Introduction

Organic photovoltaics (OPVs) based on conjugated polymeric materials have received considerable attention in recent years because of their advantages (for example, low cost, light weight, flexibility, and facile large-scale fabrication) compared with silicon-based solar cells.1, 2, 3, 4 One of the most successful systems currently utilizes a blend of regioregular poly(3-hexylthiophene) (P3HT) and [6,6]-phenyl-C61-butyric acid methyl ester (PC61BM) as the donor and acceptor materials, respectively, in the active layer, which provide power conversion efficiencies (PCEs) in the range of 3 and 4%. To further improve PCEs, the development of novel donor polymers with enhanced light-harvesting properties (that is, low band-gap polymers) has been extensively explored. For example, the benzo[1,2-b:4,5-b′]dithiophene (BDT)-,5, 6, 7, 8, 9, 10, 11 cyclopenta[2,1-b:3,4-b′]dithiophene-,12, 13 dithieno[3,2-b:2′,3′-d]silole-,14 diketopyrrolo[3,4–c]-pyrrole-1,4-dione-,15 and thieno[3,4–c]pyrrole-4,6-dione16 -based polymers have shown good promise with lower band gaps (Egopt) than that of P3HT (Egopt=1.90 eV). To date, OPVs using these low band-gap polymers have achieved excellent PCEs due to a high short-circuit current density (Jsc). In particular, a PCE of >9% was reported from an OPV with a BDT-based low band-gap polymer.11 These results indicate that low band-gap polymers are promising polymer motifs for high performance OPVs.

The major drawback of conventional donor polymers lies in their relatively high HOMO energy levels. The high HOMO levels in donor materials result in a low open-circuit voltage (Voc), which is directly related to the HOMO/LUMO offset of the donor and acceptor materials in the OPVs and prevents further improvement of the OPV performance. One method of controlling the HOMO/LUMO levels is the incorporation of electron-withdrawing units into the polymers. Previous studies have indicated that the incorporation of electron-withdrawing units into polymer backbones results in low LUMO levels. For example, donor-acceptor alternate copolymers containing arylene bisimide units are promising acceptor materials due to their low LUMO levels (ca. −4.0 eV).17, 18, 19 However, the incorporation of electron-withdrawing side chains into polymers can be an effective method for achieving low (deep) HOMO levels. Recently, Thompson and co-workers20 reported that a random P3HT-co-3-cyanothiophene) copolymer exhibited lower HOMO levels (5.30–5.34 eV) than P3HT (5.20 eV) due to the incorporation of electron-withdrawing cyano side chains, which contributed to a larger Voc in the random P3HT-co-3-cyanothiophene copolymer/PC61BM system in the range of 0.72 and 0.80 V (by contrast, the Voc with P3HT/PC61BM was 0.60 V). The OPV with the random P3HT-co-3-cyanothiophene copolymer achieved a higher PCE (3.29%) than that with P3HT (2.70%). This result suggested that the incorporation of electron-withdrawing side chains into polymers can lead to improvement of PCEs derived from the deeper HOMO levels in the donor polymers; however, further development of the polymer is essential for achieving excellent PCEs.

Herein, we describe BDT-based low band-gap polymers with electron-withdrawing side chains prepared using Stille coupling reactions between bis(trimethyltin)-BDT-based monomers (that is, 2,6-bis(trimethyltin)-4,8-bis(2-ethylhexyloxy)benzo[1,2-b:4,5-b′]dithiophene (BDT1) and 2,6-bis(trimethyltin)-4,8-bis(2-(ethylhexyloxy)thienyl)benzo[1,2-b:4,5-b′]dithiophene (BDT2)) and dibromothiophene monomers with electron-withdrawing side chains (that is, 2,5-dibromo-3-(cyanomethyl)thiophene (Th1), and 2,5-dibromo-3-cyanothiophene (Th2)). From this polymer design, low-lying HOMO levels derived from the incorporation of electron-withdrawing side chains are expected without sacrificing the excellent light-harvesting properties derived from the BDT-based structure, resulting in higher Voc values while maintaining light-harvesting properties than those of conventional donor polymers. In this study, a series of novel BDT-based low band-gap polymers with electron-withdrawing side chains (cyano or cyanomethyl group) was successfully synthesized using a combination of the aforementioned monomers in the presence of Pd(PPh3)4 under typical Stille coupling reaction conditions. Furthermore, their optical and electrochemical properties, thin film morphology and OPV performances were investigated in detail.

