Introduction

Conjugated polymers are important materials for next-generation electronics, such as organic field-effect transistors (OFETs), organic photovoltaics (OPVs), non-volatile organic memory, and so on [1,2,3,4]. In general, conjugated polymers are synthesized by cross-coupling polymerization between a dihaloarene monomer and an organometallic arene monomer, such as Suzuki and Stille cross-coupling polymerizations (Fig. 1) [5,6,7]. Cross-coupling polymerization has clearly played an essential role in the development of conjugated polymers. However, these techniques have drawbacks, such as the requirement for the prior preparation of monomers with organometallic functionalities, the toxicity of the byproducts, the low atom economy, and so on. A new methodology for the synthesis of conjugated polymers that can overcome such drawbacks is greatly needed in terms of green chemistry.

Fig. 1
figure 1

General scheme of conventional cross-coupling polymerization and direct arylation polymerization (DArP)

Direct arylation polymerization (DArP), cross-coupling polymerization between a dihaloarene monomer and a non-substituted arene monomer, is of current interest as a promising alternative methodology [8,9,10,11,12]. The synthetic advantage of DArP is that no preparation of monomers with organometallic functionality is required, which contributes to higher atom economy. Furthermore, the lower toxicity of the byproducts makes DArP more environment-friendly than conventional cross-coupling polymerizations. Since the first report of DArP, in which poly(3-alkylthiophene)s were synthesized from 2-iodo-3-alkylthiophenes [13], a variety of conjugated polymers have been synthesized by DArP [14,15,16,17,18,19,20,21,22,23,24]. Nevertheless, there are still many conjugated polymers that cannot be synthesized by DArP. This fact implies that DArP may currently be less universally applicable than conventional cross-coupling polymerizations, and more research is essential to expand its potential. This focus review describes our recent studies on the development of novel conjugated polymer synthesis by DArP, which emphasizes the potential of DArP as an alternative methodology for the synthesis of conjugated polymers.

Mechanism of direct arylation

Figure 2 shows the catalytic cycle of direct arylation under carboxylate-mediated conditions, which is commonly used for the direct arylation synthesis of conjugated polymers [25]. In this cycle, the oxidative addition of palladium to a haloarene monomer occurs as well as conventional cross-coupling reactions. With assistance from the carboxylate ligand, the unsubstituted monomer is deprotonated simultaneously with the formation of the metal–carbon bond. Reductive elimination then occurs to afford the final product. The combination of reagents (i.e., palladium catalyst, phosphine ligand, acid, base, and solvent) dominates the progress of the reaction, and exploring suitable combinations of these reagents is a fundamental approach for DArP [26,27,28].

Fig. 2
figure 2

Catalytic cycle of direct arylation in carboxylate-mediated conditions

Donor–acceptor (D–A) alternating copolymers from AA + BB system

D–A alternating copolymers (i.e., low-band-gap polymers) are a promising material for many device applications because the tuning of the polymer characteristics, such as optical properties, HOMO/LUMO levels, thermal properties, and crystallinity, can be accomplished by the combination of donor and acceptor structures [29,30,31]. Alternating copolymerization of the bifunctional donor and acceptor monomers (AA + BB system) enables the systematic synthesis of a series of low-band-gap polymers. To date, various low-band-gap polymers have been synthesized by DArP in an AA + BB system (Fig. 3) [16, 17, 20, 22,23,24].

Fig. 3
figure 3

Synthetic approach to obtain conjugated polymers by DArP

The arylene diimide (naphthalene diimide, perylene diimide, and so on) structure is a potential acceptor component because of its high electron affinity, high electron mobility, and high stability [32]. In particular, naphthalene diimide (NDI)-based low-band-gap polymers have been widely investigated as n-type semiconducting materials for OFETs [33,34,35]. In addition to OFET applications, the potential of NDI-based low-band-gap polymers as acceptor materials is also demonstrated by their use in non-fullerene solar cells (i.e., all-polymer solar cells) [36, 37]. This utility indicates that NDI-based low-band-gap polymers are high-performance acceptor polymers. Accordingly, there has been great scientific interest in and high demand for the synthesis of such high-performance acceptor polymers via DArP; however, the synthesis of arylene diimide-based low-band-gap polymers via DArP had never been accomplished.

