Introduction

In the past few years, bulk heterojunction (BHJ) polymer solar cells (PSCs) have drawn great attention because of their unique merits, such as light weight, flexibility, low cost, and easy processing [1,2,3]. The power conversion efficiencies (PCEs) of BHJ PSCs have recently exceeded 15% [4]. In the early stage, high-efficiency BHJ PSCs consisted of narrow bandgap polymer donors with soluble fullerene acceptors, namely, [6,6]-phenyl-C61-butyric acid methyl ester (PC61BM) and its C71 analog (PC71BM), owing to their isotropic electron-transporting capacity, high electron mobility, and ability to induce nanoscale phase separation [5, 6]. However, these fullerene acceptors are relatively expensive, and there are some intrinsic shortcomings of fullerene materials, for instance, limited chemical and energetic tunability, narrow absorption spectra, and morphological instability [7, 8]. Therefore, it is necessary to develop new acceptor materials to solve these problems. Nonfullerene acceptors (NFAs), including small-molecule acceptors [9, 10] and polymeric acceptors [11,12,13,14], which have the advantages of broader and stronger absorb spectra, tunable energy levels and lower cost, have drawn increasing attention in recent years.

Small-molecule NFAs exhibit good reproducibility because of their exact molecular structure and high purity. Therefore, at present, much more efforts are devoted to developing small-molecule NFAs than polymeric NFAs [15,16,17]. Diketopyrrolopyrrole (DPP) dyes have several attractive properties: strong light absorption, good photochemical stability, facile synthetic modification, and high electron mobilities [18,19,20]. Therefore, some DPP-based materials have been recently developed as acceptors for PSCs. However, not only simple linear DPP-based materials [21,22,23,24] but also three-dimensional (3D) DPP-based acceptors [25,26,27] provided relatively lower PCEs than those of other acceptors when blended with poly(3-hexylthiophene) (P3HT) as donors in BHJ devices. Chen et al. achieved the highest PCE of 5.16% when SF(DPPB)4 with a spirofluorene (SF) core was used as an acceptor and P3HT was used as a donor [27]. Recently, Jun-Hua Wan et al. reported that three NFAs that also had an SF core, SF(DPP3F)4, SF-(DPPBT)4 and SF-(DPPBF)4 [28], yielded moderate PCE values (≤ 2%) when blended with P3HT. It is not hard to observe that most of the attention has been paid to the orthogonal SF core, which is owing to its ability to decrease the crystallinity of DPP dyes, to improve phase separation when blended with P3HT [27] and to enhance multidimensional charge-transporting capacities in the quasi-3D configuration. In our former studies, we used a rotary biphenyl (BP) core to displace the SF in 3D NFAs based on perylene diimides (PDI). The BP core is much cheaper than SF [29], and the photovoltaic performance of BP–PDI4 using BP as the core outperforms that of NFAs using SF or phenyl as the core unit. The BP core can not only take advantage of the quasi-3D configuration but also result in a smaller aggregation tendency, which is also favorable for better device performance when blended with suitable donors [15].

In this study, a new nonfullerene electron acceptor, BP(DPPB)4, was designed and synthesized via the replacement of the SF core in SF(DPPB)4 with a BP core. The optical, thermal, and photovoltaic properties of BP(DPPB)4 were well characterized. After optimizing the thermal annealing conditions, PSCs based on P3HT and BP(DPPB)4 can achieve a moderate PCE of 1.43%, with a high open-circuit voltage (Voc) of 1.17 V, indicating that BP is a potential candidate to replace SF in constructing multidimensional NFAs.