Experimental procedure

Materials

All of the reagents and solvents were used as received unless otherwise stated. BDT1,21, 22 BDT2,23, 24 Th1,25, 26 Th227 and 2,5-dibromo-3-hexylthiophene (Th3)28 were synthesized according to previously published protocols.

Procedure for the synthesis of polymers (P1–6)

A typical procedure for the synthesis of the polymers is as follows (P1): The dry toluene/DMF solution (10 ml, 4:1 v/v) of BDT1 (190 mg, 0.3 mmol), Th1 (80 mg, 0.3 mmol), and Pd(PPh3)4 (3 mg, 8 mol%) were refluxed for 2 days under a nitrogen atmosphere. After the reaction, the mixture was poured into methanol (120 ml) to yield the precipitate. The obtained precipitate was purified by Soxhlet extraction with methanol followed by extraction with chloroform (CF). The CF fraction was filtered through Celite, and poured into methanol (120 ml) to yield a red solid (130 mg, 91%). Mn=11 800, Mw/Mn=1.47. 1H NMR (CDCl3, δ, p.p.m.): 8.50–6.45 (br, 3H), 4.50–3.35 (br, 6H), 2.35–0.80 (m, 30H).

P2 from BTD1 and Th2

Red solid (78%). Mn=12 400, Mw/Mn=1.23. 1H NMR (CDCl3, δ, p.p.m.): 8.60–6.10 (br, 3H), 4.55–3.65 (br, 4H), 2.50–0.80 (m, 30H).

P3 from BTD1 and Th3

Red solid (89%). Mn=10 700, Mw/Mn=1.47. 1H NMR (CDCl3, δ, p.p.m.): 7.60–7.00 (br, 3H), 4.35–4.00 (br, 4H), 3.80–3.50 (m, 2H), 2.25–0.85 (br, 41H).

P4 from BTD2 and Th1

Red solid (55%). Mn=7500, Mw/Mn=1.40. 1H NMR (CDCl3, δ, p.p.m.): 7.90–6.00 (br, 5H), 4.00–2.25 (br, 6H), 2.15–0.80 (m, 30H).

P5 from BTD2 and Th2

Red solid (20%). Mn=3200, Mw/Mn=1.21. 1H NMR (CDCl3, δ, p.p.m.): 7.90–6.50 (br, 5H), 4.00–3.25 (br, 4H), 2.35–0.80 (m, 30H).

P6 from BTD2 and Th3

Red solid (79%). Mn=6900, Mw/Mn=1.70. 1H NMR (CDCl3, δ, p.p.m.): 7.90–6.50 (br, 5H), 3.75–3.25 (br, 4H), 2.95–2.25 (br, 2H), 1.95–0.80 (br, 37H).

Device fabrication and measurements

A typical procedure for the preparation of the indium tin oxide (ITO)/poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS)/polymer:PC61BM/Ca/Al architecture is as follows: commercially available prepatterned ITO substrates with a sheet resistance of 15Ω/□ were cleaned and plasma-etched prior to coating with a 30-nm layer of PEDOT: PSS, spin coated at 4000 r.p.m. for 40 s followed by annealing under flowing nitrogen at 120 °C for 10 min. The substrates were allowed to cool under a nitrogen atmosphere and then transferred to a glovebox for active layer deposition. The polymer:PC61BM blend solution was spin coated at 700 r.p.m. for 90 s. The blend solution, which consisted of 10 mg of each polymer dissolved in 1 ml of solvent (1:1 by weight, conc.=20 mg ml−1), was prepared in a glovebox. Then, the top electrode that consisted of Ca interlayer (20 nm) and Al electrode (80 nm) was applied using vacuum deposition. The density-voltage (J–V) characteristics of the devices were measured using a direct-current voltage and a current source/monitor (BSO-X500L, Bunko-Keiki, Tokyo, Japan) in a nitrogen atmosphere under an AM1.5G simulated solar light at 100 mW cm−2. The light intensity was corrected with a calibrated silicon photodiode reference cell (BS-520, Bunko-Keiki).

