Ternary non-fullerene polymer solar cells with 13.51% efficiency and a record-high fill factor of 78.13%†

Li Nian‡ *^ab, Yuanyuan Kan‡ ^c, Haitao Wang ^d, Ke Gao ^c, Bo Xu ^c, Qikun Rong ^ab, Rong Wang ^d, Jing Wang ^e, Feng Liu *^e, Junwu Chen *^d, Guofu Zhou *^ab, Thomas P. Russell ^f and Alex K.-Y. Jen *^cg
aGuangdong Provincial Key Laboratory of Optical Information Materials and Technology & Institute of Electronic Paper Displays, South China Academy of Advanced Optoelectronics, South China Normal University, Guangzhou 510006, P. R. China. E-mail: nianli@m.scnu.edu.cn; guofu.zhou@m.scnu.edu.cn
^bNational Center for International Research on Green Optoelectronics, South China Normal University, Guangzhou 510006, P. R. China
^cDepartment of Materials Science and Engineering, University of Washington, Seattle, WA 98195, USA. E-mail: ajen@uw.edu
^dInstitute of Polymer Optoelectronic Materials and Devices, State Key Laboratory of Luminescent Materials and Devices, South China University of Technology, Guangzhou 510640, P. R. China. E-mail: psjwchen@scut.edu.cn
^eDepartment of Physics and Astronomy, and Collaborative Innovation Center of IFSA (CICIFSA), Shanghai Jiaotong University, Shanghai 200240, P. R. China. E-mail: fengliu82@sjtu.edu.cn
^fMaterials Sciences Division, Lawrence Berkeley National Lab, Berkeley, CA 94720, USA
^gDepartment of Chemistry, City University of Hong Kong, Kowloon, Hong Kong, SAR, P. R. China

First published on 27th July 2018

---

---

Broader context Recently, the ternary strategy has been widely used to improve the power conversion efficiency (PCE) of polymer solar cells (PSCs). In ternary PSCs, the short-circuit current density (Jsc) can be enhanced significantly by using a third component with a complementary absorption spectrum. Meanwhile, the open-circuit voltage (Voc) in ternary PSCs can be continuously tuned by varying the compositions. However, the fill factor (FF) is still a bottleneck in ternary PSCs that needs to be improved, especially for those with non-fullerene acceptors (NFAs) due to the more complicated dynamic process. To date, the FFs for state-of-the-art ternary PSCs are mostly below 70%. In this work, a strongly aggregating polymer P1 is added into the classic PBDB-T:IT-M and PBDB-T:ITIC blends. The addition of P1 significantly enhances the crystallization of the blend film, and a clear structure-relationship is illustrated. The ternary devices show significantly improved charge extraction and suppressed charge recombination. As a result, the PCE of the PBDB-T:ITIC based PSC increases from 10.82% to 12.70% and the FF from 71.85% to 78.07% after adding P1. For the PBDB-T:IT-M based PSC, the PCE increases from 11.71% to 13.51% and the FF from 72.07% to 78.13%. The high FFs and PCEs are the best results reported to date for ternary NF PSCs.

Introduction

Polymer solar cells (PSCs) are a promising and cost-effective alternative for future renewable energy due to their unique advantages of low-cost, light-weight and flexible form factors.^1–10 Recently, ternary mixtures, that use either two electron donor polymers or two electron acceptors, have been widely used to elevate the power conversion efficiency (PCE) of PSCs.^11–15 In ternary PSCs, the light harvesting ability of the active layer can be enhanced markedly by the additional absorber with a complementary absorption spectrum, resulting in much improved short-circuit current density (J_sc).^16–18 High J_sc over 25 mA cm⁻² have been demonstrated in ternary blends.¹⁸ Meanwhile, the open-circuit voltage (V_oc) in ternary PSCs can be continuously tuned by varying the donor or acceptor compositions, rather than being pinned to the lower V_oc of the binary blends.^19–21 By combining experiments and modelling, Kemerink and co-workers demonstrated that the enhanced absorption by introducing an additional component offered modest improvements over binary devices while the increased fill factor (FF) arising from the improved charge transport and recombination may offer more significant contributions to enhanced device performance.²² However, the FF for ternary PSCs is the least controllable among the three parameters since the FF for PSCs can be influenced by many factors, and the third component makes the dynamic process in ternary blends more complex. For many ternary PSCs reported, although the PCEs can be improved due to the increased J_sc, the FFs for the devices often stay unchanged or even reduce when the third component is introduced, due to the formation of energetic traps and disruption of molecular packing.²³ To date, the FFs for state-of-the-art ternary PSCs with PCEs over 10% are mostly below 75% (Table S1, ESI†), leaving ample room for further improvement.

