# SMALL-MOLECULE SEMICONDUCTORS FOR HIGH-EFFICIENCY ORGANIC SOLAR CELLS

EDITED BY : Chuanlang Zhan and Donghong Yu PUBLISHED IN : Frontiers in Chemistry

#### Frontiers Copyright Statement

© Copyright 2007-2019 Frontiers Media SA. All rights reserved. All content included on this site, such as text, graphics, logos, button icons, images, video/audio clips, downloads, data compilations and software, is the property of or is licensed to Frontiers Media SA ("Frontiers") or its licensees and/or subcontractors. The copyright in the text of individual articles is the property of their respective authors, subject to a license granted to Frontiers.

The compilation of articles constituting this e-book, wherever published, as well as the compilation of all other content on this site, is the exclusive property of Frontiers. For the conditions for downloading and copying of e-books from Frontiers' website, please see the Terms for Website Use. If purchasing Frontiers e-books from other websites or sources, the conditions of the website concerned apply.

Images and graphics not forming part of user-contributed materials may not be downloaded or copied without permission.

Individual articles may be downloaded and reproduced in accordance with the principles of the CC-BY licence subject to any copyright or other notices. They may not be re-sold as an e-book.

As author or other contributor you grant a CC-BY licence to others to reproduce your articles, including any graphics and third-party materials supplied by you, in accordance with the Conditions for Website Use and subject to any copyright notices which you include in connection with your articles and materials.

All copyright, and all rights therein, are protected by national and international copyright laws.

The above represents a summary only. For the full conditions see the Conditions for Authors and the Conditions for Website Use. ISSN 1664-8714 ISBN 978-2-88945-980-3 DOI 10.3389/978-2-88945-980-3

#### About Frontiers

Frontiers is more than just an open-access publisher of scholarly articles: it is a pioneering approach to the world of academia, radically improving the way scholarly research is managed. The grand vision of Frontiers is a world where all people have an equal opportunity to seek, share and generate knowledge. Frontiers provides immediate and permanent online open access to all its publications, but this alone is not enough to realize our grand goals.

#### Frontiers Journal Series

The Frontiers Journal Series is a multi-tier and interdisciplinary set of open-access, online journals, promising a paradigm shift from the current review, selection and dissemination processes in academic publishing. All Frontiers journals are driven by researchers for researchers; therefore, they constitute a service to the scholarly community. At the same time, the Frontiers Journal Series operates on a revolutionary invention, the tiered publishing system, initially addressing specific communities of scholars, and gradually climbing up to broader public understanding, thus serving the interests of the lay society, too.

#### Dedication to Quality

Each Frontiers article is a landmark of the highest quality, thanks to genuinely collaborative interactions between authors and review editors, who include some of the world's best academicians. Research must be certified by peers before entering a stream of knowledge that may eventually reach the public - and shape society; therefore, Frontiers only applies the most rigorous and unbiased reviews.

Frontiers revolutionizes research publishing by freely delivering the most outstanding research, evaluated with no bias from both the academic and social point of view. By applying the most advanced information technologies, Frontiers is catapulting scholarly publishing into a new generation.

#### What are Frontiers Research Topics?

Frontiers Research Topics are very popular trademarks of the Frontiers Journals Series: they are collections of at least ten articles, all centered on a particular subject. With their unique mix of varied contributions from Original Research to Review Articles, Frontiers Research Topics unify the most influential researchers, the latest key findings and historical advances in a hot research area! Find out more on how to host your own Frontiers Research Topic or contribute to one as an author by contacting the Frontiers Editorial Office: researchtopics@frontiersin.org

# SMALL-MOLECULE SEMICONDUCTORS FOR HIGH-EFFICIENCY ORGANIC SOLAR CELLS

Topic Editors: Chuanlang Zhan, Institute of Chemistry (CAS), China Donghong Yu, Aalborg University, Denmark

Citation: Zhan, C., Yu, D., eds. (2019). Small-Molecule Semiconductors for High-Efficiency Organic Solar Cells. Lausanne: Frontiers Media. doi: 10.3389/978-2-88945-980-3

# Table of Contents

*05 Comparison Study of Wide Bandgap Polymer (PBDB-T) and Narrow Bandgap Polymer (PBDTTT-EFT) as Donor for Perylene Diimide Based Polymer Solar Cells*

Tengling Ye, Shan Jin, Cong Kang, Changhao Tian, Xin Zhang, Chuanlang Zhan, Shirong Lu and Zhipeng Kan


Chaohua Cui

*40 Dithienonaphthalene-Based Non-fullerene Acceptors With Different Bandgaps for Organic Solar Cells*

Meiqi Zhang, Yunlong Ma and Qingdong Zheng

*49 Small-Molecule Electron Acceptors for Efficient Non-fullerene Organic Solar Cells*

Zhenzhen Zhang, Jun Yuan, Qingya Wei and Yingping Zou

*71 Effects of Alkoxy and Fluorine Atom Substitution of Donor Molecules on the Morphology and Photovoltaic Performance of all Small Molecule Organic Solar Cells*