The direct arylation synthesis of NDI-based low-band-gap polymers was firstly reported in 2012 by Horie et al. [38]. This direct arylation was carried out between a conventional dibromo NDI monomer (N,N′-di-n-hexadecyl-2,6-dibromonaphthalene-1,4,5,8-tetracarboxylic acid diimide) and 4,4′-bis(2-ethylhexyl)dithieno[3,2-b:2′,3′-d]silole in the presence of Pd(OAc)2, which is commonly used as a palladium catalyst in DArP. This reaction gave the corresponding polymer with a very low molecular weight (Mn = 2200). Due to the low reactivity of the dibromo NDI monomer, this DArP was deemed unsuccessful. Based on this result, we focused on the design of a new dibromo NDI monomer with enhanced reactivity in DArP. As explained earlier, the catalytic cycle of DArP passes through a transition state in which palladium coordinates with the carbonyl group of the carboxylate ligand. In the case of DArP using a conventional dibromo NDI monomer (e.g., N,N′-di-n-hexadecyl-2,6-dibromonaphthalene-1,4,5,8-tetracarboxylic acid diimide), the carbonyl group of the dibromo NDI monomer can coordinate to palladium in addition to the carboxylate ligand (Fig. 4). A very stable bidentate coordination would prevent the C–H bond cleavage. The formation of bidentate coordination is considered to be one of the reasons that DArP using a conventional dibromo NDI monomer did not proceed well. Accordingly, the design of a new dibromo NDI monomer that does not form the aforementioned bidentate coordination could be the key to the successful direct arylation synthesis of NDI-based low-band-gap polymers. Therefore, we synthesized a new dibromo NDI monomer, bromothiophene-end-capped NDI monomer (NDI2T-Br2) [39], and investigated DArP between NDI2T-Br2 and 3,4-dimethylthiophene under typical Pd(OAc)2-based direct arylation conditions (Fig. 4a) [40, 41]. As shown in Table 1, the progress of DArP using NDI2T-Br2 was significantly improved compared to the previous report, and the NDI-terthiophene alternating low-band-gap polymer PNDI3T with Mn = 31,000 was successfully obtained by using PtBu3·HBF4 as a result of screening phosphine ligands [41]. For comparison, DArP between N,N′-bis(2-decyl-1-tetradecyl)-2,6-dibromonaphthalene-1,4,5,8-tetracarboxylic acid diimide and 3,4-dimethylthiophene was conducted under the same conditions; however, in this case, the polymer was not obtained. The optical, electrochemical, and thermal properties of the resulting PNDI3T were consistent with those of the corresponding polymer synthesized by Stille cross-coupling polymerization. To the best of our knowledge, DArP using NDI2T-Br2 is the first example of the direct arylation synthesis of an NDI-based low-band-gap polymer with high molecular weight. Furthermore, the correlation between these results as a function of the phosphine ligand provided additional insight (Fig. 4b). The reactivity of the catalyst strongly depends on the cone angle and basicity of the ligand in the cross-coupling reaction [42]. In general, a large cone angle and strong basicity promote reductive elimination and oxidative addition, respectively. In DArP using NDI2T-Br2, the basicity tended to dominate the progress of the DArP rather than the cone angle. This fact indicates that the promotion of the oxidative addition process by a ligand with strong basicity (strong electron-donation) is highly important for DArP using NDI2T-Br2. Following our first report on DArP using NDI2T-Br2, other research groups have developed NDI-based low-band-gap polymers by this approach, thus showing its general versatility [43,44,45]. It should be noted, however, that Sommer et al. developed the direct arylation synthesis of NDI-based low-band-gap polymers from a conventional dibromo NDI monomer while exploring the Pd2dba3-based catalytic system [46].

