkylphenyl group is utilized here mainly for obtaining a much higher Voc due to its weak electron-donating property. Although the new zNDT type donor polymer 23 (Fig. 7) only exhibited a moderate PCE of 4.11% (Table 2), it indeed injected new vitality to thedevelopment of NDT-based donor materials . At the almost same time, Wei and co-workers adopted the alkylthienyl group as the conjugated side chain and synthesized a novel D-A type copolymer (polymer 24, Fig. 7) based on the 2D zNDT unit. Impressively, the optimized device performance of 24-based polymer solar cells reached 7.50% (Table 2) at an active layer thickness of 200 nm, which makes it on to the list of only a few high efficient donor polymers with active layer thickness beyond 200 nm. Such a result suggests that 2D conjugated zNDT unit might be a very promising candidate building block for construction of novel highly efficient donor materials with thick active layers in fabricating large area, flexible polymer solar cells .
|Figure 7. Chemical structures of NDT3 (using alkylphenyl/alkylthienyl groups as side chains) based polymers 23 and 24|
In this short review, we have summarized the development process on the NDT-based donor materials in OSCs. Compared with many other types of donor blocks, such as thiophene, fluorine and BDT, the amount of reports about NDT-based donor materials is relatively small. When comparing various performance parameters of those reported NDT type donor materials, we can find that zNDT unit is becoming the focus of research. For example, the best PCE of the optimized zNDT-based small molecule (6) can be up to 7.2%, and the best result reported for zNDT polymer (22) solar cells reached 8.2%. In particular, these two results were both obtained with the optimized devices using ca. 300 nm thick active layers. Such an outstanding photovoltaic property indicates that zNDT unit has far reaching application potential in manufacturing highperformance, large-area organic solar cells by large-scale printing technology. Moreover, given the successful application of the 2D conjugated structure of the state-of-the-art material (PTB7-Th, using 2D-BDT as the donor block), we should be aware that there is still a wide gap between the photovoltaic performance of zNDT and that of BDT. However, there is no denying that zNDT unit possesses some unique attributes compared with BDT, such as lower HOMO energy level, higher charge carrier transport ability, and stronger intermolecular interactions. Although lower HOMO energy level may broaden the optical bandgap and stronger intermolecular interactions may lead to the excessive aggregation, it is believed that zNDT unit could be headed for an unprecedented break in the near future by selecting the appropriate molecular structure and device preparation process.Acknowledgments This work was supported by the National Natural Science Foundation of China (No. 21474022), the Beijing Municipal Science and Technology Commission (No. D151100003315002), and the Chinese Academy of Sciences.
|||S. Shinamura, I. Osaka, E. Miyazaki, et al. Linear-and angular-shaped naphthodithiophenes: selective synthesis, properties, and application to organic fieldeffect transistors. J. Am. Chem. Soc. 133 (2011) 5024–5035. DOI:10.1021/ja110973m|
|||S. Shinamura, E. Miyazaki, K. Takimiya. Synthesis, properties, crystal structures, and semiconductor characteristics of naphtho. J. Org. Chem. 75 (2010) 1228–1234. DOI:10.1021/jo902545a|
|||I. Osaka, T. Abe, S. Shinamura, E. Miyazaki, K. Takimiya. High-mobility semiconducting naphthodithiophene copolymers. J. Am. Chem. Soc. 132 (2010) 5000–5001. DOI:10.1021/ja101125p|
|||I. Osaka, T. Abe, S. Shinamura, K. Takimiya. Impact of isomeric structures on transistor performances in naphthodithiophene semiconducting polymers. J. Am. Chem. Soc. 133 (2011) 6852–6860. DOI:10.1021/ja201591a|
|||J.H. Hou, M.H. Park, S.Q. Zhang, et al. Bandgap and molecular energy level control of conjugated polymer photovoltaic materials based on benzo. Macromolecules 41 (2008) 6012–6018. DOI:10.1021/ma800820r|
|||. Loser, C.J. Bruns, H. Miyauchi, et al. A naphthodithiophene-diketopyrrolopyrrole donor molecule for efficient solution-processed solar cells. J. Am. Chem. Soc 133 (2011) 8142–8145. DOI:10.1021/ja202791n|
|||P. Dutta, W. Yang, S.H. Eom, et al. Development of naphtho. Chem. Commun. 48 (2012) 573–575. DOI:10.1039/C1CC15465F|
|||P. Dutta, W. Yang, W.H. Lee, I.N. Kang, S.H. Lee. Lee, Novel naphtho[1, 2-b:5, 6-b'] dithiophene core linear donor-π-acceptor conjugated small molecules with thiophene-bridged bithiazole acceptor: design, synthesis, and their application in bulk heterojunction organic solar cells. J. Mater. Chem. 22 (2012) 10840–10851. DOI:10.1039/c2jm30934c|
|||S. Loser, H. Miyauchi, J.W. Hennek, et al. A "zig-zag" naphthodithiophene core for increased efficiency in solution-processed small molecule solar cells. Chem. Commun 48 (2011) 8511–8513.|