Experimental section

Instruments

1H-NMR spectra were measured on a Bruker AVANCE 400 MHz NMR (Billerica, MA, USA) spectrometer. Mass spectra were recorded on a Bruker BIFLEX III MALDI-TOF mass spectrometer. Thermogravimetric analysis (TGA) measurements were carried out on a PerkinElmer thermal analysis apparatus (Waltham, MA, USA). Differential scanning calorimetry (DSC) measurements were performed using a METTLER differential scanning calorimeter (DSC822e) (Zurich, Switzerland). TGA and DSC measurements were conducted under nitrogen at a heating rate of 10 °C min−1. Absorption spectra were acquired on a PerkinElmer Lambda 35 (Waltham, MA, USA). The cyclic voltammetry (CV) measurements were recorded on a CHI604E electrochemical workstation (Shanghai, China) by using Pt wire as the counter electrode, a glassy carbon electrode as the working electrode, and a Ag/AgCl electrode as the reference electrode in a 0.1 mol L−1 tetrabutylammonium hexafluorophosphate (Bu4NPF6) solution with a scanning rate of 100 mV s−1, and the potential of the Ag/Ag+ reference electrode was internally calibrated by using ferrocene/ferrocenium (Fc/Fc+) as the redox couple. TEM images were recorded by using a JEM-2100/HR TEM (Beijing, China). The current density–voltage (J–V) characteristics of the devices were measured under AM 1.5 G illumination using a solar simulator (SS-F5-3A, Enli Technology Co., Ltd) (Santa Clara, California, USA). The light intensity was carefully calibrated with a standard silicon reference before every test.

Materials

All the chemicals and solvents used, unless otherwise specified, were purchased from commercial suppliers (J&K, Energy Chemical, Sinopharm Chemical Reagent Co., Ltd) and used without further purification. P3HT was purchased from J&K. The syntheses of 2,5-bis-(2-ethyl-hexyl)−3-(5-bromo-thiophen-2-yl)−6-(5-phenyl-thiophen −2-yl)−2,5-dihydropyrr-olo[3,4-c]pyrrole-1,4-dione (DPPB-Br) (4) and 3,5,3’,5’-tetrakis(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)biphenyl (BP-Bpin4) (5) were conducted according to references [29,30,31,32].

Synthesis

2,5-diethylhexyl-3,6-dithiophen-2-ylpyrrolo[3,4-c]pyrrole-1,4-dione (1)

1H-NMR (400 MHz, CDCl3, ppm): δ 8.94 (d, 2 H), 7.64 (d, 2 H), 7.27 (d, 2 H), 4.02 (m, 4 H), 1.83 (m, 2 H), 1.36–1.22 (m, 16 H), 0.81 (m, 12 H).

3-(5-bromothiophen-2-yl)−2,5-bis(2-ethylhexyl)−6-(thiophen-2-yl)−2,5-dihydropy- rrolo[3,4-c]pyrrole-1,4-dione (2)

1H-NMR (400 MHz, CDCl3, ppm): δ 8.91 (dd, 1 H), 8.65 (d, 1 H), 7.64 (dd, 1 H), 7.27–7.25 (m, 1 H), 7.20 (d, 1 H), 4.01–3.99 (m, 2 H), 3.98–3.94 (m, 2 H), 1.86–1.80 (m, 2 H), 1.39–1.22 (m, 16 H), 0.91–0.83 (m, 12 H).

2,5-bis(2-ethylhexyl)−3-(5-phenylthiophen-2-yl)−6-(thiophen-2-yl)−2,5-dihydropy- rrolo[3,4-c]pyrrole-1,4-dione (3)

1H-NMR (400 MHz, CDCl3, ppm): δ 8.96 (d, 1 H), 8.92 (d, 1 H), 7.70–7.66 (m, 2 H), 7.64–7.60 (m, 1 H), 7.46–7.25 (m, 5 H), 4.05–4.01 (m, 4 H), 1.98–1.82 (m, 2 H), 1.41–1.24 (m, 16 H), 0.92–0.83 (m, 12 H).

2,5-bis-(2-ethyl-hexyl)−3-(5-bromo-thiophen-2-yl)−6-(5-phenyl-thiophen-2-yl)−2,5-dihydropyrrolo[3,4-c]pyrrole-1,4-dione (4)

1H-NMR (400 MHz, CDCl3, ppm), δ 8.95 (d, 1 H), 8.61 (d, 1 H), 7.67 (d, 2 H), 7.46–7.32 (m, 4 H), 7.19 (d, 1 H), 4.06–3.92 (m, 4 H), 1.94–1.80 (m, 2 H), 1.39–1.21 (m, 16 H), 0.91–0.82 (m, 12 H).