Characterization

The 1H (400 MHz) NMR spectra were recorded with a JEOL JNM-ECX400 (JEOL, Tokyo, Japan). The ultraviolet–visible (UV–Vis) spectra were recorded on a JASCO V-630BIO UV–Vis spectrophotometer (JASCO, Tokyo, Japan). The number-average molecular weight (Mn) and molecular weight distribution (Mw/Mn) were estimated by size exclusion chromatography using a HLC-8320 system (TOYO Corporation, Tokyo, Japan). The column setup was as follows: a guard column (TSK guard column HXL-L, (TOYO Corporation)) and three consecutive columns (G4000HXL, G3000HXL and G2000HXL) were eluted with CF at a flow rate of 1.0 ml min−1. Polystyrene standards were employed for calibration. Cyclic voltammetry experiments for the polymer thin films were performed on a BAS (Tokyo, Japan) electrochemical analyzer (model 660C). A three-electrode cell was used with platinum electrodes as both the counter and working electrodes. Silver/silver ion (Ag in 0.1 M AgNO3 solution) was used as the reference electrode. Ferrocene/ferrocenium (Fc/Fc+) was used as an internal standard. The potential values obtained in reference to Ag/Ag+ were converted to the values relative to the saturated calomel electrode. Thermal analysis was performed on a SEIKO SSC6200 and SEIKO EXSTAR 6000 DSC 6200 at a heating rate of 10 °C min−1 for thermogravimetry and differential scanning calorimetry (DSC) under a nitrogen atmosphere, respectively. Tapping mode atomic force microscopy (AFM) observation was performed with an Agilent (Agilent Technologies, Santa Clara, CA, USA) AFM 5500 using micro-fabricated cantilevers with a force constant of ~34 N m−1.

Results and Discussion

Synthesis and Characterization

The monomers for the synthesis of the polymers (that is, BDT1, BDT2, Th1, Th2 and Th3) were prepared according to previously reported protocols.18, 19, 20, 21, 22, 23, 24, 25 Then, Stille coupling reactions were carried out in the presence of Pd(PPh3)4 for 2 days in a toluene/DMF solution under refluxing conditions (Scheme 1).29 The results are summarized in Table 1. P1 was obtained as a red solid in high yield (91%) with a molecular weight that was estimated to be Mn=11 800 and Mw/Mn=1.47 based on size exclusion chromatography analysis. P2 and P3 were successfully obtained in relatively high yields (78 and 89%), and their molecular weights were Mn=12 400 (Mw/Mn=1.23) and Mn=10 700 (Mw/Mn=1.47), respectively, based on size exclusion chromatography analysis. The molecular weight of P3, which was previously reported (Mn=10 800, and Mw/Mn=1.7), was nearly the same value as that of P3 synthesized here.29 P4, P5 and P6 were also synthesized under Stille coupling reaction conditions. However, large amounts of the CF-insoluble fractions were formed in P4 and P5. The molecular weights of the CF-soluble P4 and P5 fractions were estimated to be Mn=7500 and Mn=3200, respectively. In addition, the yield of each was relatively low (55% and 20%, respectively). Previously reported thienyl-substituted BDT-based polymers also exhibited relatively low yields (58–29%).21 Therefore, this problem may be due to the low solubility of the high-molecular-weight products (that is, CF-insoluble fractions). The structures of P1–6 were characterized by 1H NMR spectroscopy. As shown in Figure 1, all of the peaks were assigned in the 1H NMR spectrum of P1 (see the 1H NMR spectra of P2–6 in Supplementary information).

Table 1 Polymerization results and thermal properties of P1–6
Full size table
Figure 1
figure 1

1H NMR spectrum of P1 in CDCl3. The residual CHCl3 peak is marked.

Full size image

The solubility of P1–6 in common organic solvents was evaluated, and the results are summarized in Table 2. P1 and P3 exhibited good solubility in CF and o-dichlorobenzene (DCB), whereas the solubility of P2 in CB and DCB was lower than that of P1 and P3 due to the short cyano side chain. For P4–6, P4 and P5 exhibited much lower solubility than P1–3, which may be due to the planar thienyl side chain and other short side chains (that is, cyano or cyanomethyl groups).

Table 2 Solubility of P1–6
Full size table

Thermal properties

The thermal properties of P1–6 were investigated by thermogravimetry and DCS measurements. The results are summarized in Table 1. As shown in Figure 2, the 5% weight decomposition temperatures (T5d) were >300 °C in P1–6. In particular, P4–6 exhibited much higher thermal stability (T5d>370 °C) than P1–3 (T5d≈320 °C), due to the thermal stability of the thienyl side chains. In the DSC profiles (Figure 3), the glass transition temperatures (Tg) of P1–3 were observed in the range of 110 to 117 °C. However, the Tg of P4–6 could not be observed.

Figure 2
figure 2

TG analysis of P1–6. TG, thermogravimetry. A full color version of this figure is available at Polymer Journal online.

Full size image
Figure 3
figure 3

DSC profiles of P1–6. DSC, differential scanning calorimetry. A full color version of this figure is available at Polymer Journal online.