Fig. 4
figure 4

a Direct arylation synthesis of PNDI3T from NDI2T-Br2. b Two types of transition state expected from each NDI monomer. c Correlation between the molecular weight and phosphine ligand in PNDI3T. JP-Se is the coupling constant of the corresponding phosphine selenides Se = PR3 (smaller values correspond to higher basicity). The cone angle (deg) of the phosphine ligand is given in parentheses

Table 1 Direct arylation synthesis of PNDI3Ta

In addition, perylene diimide (PDI)-based low-band-gap polymers were successfully synthesized by using a bromothiophene-end-capped PDI monomer (PDI2T-Br2) [47]. The results of condition screening showed that the tendency of suitable phosphine ligands for DArP with the PDI monomer was similar to that with the NDI monomer; a PDI-terthiophene alternating low-band-gap polymer PPDI3T with Mn = 13,600 was obtained in the presence of PCy3·HBF4 with strong basicity (Fig. 5a). Furthermore, random copolymerization of NDI2T-Br2, PDI2T-Br2, and 3,4-dimethylthiophene successfully afforded the random copolymer with NDI and PDI components, P(NDI3T-PDI3T), and the component ratio (m/n) could be tuned by adjusting the feed ratios of NDI2T-Br2 and PDI2T-Br2 (Fig. 5b). The light harvesting property was tuned in the range of 400–800 nm by adjusting the m/n component ratio.

Fig. 5
figure 5

Direct arylation synthesis of (a) PPDI3T and (b) P(NDI3T-PDI3T)

D–A alternating copolymers from AB system

The synthesis of low-band-gap polymers by using an AA + BB system is a good method in terms of systematic sample preparation. On the other hand, a common problem in the AA + BB system is that undesirable side reactions due to homo-couplings might occur [48, 49]. It is possible to minimize such structural defects by optimizing the reaction conditions in the AA + BB system [50, 51]. However, DArP using an asymmetric AB-type monomer (i.e., self-condensation-type DArP) could be a more effective way to fundamentally avoid this drawback. Furthermore, strict stoichiometric control can also be avoided in the AB system. Therefore, self-condensation-type DArP is more attractive than DArP in the AA + BB system from a practical point of view. However, the synthesis of low-band-gap polymers via self-condensation-type DArP is far less developed than that via DArP in the AA + BB system. Therefore, we investigated the synthesis of low-band-gap polymers via self-condensation-type DArP. Our designed NDI monomer end-capped with thiophene was considered a monomer with the D–A architecture. The self-condensation-type DArP using an asymmetric NDI-based D–A monomer NDI2T-Br was investigated (Fig. 6a) [52]. The combination of reagents (palladium catalyst, phosphine ligand, base, and solvent) for polymerization was optimized based on the direct arylation synthesis of NDI-based low-band-gap polymers described earlier. As a result, the NDI-dithiophene alternating low-band-gap polymer PNDI2T with the highest Mn (Mn = 13,000) was obtained in high yield under the Pd(OAc)2/PCy3·HBF4/K2CO3/PivOH/DMAc condition.

Fig. 6
figure 6

Direct arylation synthesis of (a) PNDI2T and (b) PTIG2T from asymmetric D–A monomer

The thienoisoindigo (TIG) structure has been investigated intensively as a new class of acceptor components for low-band-gap materials [53]. The high molecular ordering, expanded π-conjugation, and efficient charge transport due to the high coplanarity via S…O interactions, and the high charge delocalization via the quinoidal structure of the backbone are characteristics of the TIG structure. By utilizing these advantages, an OFET with a TIG-based low-band-gap polymer achieved ultrahigh hole mobility above 14 cm2/V·s [54]. showing the high potential of TIG-based low-band-gap polymers for opto-electrical applications. We herein investigated the direct arylation synthesis of new TIG-based low-band-gap polymers from an asymmetric TIG-based D-A monomer TIG2T-Br [55]. By optimizing the reaction conditions, a TIG-dithiophene alternating low-band-gap polymer PTIG2T with Mn = 8300 was obtained in the presence of Pd2dba3/PtBu2Me·HBF4/K2CO3/PivOH in toluene (Fig. 6b). The direct arylation synthesis of TIG-based low-band-gap polymers by the AA + BB system was reported by Wang et al. and Mei et al.; however, the polydispersities of the obtained polymers were relatively high in some cases, implying some undesirable side reactions [56, 57]. Furthermore, we investigated the photovoltaic property of a PTIG2T:PC61BM active layer in a conventional device configuration of ITO/PEDOT:PSS/PTIG2T:PC61BM/Ca/Al. A power conversion efficiency (PCE) of 3.19% was achieved with a Voc of 0.52 V, Jsc of 10.58 mA/cm2, and FF of 0.58. TIG-based low-band-gap materials synthesized by conventional cross-coupling reactions have been used in OPVs as donor materials; most of those OPVs exhibited PCE values of 1–3% [58,59,60]. The PCE of 3.19% from PTIG2T was one of the highest PCEs achieved by OPVs with TIG-based low-band-gap materials. Furthermore, this fabricated OPV is the first example that used a TIG-based low-band-gap material synthesized via DArP. These results demonstrate that a TIG-based low-band-gap polymer obtained by DArP can be utilized for OPV applications as effectively as those synthesized via conventional cross-coupling reactions.