3,5,3’,5’-tetrakis(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)biphenyl (5)

1H-NMR (400 MHz, CDCl3, ppm): δ 8.23 (s, 2 H), 8.12 (d, 4 H), 1.38 (s, 48 H).

6,6’,6”,6”‘-([1,1’-biphenyl]−3,3’,5,5’-tetrayltetrakis(thiophene-5,2-diyl))tetrakis (2,5-bis(2-ethylhexyl)−3-(5-phenylthiophen-2-yl)−2,5-dihydropyrrolo[3,4-c] pyrrole-1,4-dione) (6) (BP(DPPB)4)

5 (150 mg, 0.23 mmol), 4 (774.75 mg, 1.14 mmol), tetrakis(triphenylphosphine)palladium(0) (21.07 mg, 118.24 µmol), and tetrabutylammonium bromide (100 mg, 0.31 µmol) were added to a three-necked 100 mL round bottom flask under no light. The flask was purged by eight vacuum/nitrogen fill cycles to remove oxygen and fill the flask with nitrogen. K2CO3 aqueous solution (2 mol L−1, 4 mL) and toluene (20 mL) were added to the flask. However, all solutions should be blown with nitrogen for 20 min in advance. The mixture was heated to 80 °C and stirred for 4 days. After cooling to room temperature, the reaction solution was washed with water and extracted repeatedly with dichloromethane. The organic layer was combined and dried with anhydrous sodium sulfate. Then, the solvent was removed from the crude product, which was purified by silica gel chromatography with dichloromethane/methanol (150:1, v/v) as the eluent. The product was obtained as magenta powder (261 mg, yield 45%). 1H-NMR (400 MHz, CDCl2, ppm): δ 9.11–9.01 (dr, 8 H), 7.90–7.80 (dr, 6 H), 7.58 (dr, 12 H), 7.37–7.34 (dr, 16 H), 3.94–3.89 (d, 16 H), 1.90 (m, 8 H), 1.38–1.16 (m, 64 H), 0.90–0.74 (m, 48 H). MS spectrum (MALDI, M+): calculated for C156H178N8O8S8: 2549.7, found: 2571.4([M + Na+]). The elemental analysis: calculated: C (73.49%), H (7.04%), N (4.39%), found: C (73.66%), H (6.81%), N (4.39%) (Supporting information Figure S1 and Table S1).

DFT calculation

DFT calculations were performed at the B3LYP/6-31 g(d,p) level to determine the geometric configurations and then analyze the electronic structures of BP(DPPB)4. All the calculations were performed using Gaussian 09 (version A.01). All N-ethylhexyl substituents were replaced with methyl groups to simplify the calculations.

Device fabrication

The PSC devices were fabricated with the following device structure: ITO/PEDOT:PSS/P3HT:BP(DPPB)4/Ca/Al, and the process of specific device fabrication was performed according to references [26,27,28,29]. The difference is that for the active layer, a solution of P3HT and BP(DPPB)4 dissolved in chloroform with 2:1 blend weight ratios at a total solid concentration of 15 mg ml−1 was spin-cast at 3500 rpm onto the substrates. In addition, the samples were annealed at different annealing temperatures (80 °C, 120 °C, 150 °C) for 10 min. The effective device area was 0.04 cm2. Device characterization was carried out under an AM 1.5 G solar simulator (SS-F5-3A, Enli Technology Co., Ltd) (Beijing, China) with an intensity of 100 mW cm−2. The charge carrier mobilities of the P3HT:BP(DPPB)4 films were measured using the space-charge-limited current (SCLC) method. Hole-only devices with a structure of ITO/PEDOT:PSS/active layer/MoO3/Ag and electron-only devices with a structure of ITO/ZnO/active layer/Ca/Al were also fabricated.