Full size image

Optical properties

The optical absorption spectra of P1–6 in a dilute CF solution and as thin films were investigated (Figure 4), and the results are summarized in Table 3. In the CF solutions (Figure 4a), the maximum absorption of P1 and P2 was observed at 496 nm and 520 nm, respectively, which were red-shifted compared with P3 (473 nm) due to the bathochromic effect of the electron-withdrawing side chains. The optical band gaps (Egopt) of the P1–3 solutions were estimated to be 2.03 eV, 1.88 eV and 2.21 eV, respectively, from the onset absorption (λonset) at 640 nm, 660 nm and 560 nm, respectively. For thin films, P1–3 exhibited narrower Egopt than that in solution, and Egopt was estimated to be 1.65 eV, 1.51eV and 1.76 eV, respectively (Figure 4b). P4–6 exhibited the same absorption behaviors as P1–3, and the bathochromic effects in P4 and P5 led to a red-shifted maximum absorption (498 nm and 529 nm, respectively) compared with P6 (487 nm). In addition, the absorption due to the thienyl side chain was observed in the range of 350 to 400 nm in P4–6, suggesting that the polymer side chains are an important factor for achieving light-harvesting properties over a wide range. The Egopt of the P4–6 solutions was estimated to be 2.00 eV, 1.85 eV and 2.14 eV, respectively, which were narrower in the thin films (1.61 eV, 1.49 eV and 1.75 eV, respectively). Among P1–6, the P5 thin film exhibited the narrowest Egopt, and the enhanced planarity due to the thienyl and short cyano side chains contributed to the narrowing of the optical band gap.

Figure 4
figure 4

UV–Vis absorption of P1–6 in chloroform solutions (a) and thin films (b). UV–Vis, ultraviolet–visible. A full color version of this figure is available at Polymer Journal online.

Full size image
Table 3 Optical properties of P1–6
Full size table

Electrochemical properties

The electronic states of the polymer thin films were investigated by cyclic voltammetry) measurements. To determine the HOMO levels, each measurement was calibrated using the saturated calomel electrode. The electrochemical data are summarized in Figure 5. The HOMO levels of P1–3 were estimated to be −5.53 eV, −5.61 eV and −5.28 eV, respectively, based on their onset oxidation potentials (Eoxonset) of 1.13 eV, 1.21 eV and 0.88 eV, respectively. As expected, the incorporation of electron-withdrawing side chains led to deep HOMO levels in P1 and P2 compared with P3. P4–6 exhibited the same electrochemical behaviors as P1–3. Deep HOMO levels were observed in P4 and P5 and were estimated to be −5.55 eV and −5.65 eV, respectively. In a previous study, Prasad and co-workers30 also reported that polymers with electron-withdrawing carboxylate side chains possessed deep HOMO levels between −5.41 and −5.50 eV. These results demonstrated that the incorporation of electron-withdrawing side chains effectively resulted in deep HOMO levels, and the electrochemical properties of our polymers were independent of the differences in the side chains in the BDT units based on the cyclic voltammetry results.

Figure 5
figure 5

HOMO/LUMO diagram of P1–6 thin films. A full color version of this figure is available at Polymer Journal online.

Full size image

Photovoltaic performance

The photovoltaic properties of P1–3 were studied using PC61BM as an acceptor material in a conventional device configuration (that is, ITO/PEDOT:PSS/polymer:PC61BM/Ca/Al). It is important to note that the photovoltaic properties of P4 and P5 were not investigated because it was difficult to fabricate OPVs because of their low solubility and low molecular weights. The current J–V curves of the fabricated photovoltaics under AM1.5G solar irradiation (100 mW cm−2) are shown in Figure 6. The Voc, Jsc, fill factor (FF) and PCE values are summarized in Table 4. In the P1/PC61BM system, the OPVs with active layers fabricated from CF and CB solutions exhibited low PCEs (<1%). By contrast, the OPV with an active layer fabricated from the DCB solution exhibited a much higher PCE (2.34%) than that of the OPVs from the CF and CB solutions. Furthermore, the thermal annealing at 140 °C led to the highest PCE of 2.68% with Voc of 0.71 V, Jsc of 10.64 mA cm−2 and FF of 0.35. To date, the highest PCE of the OPV with the P3/PC61BM active layer is 0.86% with Voc of 0.59 V, Jsc of 4.93 mA cm−2 and FF of 0.30.26 The larger Voc of the P1/PC61BM system (0.71 V) compared with the P3/PC61BM system (0.59 V) is most likely due to the deeper HOMO level of P1 due to the incorporation of the electron-withdrawing side chains, which contributed to the higher PCEs. To improve the PCEs, 1,8-diiodooctane was introduced into the P1/PC61BM system as an additive.31, 32 The device with the P1/PC61BM system was fabricated using the DCB/1,8-diiodooctane solution (97/3 vol%) followed by thermal annealing at 140 °C. Surprisingly, the PCE dramatically decreased to 1.10% with Voc of 0.76 V, Jsc of 4.96 mA cm−2 and FF of 0.29 (Supplementary Figure S7). This result may be due to the inferior morphology, which led to ineffective charge separation.