Generally, the FF of a PSC is determined by the competition between charge transport/extraction and recombination. An efficient way to improve these processes simultaneously is to increase the crystallization (but with a reasonable domain size to ensure exciton dissociation) of the active layer.²⁴ Highly crystalline materials/domains show higher charge carrier mobility and better domain purity, which will facilitate the charge extraction and reduce charge carrier recombination in the active layer. For instance, by using a highly crystalline polymer (PffBT4T-2OD and PTPD3T) or a highly crystalline non-fullerene acceptor (IDTN) with a corresponding fullerene acceptor or polymer donor, high FFs of 77%, 79.6% and 78% were achieved in binary PSCs, respectively.^24–26 In addition, commonly used device optimization methods to improve the FF, such as thermal annealing, solvent–vapor annealing and solvent additive, are also aiming at optimizing the crystallization or ordered molecular packing in the active layer.^27–29 Hence, the development/selection of a suitable third component to improve the process of crystallization in the active layer while maintaining proper sized phase separation is of great importance for further FF and PCE breakthroughs in ternary PSCs.

In this work, we demonstrate two highly efficient ternary non-fullerene (NF) PSCs with FFs over 78% and PCEs up to 13.52% and 12.70%, respectively. P1 (also known as PDTfBO-TT), a strongly aggregating polymer with a hole mobility of 0.54 cm² V⁻¹ s⁻¹, is used as the additional donor into PBDB-T:ITIC and PBDB-T:IT-M blends.³⁰ The addition of P1 significantly enhances the crystallization of the blend film while maintaining proper morphology. The ternary devices show highly improved charge extraction and suppressed charge recombination in comparison to the binary mixture. As a result, the PCE of the PBDB-T:ITIC based NF PSC increases from 10.82% to 12.70% and the FF from 71.85% to 78.07% after adding P1. For the PBDB-T:IT-M based NF PSC, the PCE increases from 11.71% to 13.52% and the FF from 72.07% to 77.83%. The high FFs and PCEs are the best results reported for ternary NF PSCs to date.

Results and discussion

The chemical structures of PBDB-T, P1, ITIC and IT-M are shown in Fig. 1a. Fig. 1b shows their energy levels reported in previous studies.^30–32 We have employed photoluminescence (PL) spectroscopy to study the possible charge/energy transfer between PBDB-T and P1 by measuring the PL signal of PBDB-T, P1 and PBDB-T:P1 blended films as shown in Fig. S1 (ESI†). The increased PL signal of PBDB-T, concomitant with a quenching of the P1 PL signal in blends, provides evidence for the energy transfer process from P1 to PBDB-T. The energy transfer contributes to the extraction of inner P1 excitons (if a small amount of P1 was blended in PBDB-T), followed by efficient charge separation at the PBDB-T:ITIC or the PBDB-T:IT-M interface.³³Fig. 1c shows the grazing incidence wide-angle X-ray scattering (GIWAXS) of pure P1. Three orders of diffraction characteristic of alkyl-to-alkyl (center-to-center distance) stacking are seen, indicating that P1 is a strongly crystallizing polymer. The strong aggregation property of P1 is also confirmed by the temperature-dependent absorption spectra (Fig. 1d), in which the P1 solution (10⁻⁵ M in chlorobenzene) still showed a shoulder peak at 90 °C.