|||X.W. Zhu, B.Z. Xia, K. Lu, et al. Naphtho. Chem. Mater. 28 (2016) 943–950. DOI:10.1021/acs.chemmater.5b04668|
|||D. Deng, Y.J. Zhang, L. Yuan, et al. Effects of shortened alkyl chains on solutionprocessable small molecules with oxo-alkylated nitrile end-capped acceptors for high-performance organic solar cells. Adv. Energy Mater. 4 (2014) 1400538. DOI:10.1002/aenm.201400538|
|||Q. Peng, Q. Huang, X.B. Hou, et al. Enhanced solar cell performance by replacing benzodithiophene with naphthodithiophene in diketopyrrolopyrrole-based copolymers. Chem. Commun. 48 (2012) 11452–11454. DOI:10.1039/c2cc36324k|
|||K. Li, Z.J. Li, K. Feng, et al. Development of large band-gap conjugated copolymers for efficient regular single and tandem organic solar cells. J. Am. Chem. Soc. 135 (2013) 13549–13557. DOI:10.1021/ja406220a|
|||S. Sanjaykumar, C.E. Song, W.S. Shin, et al. Synthesis and characterization of a novel naphthodithiophene-based copolymer for use in polymer solar cells. Macromolecules 45 (2012) 6938–6945. DOI:10.1021/ma301312d|
|||R.S. Koti, S.R. Sanjaykumar, S.J. Hong, et al. 3, 8-Dialkoxynaphthodithiophene based copolymers for efficient polymer solar cell. Sol. Energy Mater. Sol. Cells 108 (2013) 213–222. DOI:10.1016/j.solmat.2012.09.021|
|||P. Dutta, H. Park, W.H. Lee, et al. Synthesis characterization and bulk-heterojunction photovoltaic applications of new naphtho. Polym. Chem. 5 (2014) 132–143. DOI:10.1039/C3PY00911D|
|||P. Dutta, H. Park, M. Oh, et al. Modulation of electronic properties of π-conjugated copolymers derived from naphtho. J. Polym. Sci. Part A: Polym. Chem. 51 (2013) 2948–2958. DOI:10.1002/pola.26691|
|||I. Osaka, T. Abe, M. Shimawaki, T. Koganezawa, K. Takimiya. Naphthodithiophenebased donor-acceptor polymers: versatile semiconductors for OFETs and OPVs. ACS Macro Lett. 1 (2012) 437–440. DOI:10.1021/mz300065t|
|||C. Bathula, S. Badgujar, C.E. Song, et al. Effect of backbone structures on photovoltaic properties in naphthodithiophene-based copolymers. J. Polym. Sci. Part A: Polym. Chem. 52 (2014) 305–312. DOI:10.1002/pola.27005|
|||C. Bathula, C.E. Song, S. Badgujar, et al. Naphtho. Polym. Chem. 4 (2013) 2132–2139. DOI:10.1039/c3py21062f|
|||S.W. Shi, P. Jiang, S.Q. Yu, et al. Efficient polymer solar cells based on a broad bandgap D-A copolymer of "zigzag" naphthodithiophene and thieno. J. Mater. Chem. A 1 (2013) 1540–1543. DOI:10.1039/C2TA01143C|
|||S.W. Shi, X.D. Xie, P. Jiang, et al. Naphtho. Macromolecules 46 (2013) 3358–3366. DOI:10.1021/ma400177w|
|||J. Lee, H. Ko, E. Song, et al. Naphthodithiophene-based conjugated polymer with linear, planar backbone conformation and strong intermolecular packing for efficient organic solar cells. ACS Appl. Mater. Interfaces 7 (2015) 21159–21169. DOI:10.1021/acsami.5b04884|
|||S.W. Cheng, C.E. Tsai, W.W. Liang, et al. Angular-shaped 4, 9-dialkylnaphthodithiophene-based donor-acceptor copolymers for efficient polymer solar cells and high-mobility field-effect transistors. Macromolecules 48 (2015) 2030–2038. DOI:10.1021/acs.macromol.5b00098|
|||S.W. Cheng, D.Y. Chiou, C.E. Tsai, et al. Angular-shaped 4, 9-dialkyl a-and bnaphthodithiophene-based donor-acceptor copolymers: investigation of isomeric structural effects on molecular properties and performance of field-effect transistors and photovoltaics. Adv. Funct. Mater. 25 (2015) 6131–6143. DOI:10.1002/adfm.v25.38|
|||S.W. Cheng, D.Y. Chiou, Y.Y. Lai, et al. Synthesis and molecular properties of four isomeric dialkylated angular-shaped naphthodithiophenes. Org. Lett. 15 (2013) 5338–5341. DOI:10.1021/ol4025953|
|||S. Shinamura, R. Sugimoto, N. Yanai, et al. Orthogonally functionalized naphthodithiophenes: selective protection and borylation. Org. Lett. 14 (2012) 4718–4721. DOI:10.1021/ol301797g|
|||I. Osaka, T. Kakara, N. Takemura, T. Koganezawa, K. Takimiya. Naphthodithiophene-naphthobisthiadiazole copolymers for solar cells: alkylation drives the polymer backbone flat and promotes efficiency. J. Am. Chem. Soc. 135 (2013) 8834–8837. DOI:10.1021/ja404064m|
|||L.J. Huo, J.H. Hou, S.Q. Zhang, H.Y. Chen, Y. Yang. A polybenzo[1, 2-b:4, 5-b'] dithiophene derivative with deep HOMO level and its application in highperformance polymer solar cells. Angew. Chem. Int. Ed 49 (2010) 1500–1503. DOI:10.1002/anie.200906934|
|||L.J. Huo, S.Q. Zhang, X. Guo, et al. Replacing alkoxy groups with alkylthienyl groups: a feasible approach to improve the properties of photovoltaic polymers. Angew. Chem. Int. Ed. 50 (2011) 9697–9702. DOI:10.1002/anie.201103313|
|||S.W. Shi, K.L. Shi, R. Qu, et al. Alkylphenyl substituted naphthodithiophene: a new building unit with conjugated side chains for semiconducting materials. Macromol. Rapid Commun. 35 (2014) 1886–1889.|
|||X.W. Zhu, J. Fang, K. Lu, et al. Naphtho. Chem. Mater. 26 (2014) 6947–6954. DOI:10.1021/cm5033223|