Fig. 1
figure 1

a Chemical structures of BP(DPPB)4 and P3HT; b energy level diagram; c device structure

Results and discussion

Synthesis and characterization

The synthetic route for BP(DPPB)4 is shown in Scheme 1, and the chemical structures, energy levels of BP(DPPB)4 and P3HT are shown in Fig. 1. BP(DPPB)4 was synthesized via the Suzuki coupling reaction between BP-pin4 and DPPB-Br [27] using Pd(PPh3)4 as a catalyst in a moderate yield of 45%. BP(DPPB)4 was characterized by proton nuclear magnetic resonance (1H-NMR), mass spectrometry, and elemental analysis. Thermal properties were studied by TGA and DSC. BP(DPPB)4 exhibited excellent thermal stability with a high decomposition temperature (Td) of 409 °C at 5% weight loss under a nitrogen atmosphere (as shown in Fig. 2a). The DSC traces for this compound showed a glass transition at ca. 120 °C, solid-phase conversion at ca. 175 °C and a melting peak at ca. 246 °C during the heating process, suggesting that BP(DPPB)4 have a certain degree of crystallinity.

Scheme 1
scheme 1

Synthetic route for BP(DPPB)4

Fig. 2
figure 2

a TGA thermograms of BP(DPPB)4; b DSC thermograms of BP(DPPB)4

Theoretical calculation

To analyze the performance at the molecular level, the optimization of molecular geometries and successive analysis of frontier molecular orbital electron density distributions were calculated by density functional theory (DFT). As shown in Fig. 3a, the local optimal configuration of BP(DPPB)4 was a three-dimensional molecule with an “X” configuration (see energy variations of different possible configurations derived from the connection of DPPB chains, Supporting information Figure S3), and the dihedral angle between the two DPP arms was 57.3°. Furthermore, the dihedral angle between the two benzene planes of the BP core was 143.65°. This molecular configuration can inhibit the strong intermolecular aggregation of BP(DPPB)4. In addition, the two dihedral angles between the two thiophene planes and the DPP building block were 169.47° and 168.34°, and the two dihedral angles between DPP and the two benzene planes, i.e., the benzene group of the BP core and the terminal benzene, were 27.30° and 25.26°, respectively, as shown in Fig. 3b. This result indicates that each arm maintains effective conjugation for photon harvesting and charge transport. Compared with the SF core, the BP core is much smaller in size; additionally, the dihedral angle between the two benzene planes of the BP core is 143.65°, which indicates that the geometry of BP(DPPB)4 is much more compact than that of SF(DPPB)4.

Fig. 3
figure 3

Local optimized geometries and frontier molecular orbital distributions of BP(DPPB)4

The distributions of the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) as well as LUMO+1 can also be seen in Fig. 3c and Fig. 3d. The calculated LUMO and LUMO+1 levels are −2.62 eV and −2.59 eV, respectively. The gap between the LUMO and LUMO+1 (ΔLUMO) is 0.03 eV (< 0.5 eV ΔLUMO). The results indicate that BP(DPPB)4 is a highly efficient electron acceptor because the low-lying LUMO+1 may accelerate the charge separation at the interface and induce efficient electron acceptance [33,34,35].

Optical and electrochemical properties

Figure 4a depicts the molar extinction coefficient of BP(DPPB)4 in CHCl3 solution and the normalized UV spectrum of BP(DPPB)4 in a thin film. BP(DPPB)4 showed two distinct absorption bands, including a weak band in the shorter wavelength region (300–450 nm) and a strong band with a typical aggregation peak in the longer wavelength region (500–700 nm). This spectral profile is well complementary with that of P3HT (400–600 nm), which promotes good light harvesting capability. In solution, the absorption maximum was 607 nm, whereas the absorption maximum in the thin film was 641 nm. This redshift (34 nm) in the maximum absorption peak from the solution to film indicates effective π–π stacking in the thin film. The molar extinction coefficient of BP(DPPB)4 in solution reached up to 2.0 × 105m−1 cm−1, which is beneficial to device performance. Moreover, from the absorption band edges (λonset, 693 nm) of the BP(DPPB)4 film, the optical band gap (Egopt) is calculated to be 1.79 eV.