The BHJ thin film structure order was studied by GIWAXS, with diffraction patterns and line-cut profiles shown in Fig. 2. GIWAXS of the pure films of PBDB-T, ITIC and IT-M is shown in Fig. S2 (ESI†). PBDB-T is a semi-crystalline polymer with well-defined lamellae and π–π stacking. However, when blended with ITIC, even though processed with a DIO additive, a poor structure order is seen. Weak and broad polymer (100) diffraction is observed and enhanced in the out-of-plane direction at 0.29 Å⁻¹. The π–π stacking is observed at 1.73 Å⁻¹. The crystallization of ITIC is fully suppressed by the presence of PBDB-T, indicating a good mixing of these two components. The binary blend P1:ITIC thin film shows quite strong crystallization of the conjugated polymers, showing the third-order diffraction of alkyl-to-alkyl interaction (0.31 Å⁻¹). Its π–π stacking peak is located in the in-plane direction at 1.76 Å⁻¹, coming from P1. Thus, P1 could also retard the crystallization of ITIC, forming mixed regions to facilitate electron transport.

However, when the ternary blend was used, BHJ thin films showed quite prominent ordering of ITIC. With 5% addition of P1 (the mass ratio of P1 in overall donors, similarly hereinafter), a new diffraction peak at 0.50 Å⁻¹ in the out-of-plane direction and 0.37 Å⁻¹ diffraction in the in-plane direction are seen, corresponding to a different packing motif of ITIC. Polymer (100) diffraction remained at 0.29 Å⁻¹, dominated by PBDB-T structure features. The polymer (100) diffraction became sharper and more intense, and thus PBDB-T crystallization and size were enhanced. Quite strong π–π stacking diffraction peaks of ITIC are seen in the out-of-plane direction, located at 1.55 Å⁻¹ and 1.72 Å⁻¹, corresponding to different packing of ITIC. Such a feature can enhance the transport of both electrons and holes, leading to an improved FF in a device. Adding more P1 (10%) in the ternary blend yields similar diffraction features. However, as seen from the π–π stacking region, the peak intensity is reduced when more P1 (10%) is added, but it is still greater than that for the binary blend. Although P1 itself cannot induce face-on ordering of ITIC, as seen from P1:ITIC blend GIWAXS data, its mixture with PBDB-T changed the fundamental physical properties, which allows ITIC to order more readily. Therefore, higher transport can be achieved to improve device performance. When using IT-M as the acceptor in binary and ternary blends, adding 5% of P1 also significantly improves the crystallization, as seen from the sharp increase and narrowing of the PBDB-T (100) peak that has a crystalline correlation length (CCL) of 17.5 nm. The π–π stacking also improves, where the peak position slightly shifts to lower q for the PBDB-T:P1:IT-M blends.

The morphologies of the PBDB-T:P1:ITIC films were studied using atomic force spectroscopy (AFM), transmission electron microscopy (TEM) and resonant soft X-ray scattering (RSoXS) as shown in Fig. 3. As can be seen from the AFM and TEM images, all films show a uniform morphology with some fibrillar textures. From the AFM images, the root-mean-square roughness (RMS) increases from 1.97 nm for the PBDB-T:ITIC blend to 2.55 nm for the 5% ternary blend and 2.18 nm for the 10% ternary blend, which is consistent with the increased ternary blends showing similar morphologies to the binary blend. The AFM phase images are shown in Fig. S3 (ESI†). The phase separation of BHJ thin films was investigated by RSoXS. All BHJ thin films showed quite broad scattering profiles, indicating a multi-length scaled morphology. PBDB-T:ITIC blends showed a small length scale of phase separation at ∼0.006 Å⁻¹, giving a distance of 105 nm by using double tangent lines to estimate peak positions. P1:ITIC blends showed a phase separation of ∼300 nm, which is much larger than the ideal length scale. Ternary blends showed a scattering hump similar to that of PBDB-T:ITIC blends, and thus the small length scale phase separation was locked by PBDB-T:ITIC blends. The 5% ternary blends showed the best scattering intensity, and thus the most pronounced phase separation. The Debye Bucherer model was used to fit the RSoXS data, which yield correlation lengths of 15.2, 14.4, 12.5, and 38.1 nm for PBDB-T:ITIC, PBDB-T:P1(5%):ITIC, PBDB-T:P1(10%):ITIC, and P1:ITIC respectively. Thus, adding P1 into BHJ blends could reduce the size of phase separation. Thus, the improved crystallinity and smaller length scale of phase separation reflects a more optimal morphology for this ternary blend, and an improved device efficiency is expected.