Figure 6
figure 6

J–V characteristics of P1/PC61BM system fabricated from the DCB solutions under AM1.5G irradiation (100 mW cm−2). The active layer was annealed at 100 °C (solid line) and 140 °C (dotted line). DCB, o-dichlorobenzene; PC61BM, P1/[6,6]-phenyl-C61-butyric acid methyl ester. A full color version of this figure is available at Polymer Journal online.

Full size image
Table 4 OPV performance of P1–3 with PC61BM
Full size table

For the P2/PC61BM system, low PCEs were observed for all of the devices. This unfavorable performance was due to two reasons. The first reason was the low solubility of P2 compared with P1, preventing the thorough mixing of P2 and PC61BM in the active layer (see next section in detail). The second reason is the LUMO level of P2 (−4.1 eV), which was nearly the same as that of PC61BM (ca. −4.1 eV) and was likely too deep for use with PC61BM. However, the incorporation of electron-withdrawing side chains is an efficient method for achieving the large Voc and high PCE values. Tuning the depth of the HOMO/LUMO levels within the proper range by the incorporation of suitable electron-withdrawing side chains may be a promising approach for improving PCEs.

Thin film morphology

The effect of the surface morphology on the photovoltaic performance was investigated by AFM. As shown in Figure 7a, the root-mean-square roughness of the P1/PC61BM OPV with PCE=2.34% was 4.62 nm. In comparison to Figure 7a, the device with the higher PCE (2.68%) exhibited a smaller root-mean-square roughness (3.76 nm in Figure 7b). By contrast, the P2/PC61BM OPVs with the low PCEs (0.56 and 0.63%) exhibited a much higher root-mean-square (7.45 nm and 6.07 nm, respectively). Furthermore, the high root-mean-square (9.05 nm) was also observed in the P1/PC61BM OPVs fabricated from the DCB/1,8-diiodooctane solution (PCE=1.10%, Supplementary Figure S8). Therefore, the OPV performance was primarily dependent on the surface roughness of the active layers. The high surface roughness of the P2/PC61BM system may be due to the low solubility of P2, as previously described, which resulted in a much lower PCE than the P1/PC61BM system.

Figure 7
figure 7

AFM height images of the active layers fabricated from the CDB solutions. (a) P1/PC61BM system annealed at 100 °C, (b) P1/PC61BM system annealed at 140 °C, (c) P2/PC61BM system annealed at 100 °C, and (d) P2/PC61BM system annealed at 140 °C. AFM, atomic force microscopy; PC61BM, P1/[6,6]-phenyl-C61-butyric acid methyl ester. A full color version of this figure is available at Polymer Journal online.

Full size image

Conclusion

A series of novel BDT-based low band-gap polymers P1–6 were successfully synthesized using the Stille coupling reaction. P1, 2, 4 and 5 exhibited wide light absorption due to the bathochromic effect resulting from the incorporation of electron-withdrawing side chains (that is, cyano or cyanomethyl groups) compared with P3 and P6 with hexyl side chains. The P5 thin film exhibited the narrowest optical band gap of 1.49 eV. The cyclic voltammetry measurements demonstrated that the incorporation of electron-withdrawing side chains led to deep HOMO levels, and these levels were estimated to be −5.53 eV, −5.61 eV, −5.55 eV and −5.65 eV in P1, 2, 4 and 5, respectively. The OPV with the P1/PC61BM system achieved a PCE of 2.68% with Voc of 0.71 V, Jsc of 10.64 mA cm−2 and FF of 0.35. By contrast, the OPV with the P3/PC61BM system exhibited a PCE of 0.86% with Voc of 0.59 V, Jsc of 4.93 mA cm−2 and FF of 0.30. These results demonstrate that improvement in the PCEs was accomplished by the increased Voc due to the deeper HOMO levels. Therefore, the incorporation of electron-withdrawing side chains is a promising strategy for improving PCEs.

scheme 1

Synthesis of P1–6.