Fig. 4
figure 4

a Absorption spectra of BP(DPPB)4 in film and solution; b CV curve of BP(DPPB)4

CV measurements were performed to characterize the electrochemical properties of BP(DPPB)4. The LUMO and HOMO energy levels can be estimated from the onset potentials for reduction and oxidation with ferrocene/ferrocenium (Fc/Fc+) as the internal standard, respectively. The CV curves are shown in Fig. 4, and the date are summarized in Table 1. As shown in Fig. 4b, the onset reduction potential was −1.13 V, and the onset oxidation potential was 0.75 V. Thus, the LUMO level is calculated to be −3.63 eV, and the HOMO level is calculated to be −5.31 eV for BP(DPPB)4. The energy levels are suitable for a high-performance polymer donor (P3HT, as shown in Fig. 1b). The energy offsets (Δ ELUMO) between the LUMO of P3HT and that of BP(DPPB)4 was 0.69 eV, and the relative HOMO offsets (Δ EHOMO) was 0.37 eV, both of which are large enough to guarantee efficient exciton dissociation [37]. The difference between the LUMO of BP(DPPB)4 and the HOMO of P3HT was as large as 1.31 eV, which may result in PSCs with a high Voc.

Table 1 Optical and electrochemical properties of BP(DPPB)4

Morphology

The morphology of the blend film was characterized by transmission electron microscopy (TEM). The TEM images (Fig. 5) of as-cast devices do not show obvious D–A phase separation (Fig. 5a). When the devices were annealed by thermal annealing at 80 °C for 10 min, slightly interpenetrating structures and phase separation appeared (Fig. 5b). As shown in Fig. 5c, upon thermal annealing at 120 °C for 10 min, obvious nanoscale interpenetrating structures and improved phase separation between the donor and acceptor formed. This morphology would lead to more-efficient exciton dissociation and charge transfer and result in higher Jsc and FF values for annealed devices. However, when the thermal annealing temperature was increased to 150 °C, the nanoscale structures disappeared, and phase separation decreased, as shown in Fig. 5d. This effect possibly would result in lower Jsc and FF values. Clearly, the results are closely associated with the DSC results. Because the glass transition temperature (Tg) of BP(DPPB)4 is 120 °C, when the annealing temperature reaches 120 °C, BP(DPPB)4 is able to further self-aggregate, resulting in good morphology with obvious nanoscale interpenetrating structures and appropriate phase separation [38, 39]. However, when the annealing temperature is higher than the Tg of BP(DPPB)4 at 150 °C, BP(DPPB)4 diffuses into the P3HT matrix, resulting in inconspicuous phase separation between the donor and acceptor.

Fig. 5
figure 5

TEM images of P3HT:BP(DPPB)4 blend films that are a as cast; b thermally annealed at 80 °C for 10 min; c thermally annealed at 120 °C for 10 min; and d thermally annealed at 150 °C for 10 min

Photovoltaic performance

Because the absorption spectrum and the energy levels of BP(DPPB)4 matched well with those of P3HT, P3HT and BP(DPPB)4 were used as the electron donor and electron acceptor, respectively, to fabricate PSCs. The structure of the PSCs was ITO/PEDOT:PSS/P3HT:BP(DPPB)4/Ca/Al with a donor:acceptor weight ratio of 2:1. Different annealing temperatures are assessed to optimize device performance. Figure 6a shows the J–V)characteristics of the PSCs under AM 1.5G illumination at an intensity of 100 mW cm−2, and all photovoltaic data are summarized in Table 2. Thermal annealing can significantly improve device performance. A low PCE of 0.35% with a Voc of 1.19 V, a Jsc of 1.15 mA cm−2, and an FF of 0.25 was obtained for the as-cast devices. When there was thermal annealing and the temperature increased to 80 °C, the photovoltaic performance improved sharply to a PCE of 0.83% with a Voc of 1.19 V, a Jsc of 2.43 mA cm−2, and an FF of 0.29.