Since the addition of P1 into the binary blend could enhance crystallization while maintaining a proper morphology, it is reasonable to use the ternary blend to fabricate ternary PSCs. The photovoltaic performance of the ternary PSCs was investigated based on the following device structure: ITO/ZnO/PFN/PBDB-T:P1:IT-X (ITIC or IT-M)/MoO₃/Al. The overall donor to acceptor ratio was kept at 1:1 in this study. Poly[(9,9-bis(3′-(N,N-dimethylamino)propyl)-2,7-fluorene)-alt-2,7-(9,9-dioctylfluore ne)] (PFN) was used to improve the device performance and its structure is shown in Fig. S4 (ESI†). The photovoltaic parameters of the devices are shown in Table 1. Fig. 4a illustrates the representative current density versus voltage (J–V) characteristics of devices with different PBDB-T:P1 weight ratios (0%, 5% and 10% P1) under simulated AM 1.5 G illumination at 100 mW cm⁻².

---

The PBDB-T:ITIC binary control devices exhibit an average PCE (PCE_ave) of 10.66 ± 0.10% with a V_oc of 0.90 ± 0.00 V, a J_sc of 16.64 ± 0.16 mA cm⁻² and a FF of 71.24 ± 0.54%. Adding 5% of P1 (the amount of P1 added to the donor, hereinafter) into the PBDB-T:ITIC blend dramatically increases the J_sc to 17.98 ± 0.15 mA cm⁻² and FF to 77.33 ± 0.66%, resulting in a promising PCE_ave of 12.51 ± 0.12%. Further addition of P1 (10%) led to a decrease in the J_sc (17.48 ± 0.17 mA cm⁻²) and FF (75.78 ± 0.72%), but the PCE remained higher (11.92 ± 0.13%) than that of the PBDB-T:ITIC blend. These results show that using a small amount of strongly aggregating polymer to enhance the crystallization of the active layer can dramatically improve the ternary device performance. The P1:ITIC binary device was also fabricated and it showed a low PCE of 3.85 ± 0.37%, due to the much lower J_sc and FF (Fig. S5, ESI†) caused by too large phase separation and surface roughness (Fig. S6, ESI†). For the PBDB-T:IT-M blend, 5% P1 addition shows a similar behaviour to that of the PBDB-T:ITIC blend, enhancing the PCE from 11.57 ± 0.09% to 13.31 ± 0.14% and the FF from 71.46 ± 0.56% to 76.86 ± 0.79%.

External quantum efficiency (EQE) measurements were conducted to confirm the J_sc of the PSCs. As shown in Fig. 4b, the ternary blend based devices show remarkable enhancement in comparison to the corresponding control devices, contributing to the increased J_sc value. The much improved charge carrier collection and reduced recombination contribute to the enhanced EQE values which will be discussed below. The integrated J_sc values from EQE measurements are shown in Table 1, agreeing well with J–V measurements (the deviations are within 3%). To further confirm the effect of 5% P1 addition, the statistical FF distribution histograms of the 5% P1 ternary PSCs based on PBDB-T:ITIC and PBDB-T:IT-M are shown in Fig. 4c. The average FFs are calculated from 20 identical devices prepared from different batches. The J–V curves and photovoltaic parameters of these ternary devices are summarized in Fig. S7 and Tables S2, S3 (ESI†). Both of the ternary systems show FF_max over 78%, which are the highest FF values reported for ternary NF PSCs to date.

To understand the high FF values achieved in the PBDB-T:P1:ITIC ternary blend, the charge transport and extraction properties were investigated initially. The charge carrier mobility was measured by using the space-charge-limited-current (SCLC) method (Fig. S8, ESI†). The hole-only devices and electron-only devices were fabricated by using the device architectures ITO/PEDOT:PSS/active layer/MoO₃/Au and Al/active layer/Ca/Al, respectively. The mobility is extracted from the slopes of J^1/2–V curves by modelling the dark current in the SCLC region. The hole mobilities of the PBDB-T:P1:ITIC = (1:0:1), (0.95:0.05:1) and (0.9:0.1:1) devices are 1.8 × 10⁻⁴, 4.7 × 10⁻⁴ and 3.7 × 10⁻⁴ cm² V⁻¹ s⁻¹, respectively, while the electron mobilities of them are 2.3 × 10⁻⁴, 6.4 × 10⁻⁴ and 4.9 × 10⁻⁴ cm² V⁻¹ s⁻¹, respectively. The increased crystallinity arising from the P1 addition contributed to the much increased hole and electron mobilities, leading to a better FF for the device.