Regioregular polyselenophenes from AB system

We are currently interested in polyselenophenes. Polyselenophenes have comparable HOMO levels to those of polythiophenes, while the LUMO levels of polyselenophenes are lower than those of polythiophenes [61,62,63]. Therefore, it is expected that polyselenophenes have the reduced optical band gaps, enhanced photostability, and improvement of electron transportation compared to polythiophenes. Considering these advantages, polyselenophenes have the potential to be more promising materials for organic electronics than polythiophenes. Nevertheless, studies of polyselenophenes, covering the synthesis, characterization, morphological analysis, and device applications, are far less developed than the research on polythiophenes. As for the synthesis of poly(3-alkylselenophene) which is one of basic polyselenophenes, oxidative polymerization [64], and Grignard-metathesis (GRIM) polymerization [62] have been reported. Unfortunately, the feature of the polymer structure, such as molecular weight, polydispersity, reaction sites, and so on, could not be controlled by oxidative polymerization. GRIM polymerization requires completely anhydrous reaction conditions to obtain regioregular poly(3-alkylselenophene)s with high molecular weight (Mn > 10,000) and narrow polydispersity (Mw/Mn < 1.2). Accordingly, there is a great need to develop more facile and efficient methods to synthesize polyselenophenes. We successfully synthesized regioregular poly(3-hexylselenophene) rr-P3HS (regioregularity: rr = 92–96%) under the simple phosphine-free DArP condition by using PdCl2 as a catalyst (Fig. 7a) [65]. The effectiveness of PdCl2 under the phosphine-free condition can be explained by the previously reported chloride-promoted direct arylation mechanism; chloride ion acted as a ligand, resulting in an efficient concerted metalation-deprotonation catalytic cycle [66]. Cheng et al. recently reported the synthesis of rr-P3HS via DArP under Pd(OAc)2-based conditions; however, both the molecular weight and the regioregularity were lower than those we obtained [67]. This fact supports the advantage of PdCl2 for the phosphine-free direct arylation synthesis of rr-P3HS. The assembly of nanofibers was observed by atomic force microscopy, and the assembly behavior (e.g., forming non-woven fibrous and bundle-like spherulitic self-assembled nanostructures) depended on the molecular weight of rr-P3HS (Fig. 7b). Furthermore, according to X-ray diffraction analysis, the formation of fibrous nanostructures was attributed to the pure crystalline structure with significant interdigitation of the hexyl side chains (Fig. 7c).

Fig. 7
figure 7

a Direct arylation synthesis of rr-P3HS. b Topographic AFM images (30 × 30 μm) of drop-cast rr-P3HS films prepared from chlorobenzene solutions. c XRD profiles of drop-cast rr-P3HS films prepared from chlorobenzene solutions

Conclusions

DArP is a promising candidate for an alternative method of synthesizing conjugated polymers in terms of efficiency and environmental friendliness. We developed a direct arylation methodology using acceptor monomers end-capped with electron-rich units (e.g., D–A monomers for AA + BB and AB system, respectively). By using these newly designed monomers under optimized direct arylation conditions, a variety of low-band-gap polymers, including NDI, PDI, and TIG-based low-band-gap polymers, were successfully synthesized. Furthermore, the direct arylation synthesis of rr-P3HS (rr = 92–96%) was achieved under PdCl2-based phosphine-free conditions. Our results demonstrate that DArP yields various desirable conjugated polymers with appropriate molecular design and adjustment of the catalytic system. There is still much room to pursue fundamental research with regard to DArP. Therefore, it is expected that further development by energetic investigation will lead to greater utility of DArP as an alternative conjugated polymer synthesis methodology.