Fig. 6
figure 6

a JV characteristics of solar cells based on P3HT:BP(DPPB)4 (D:A = 2:1) with different thermal annealing temperatures; photocurrent density versus effective voltage characteristics of b EQE curves of solar cells based on P3HT:BP(DPPB)4

Table 2 Summary of photovoltaic performance of PSCs based on P3HT:BP(DPPB)4 after thermal annealing at different temperatures (D:A = 2:1)

Furthermore, upon thermal annealing at 120 °C for 10 min, the device could reach a maximum PCE of 1.43% with a Voc of 1.17 V, a Jsc of 3.41 mA cm−2, and an FF of 0.36, which is consistent with previous reports in the literature [21,22,23,24,25,26,27,28, 40,41,42] (Supporting information Table S2 and Scheme S1). However, when the thermal annealing temperature was further increased to 150 °C, the performance decreased notably, with a PCE of 0.61%, a Voc of 1.12 V, a Jsc of 1.53 mA cm−2, and an FF of 0.36. Apparently, thermal annealing significantly improves the Jsc and FF, thereby further improving device performance. In addition, the enhanced photovoltaic performance of the device that was thermally annealed at 120 °C for 10 min should also be ascribed to the lower series resistance (Rs) of 3992.90 ohm because of the interpenetrating network in the active layer, as proved by the TEM results, and the higher shunt resistance (Rsh) of 17373.48 ohm, indicating reduced leakage current.

The photoresponses of PSCs based on P3HT:BP(DPPB)4 with or without thermal annealing at 120 °C for 10 min were investigated by determining the external quantum efficiency (EQE) covering the wavelength range from 300 to 800 nm. As shown in Fig. 6b, the EQE curves of the two PSCs showed a similar broad range from 300 nm to 720 nm. It is obvious that the EQE value of the annealed devices is 4 times higher than that of the as-cast device. The maximum EQE of the annealed devices reached 20% at 520 nm, which is in agreement with the variations in Jsc values in the devices. We also used the SCLC method to measure the charge carrier mobility in order to investigate the electron transport properties of the devices. Figure 7 shows the dark JV curves of the hole-only and electron-only devices after thermal annealing at different temperatures, and all data are summarized in Table 3. It is clear that thermal annealing promotes the charge carrier mobility. The as-cast devices provided the lowest hole and electron mobility. As the annealing temperature increased, the charge carrier mobility increased. When the annealing temperature increased to 120 °C, the highest hole and electron mobility of the blend film were calculated to be 4.83 × 10−4 cm2 V−1 s−1 and 2.45 × 10−4 cm2 V−1 s−1, respectively. However, when the thermal annealing temperature was further increased to 150 °C, the charge carrier mobility decreased, which is consistent with the trend in the PCE values of the devices. In addition, the charge carrier mobility data in this work are comparable with the data reported by other research groups, as listed in Table 4. Thus, the charge carrier mobility values are appropriate and believable. Compared with those without thermal annealing and those subjected to different annealing temperatures, devices after thermal annealing at 120 °C for 10 min showed more balanced hole/electron transport in the active layer, which may account for the relatively higher FF.

Fig. 7
figure 7

J1/2–V curves for hole-only (the first line) and electron-only (the second line) devices based on 2:1 P3HT:BP(DPPB)4 films subjected to different thermal annealing temperatures: a, e as-cast; b, f thermal annealing at 80 °C; c, g thermal annealing at 120 °C; and d, h thermal annealing at 150 °C. All of the samples were maintained at the corresponding temperatures for 10 min

Table 3 Charge carrier mobilities of devices based on P3HT:BP(DPPB)4 films
Table 4 Charge carrier mobilities of small-molecule acceptors based on DPP units

Conclusion

A new nonfullerene electron acceptor, BP(DPPB)4, with a diphenyl core and four diketopyrrolopyrrole arms was designed, synthesized, and characterized. BP(DPPB)4 shows appropriate absorption bands and matched energy levels with P3HT. Moreover, PSCs based on P3HT and BP(DPPB)4 blend films can achieve a moderate PCE of 1.43% with an outstandingly high Voc of 1.17 V after thermal annealing at 120 °C for 10 min. These results demonstrate that the BP unit is a useful fragment to construct multidimensional NFAs and that BP(DPPB)4 is a promising NFA for future practical applications in PSCs.