Detailed information about the charge extraction process in devices was probed by using transient photocurrent (TPC) measurements (Fig. 5a). The charge extraction time decreased from 0.29 μs for the PBDB-T:ITIC blend to 0.13 μs for the 5% ternary blend and 0.18 μs for the 10% ternary blend. The reduced charge extraction time in the ternary device indicates more efficient charge extraction from the active layer, which is beneficial to the improvement in J_sc and FF. The charge recombination dynamics in devices were probed using transient photovoltage (TPV) measurements. Fig. S9 (ESI†) shows the typical results of TPV measurements for PBDB-T:P1:ITIC = 1:0:1, 0.95:0.05:1 and 0.9:0.1:1 devices. The extracted charge carrier lifetime increased from 2.04 μs for the PBDB-T:ITIC device to 4.18 μs for the PBDB-T:P1:ITIC (5% P1) device and 3.36 μs for the PBDB-T:P1:ITIC (10% P1) device. The observed longer carrier lifetime suggested a reduced bimolecular recombination loss in the P1 addition devices, which is beneficial to the FF and J_sc values for the ternary device.

To gain more insight into the exciton dissociation and charge extraction, we measured the photocurrent density (J_ph) as a function of the effective voltage (V_eff) as shown in Fig. 5b.³⁴J_ph = J_L − J_D, where J_L and J_D are the current densities under illumination and in the dark, respectively. V_eff = V₀ − V_A, where V₀ is the zero J_ph voltage and V_A is the applied bias voltage. At high V_eff (3.2 V in this study), it is assumed that all of the generated excitons are dissociated and collected at the electrodes, resulting in a saturation current (J_sat). The charge collection probability P(E,T) is determined by normalizing J_ph with J_sat. The P(E,T) values under the short circuit condition for PBDB-T:P1:ITIC = (1:0:1), (0.95:0.05:1) and (0.9:0.1:1) devices are 88.85%, 96.27% and 93.46%, respectively. More interestingly, P(E,T) at the maximum power point (M_pp) dramatically increases from 75.46% to 89.03% (5% P1) and to 84.80% (10% P1). The higher P(E,T) value for the ternary blend device again suggests that P1 addition is efficient in promoting charge extraction, leading to an increased J_sc and a record-high FF > 78% for the ternary PSCs. The enhanced crystallization and charge carrier mobilities in the ternary blend device contribute to the increased charge extraction.

We next analysed the charge recombination behaviour of the binary and ternary devices to understand the high FF by measuring J_sc and V_oc as a function of the light intensity (P_light). At V_oc conditions, all of the charge carriers recombine inside the active layer and the recombination mechanism can be studied by analysing the slope of V_ocversus the natural logarithm of P_light.³⁵ For a bimolecular recombination system, the slope is close to k_BT/q (k_B is Boltzmann's constant, T is absolute temperature and q is the elementary charge). For a trap-assisted recombination system, V_oc shows enhanced dependence on P_light (2k_BT/q). As shown in Fig. 5c, the slope decreases from 1.26k_BT/q for the PBDB-T:ITIC control device to 1.03k_BT/q for the 5% ternary device and to 1.10 k_BT/q for the 10% ternary device, indicating that P1 addition suppresses trap-assisted recombination. Decreased trap-assisted recombination may arise from a decrease in the trap density due to the enhanced crystallinity of the active layer.

Since bimolecular recombination is the dominant charge loss mechanism (the slopes are close to k_BT/q rather than 2k_BT/q as mentioned above), we analysed the relationship between J_sc and P_light to evaluate the degree of bimolecular recombination in the devices.³⁶ In PSCs, J_sc has a power-law dependence on light intensity, J_sc ∝ (P_light)^S. For a device with weak bimolecular recombination, the S value is close to 1. Fig. 5d shows the J_sc ∼ P_light relationship for PBDB-T:P1:ITIC = (1:0:1), (0.95:0.05:1) and (0.9:0.1:1) devices. The extracted S values are 0.960, 0.999 and 0.984, respectively, indicating that bimolecular recombination is highly restrained in the devices with 5% P1 content. Therefore, both trap-assisted recombination and bimolecular recombination are highly suppressed in the 5% ternary devices, which are dominant factors contributing to the state-of-the-art FF values.

Fig. 5e shows the relationship between the FF and J_sc as a function of the normalized area of the (010) reflection for PBDB-T:P1:ITIC devices. The normalized (010) peak area for 5% P1 and 10% P1 ternary devices is 54% and 36% greater than that for the PBDB-T:ITIC binary blend. The FF and J_sc values correlate well with the (010) peak area, which is consistent with the findings of Hou and co-workers.³⁷ The results shown here show that the strategy of using a strongly aggregating polymer to enhance the crystallization of the active layer can remarkably improve the device performance.

To prove the general applicability of our method, we have used two other strongly aggregating polymers, P2 and P3 (also known as PDTfBO-DT and PDTfBO-T) as the additional donor into the PBDB-T:ITIC blend. The chemical structures of P1, P2, and P3 are shown in Fig. S10 (ESI†). As shown in our previously study,³⁰ P1 and P2 showed similar aggregation ability while P3 shows much stronger aggregation ability compared to P1 and P2, deduced from their de-aggregation behaviors during the heating process. Fig. S10 (ESI†) illustrates the representative J–V characteristics of devices based on PBDB-T:PX (P1, P2 or P3):ITIC under simulated AM 1.5 G illumination at 100 mW cm⁻². The mass ratio of PX in each donor is kept at 5%. The best photovoltaic parameters of the devices are shown in Table S4 (ESI†). Although the device with P1 addition gives the best photovoltaic performance, P2 addition also leads to greatly improved performance. However, the addition of P3 decreases the J_sc and FF in the device, which may be due to the too strong aggregation nature of P3. These results showed that using strongly aggregating polymers as the third component could be an effective method to improve the device performance and the aggregation ability of the blend should be well-tuned to optimize the improvement.

Conclusions

In conclusion, high efficiency ternary non-fullerene solar cells were fabricated by adding a small amount of P1 into PBDB-T:ITIC and PBDB-T:IT-M mixtures. The small portion of added P1 does not change the length scale of the phase separation of PBDB-T:ITIC, but the ordering is much improved and the donor and acceptors showed enhancement in the hole and electron mobilities. More interestingly, such a modification leads to an impressive improvement in the device FF to over 78%. Enhanced charge transport also contributes to elevated charge collection, and the short circuit current density is markedly improved. Impressive 12.70% PCE and 78.07% FF could be obtained for a traditional non-fullerene PBDB-T:ITIC mixture and 13.52% PCE and 77.83% FF could be achieved in the PBDB-T:IT-M blend, which are the highest reported to date, showing the importance of fine-tuning the crystallization of ternary blend films.

Author contributions

L. N. conceived the idea. Device fabrication and characterization were carried out by L. N. The design of experiments, discussion of results and manuscript writing were done by L. N., F. L., J. C., G. Z., T. P. R. and A. K.-Y. J. The donor of P1 was synthesized by H. W. Y. K., K. G., B. X., Q. R., R. W. and J. W. were contributed to the device measurement and analysis of data. All the authors discussed the results and contributed to the writing of the manuscript.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was financially supported by National Key R&D Program of China (2016YFB0401501), National Natural Science Foundation of China (51561135014, U1501244). Program for Changjiang Scholars and Innovative Research Teams in Universities (No. IRT_17R40), Guangdong Innovative Research Team Program (No. 2013C102), Guangdong Provincial Key Laboratory of Optical Information Materials and Technology (Grant No. 2017B030301007), MOE International Laboratory for Optical Information Technologies and the 111 project. A. K.-Y. J. acknowledges financial support from the Asian Office of Aerospace R&D (FA2386-15-1-4106) and the Office of Naval Research (N00014-17-1-2201, N00014-14-1-0246, N00014-17-1-2260). J. C. acknowledges financial support from National Natural Science Foundation of China (U1401244) and the National Basic Research Program of China (973 program 2014CB643505). F. L. was supported by the Young 1000 Talents Global Recruitment Program of China. T. P. R were supported by the U.S. Office of Naval Research under contract N00014-17-1-2244. Portions of this research were carried out at beamline 7.3.3 and 11.0.1.2 at the Advanced Light Source, and Molecular Foundry, Lawrence Berkeley National Laboratory, which was supported by the DOE, Office of Science, and Office of Basic Energy Sciences.

Notes and references

1. L. Dou, J. You, J. Yang, C.-C. Chen, Y. He, S. Murase, T. Moriarty, K. Emery, G. Li and Y. Yang, Nat. Photonics, 2012, 6, 180 CrossRef .
2. Y. Cui, H. Yao, B. Gao, Y. Qin, S. Zhang, B. Yang, C. He, B. Xu and J. Hou, J. Am. Chem. Soc., 2017, 139, 7302 CrossRef PubMed .
3. J. Zhao, Y. Li, G. Yang, K. Jiang, H. Lin, H. Ade, W. Ma and H. Yan, Nat. Energy, 2016, 1, 15027 CrossRef .
4. L. Nian, W. Zhang, N. Zhu, L. Liu, Z. Xie, H. Wu, F. Würthner and Y. Ma, J. Am. Chem. Soc., 2015, 137, 6995 CrossRef PubMed .
5. L. H. Rossander, H. F. Dam, J. E. Carlé, M. Helgesen, L. Rajkovic, M. Corazza, F. C. Krebs and J. W. Andreasen, Energy Environ. Sci., 2017, 10, 2411 RSC .
6. J. Wan, X. Xu, G. Zhang, Y. Li, K. Feng and Q. Peng, Energy Environ. Sci., 2017, 10, 1739 RSC .
7. L. Nian, Z. Chen, S. Herbst, Q. Li, C. Yu, X. Jiang, H. Dong, F. Li, L. Liu, F. Würthner, J. Chen, Z. Xie and Y. Ma, Adv. Mater., 2016, 28, 7521 CrossRef PubMed .
8. S.-J. Ko, Q. V. Hoang, C. E. Song, M. A. Uddin, E. Lim, S. Y. Park, B. H. Lee, S. Song, S.-J. Moon, S. Hwang, P.-Q. Morin, M. Leclerc, G. M. Su, M. L. Chabinyc, H. Y. Woo, W. S. Shin and J. Y. Kim, Energy Environ. Sci., 2017, 10, 1443 RSC .
9. L. Nian, K. Gao, Y. Jiang, Q. Rong, X. Hu, D. Yuan, F. Liu, X. Peng, T. P. Russell and G. Zhou, Adv. Mater., 2017, 29, 1700616 CrossRef PubMed .
10. A. Kuzmich, D. Padula, H. Ma and A. Troisi, Energy Environ. Sci., 2017, 10, 395 RSC .
11. N. Gasparini, X. Jiao, T. Heumueller, D. Baran, G. J. Matt, S. Fladischer, E. Spiecker, H. Ade, C. J. Brabec and T. Ameri, Nat. Energy, 2016, 1, 16118 CrossRef .
12. L. Nian, K. Gao, F. Liu, Y. Kan, X. Jiang, L. Liu, Z. Xie, X. Peng, T. P. Russell and Y. Ma, Adv. Mater., 2016, 28, 8184 CrossRef PubMed .
13. D. Baran, R. S. Ashraf, D. A. Hanifi, M. Abdelsamie, N. Gasparini, J. A. Röhr, S. Holliday, A. Wadsworth, S. Lockett, M. Neophytou, C. J. M. Emmott, J. Nelson, C. J. Brabec, A. Amassian, A. Salleo, T. Kirchartz, J. R. Durrant and I. McCulloch, Nat. Mater., 2017, 16, 363 CrossRef PubMed .
14. M. Zhang, W. Gao, F. Zhang, Y. Mi, W. Wang, Q. An, J. Wang, X. Ma, J. Miao, Z. Hu, X. Liu, J. Zhang and C. Yang, Energy Environ. Sci., 2018, 11, 841 RSC .
15. L. Lu, M. A. Kelly, W. You and L. Yu, Nat. Photonics, 2015, 9, 491 CrossRef .
16. X. Xu, Z. Bi, W. Ma, Z. Wang, W. C. H. Choy, W. Wu, G. Zhang, Y. Li and Q. Peng, Adv. Mater., 2017, 29, 1704271 CrossRef PubMed .
17. W. Jiang, R. Yu, Z. Liu, R. Peng, D. Mi, L. Hong, Q. Wei, J. Hou, Y. Kuang and Z. Ge, Adv. Mater., 2017, 29, 1703005 Search PubMed .
18. H. Yao, Y. Cui, R. Yu, B. Gao, H. Zhang and J. Hou, Angew. Chem., Int. Ed., 2017, 56, 3045 CrossRef PubMed .
19. P. P. Khlyabich, A. E. Rudenko, B. C. Thompson and Y.-L. Loo, Adv. Funct. Mater., 2015, 25, 5557 CrossRef .
20. R. A. Street, D. Davies, P. P. Khlyabich, B. Burkhart and B. C. Thompson, J. Am. Chem. Soc., 2013, 135, 986 CrossRef PubMed .
21. P. P. Khlyabich, B. Burkhart and B. C. Thompson, J. Am. Chem. Soc., 2012, 134, 9074 CrossRef PubMed .
22. N. Felekidis, E. Wang and M. Kemerink, Energy Environ. Sci., 2016, 9, 257 RSC .
23. W. Huang, P. Cheng, Y. Yang, G. Li and Y. Yang, Adv. Mater., 2018, 30, 1705706 CrossRef PubMed .
24. Y. Liu, J. Zhao, Z. Li, C. Mu, W. Ma, H. Hu, K. Jiang, H. Lin, H. Ade and H. Yan, Nat. Commun., 2014, 5, 5293 CrossRef PubMed .
25. X. Guo, N. Zhou, S. J. Lou, J. Smith, D. B. Tice, J. W. Hennek, R. P. Ortiz, J. T. L. Navarrete, S. Li, J. Strzalka, L. X. Chen, R. P. H. Chang, A. Facchetti and T. J. Marks, Nat. Photonics, 2013, 7, 825 CrossRef .
26. S. Li, L. Ye, W. Zhao, X. Liu, J. Zhu, H. Ade and J. Hou, Adv. Mater., 2017, 29, 1704051 CrossRef PubMed .
27. W. Ma, C. Yang, X. Gong, K. Lee and A. J. Heeger, Adv. Funct. Mater., 2005, 15, 1617 CrossRef .
28. K. Sun, Z. Xiao, E. Hanssen, M. F. G. Klein, H. H. Gam, M. Pfaff, D. Gerthsen, W. W. H. Wong and D. J. Jones, J. Mater. Chem. A, 2014, 2, 9048 RSC .
29. X. Guo, C. Cui, M. Zhang, L. Huo, Y. Huang, J. Hou and Y. Li, Energy Environ. Sci., 2012, 5, 7943 RSC .
30. H. Wang, Y. Zhu, Z. Liu, L. Zhang, J. Chen and Y. Cao, Org. Electron., 2016, 31, 1 CrossRef .
31. W. Zhao, S. Li, S. Zhang, X. Liu and J. Hou, Adv. Mater., 2016, 28, 1604059 Search PubMed .
32. S. Li, L. Ye, W. Zhao, S. Zhang, S. Mukherjee, H. Ade and J. Hou, Adv. Mater., 2016, 28, 9423 CrossRef PubMed .
33. V. Gupta, V. Bharti, M. Kumar, S. Chand and A. J. Heeger, Adv. Mater., 2015, 27, 4398 CrossRef PubMed .
34. L. Lu, Z. Luo, T. Xu and L. Yu, Nano Lett., 2012, 13, 59 CrossRef PubMed .
35. L. J. A. Koster, V. D. Mihailetchi, R. Ramaker and P. W. M. Blom, Appl. Phys. Lett., 2005, 86, 123509 CrossRef .
36. I. Riedel, J. Parisi, V. Dyakonov, L. Lutsen, D. Vanderzande and J. C. Hummelen, Adv. Funct. Mater., 2004, 14, 38 CrossRef .
37. Z. Zheng, O. M. Awartani, B. Gautam, D. Liu, Y. Qin, W. Li, A. Bataller, K. Gundogdu, H. Ade and J. Hou, Adv. Mater., 2016, 28, 1604241 Search PubMed .

---

Footnotes

† Electronic supplementary information (ESI) available: Experimental details, GIWAXS, J–V curves, EQE spectra, AFM images, and SCLC measurements. See DOI: 10.1039/c8ee01564c

‡ These authors contributed equally to this work.

This journal is © The Royal Society of Chemistry 2018