Beibei Qiu, Shanshan Chen, Lingwei Xue, Chenkai Sun, Xiaojun Li, Zhi-Guo Zhang, Changduk Yang and Yongfang Li

*80 Comparison of the Solution and Vacuum-Processed Squaraine: Fullerene Small-Molecule Bulk Heterojunction Solar Cells*

Guo Chen, Zhitian Ling, Bin Wei, Jianhua Zhang, Ziruo Hong, Hisahiro Sasabe and Junji Kido

*89 Urea-Doped ZnO Films as the Electron Transport Layer for High Efficiency Inverted Polymer Solar Cells*

Zongtao Wang, Zhongqiang Wang, Ruqin Zhang, Kunpeng Guo, Yuezhen Wu, Hua Wang, Yuying Hao and Guo Chen


Ruimin Zhou, Benzheng Xia, Huan Li, Zhen Wang, Yang Yang, Jianqi Zhang, Bo W. Laursen, Kun Lu and Zhixiang Wei

*117 Constructing Desired Vertical Component Distribution Within a PBDB-T:ITIC-M Photoactive Layer via Fine-Tuning the Surface Free Energy of a Titanium Chelate Cathode Buffer Layer* Yiming Bai, Bo Yang, Xiaohan Chen, Fuzhi Wang, Tasawar Hayat, Ahmed Alsaedi and Zhan'ao Tan *127 BN Embedded Polycyclic* p*-Conjugated Systems: Synthesis, Optoelectronic Properties, and Photovoltaic Applications* Jianhua Huang and Yuqing Li *149 Efficient Non-fullerene Organic Solar Cells Enabled by Sequential Fluorination of Small-Molecule Electron Acceptors* Ruihao Xie, Lei Ying, Hailong Liao, Zhongxin Chen, Fei Huang and Yong Cao *158 Two Novel Small Molecule Donors and the Applications in Bulk-Heterojunction Solar Cells* Xin Qi, Yuan-Chih Lo, Yifan Zhao, Liyang Xuan, Hao-Chun Ting, Ken-Tsung Wong, Mostafizur Rahaman, Zhijian Chen, Lixin Xiao and Bo Qu *166 Insight Into the Role of PC71BM on Enhancing the Photovoltaic Performance of Ternary Organic Solar Cells* Bei Wang, Yingying Fu, Chi Yan, Rui Zhang, Qingqing Yang, Yanchun Han and Zhiyuan Xie *174 Utilizing Benzotriazole and Indacenodithiophene Units to Construct Both Polymeric Donor and Small Molecular Acceptors to Realize Organic Solar Cells With High Open-Circuit Voltages Beyond 1.2 V*

Ailing Tang, Fan Chen, Bo Xiao, Jing Yang, Jianfeng Li, Xiaochen Wang and Erjun Zhou

## Comparison Study of Wide Bandgap Polymer (PBDB-T) and Narrow Bandgap Polymer (PBDTTT-EFT) as Donor for Perylene Diimide Based Polymer Solar Cells

Tengling Ye<sup>1</sup> \*, Shan Jin<sup>1</sup> , Cong Kang<sup>1</sup> , Changhao Tian<sup>1</sup> , Xin Zhang<sup>3</sup> , Chuanlang Zhan<sup>3</sup> \*, Shirong Lu<sup>2</sup> and Zhipeng Kan<sup>2</sup> \*

#### Edited by:

*Iwao Ojima, Stony Brook University, United States*

#### Reviewed by:

*Gregory C. Welch, University of Calgary, Canada Liming Ding, National Center for Nanoscience and Technology (CAS), China Jonathan G. Rudick, Stony Brook University, United States*

#### \*Correspondence:

*Tengling Ye ytl@hit.edu.cn Zhipeng Kan kanzhipeng@cigit.ac.cn Chuanlang Zhan clzhan@iccas.ac.cn*

#### Specialty section:

*This article was submitted to Organic Chemistry, a section of the journal Frontiers in Chemistry*

Received: *18 June 2018* Accepted: *28 November 2018* Published: *10 December 2018*

#### Citation:

*Ye T, Jin S, Kang C, Tian C, Zhang X, Zhan C, Lu S and Kan Z (2018) Comparison Study of Wide Bandgap Polymer (PBDB-T) and Narrow Bandgap Polymer (PBDTTT-EFT) as Donor for Perylene Diimide Based Polymer Solar Cells. Front. Chem. 6:613. doi: 10.3389/fchem.2018.00613* *<sup>1</sup> MIIT Key Laboratory of Critical Materials Technology for New Energy Conversion and Storage, School of Chemistry and Chemical Engineering, Harbin Institute of Technology, Harbin, China, <sup>2</sup> Organic Semiconductor Research Center, Chongqing Institute of Green and Intelligent Technology, Chinese Academy of Sciences, Chongqing, China, <sup>3</sup> Beijing National Laboratory for Molecular Sciences, CAS Key Laboratory of Photochemistry, Institute of Chemistry, Chinese Academy of Sciences, Beijing, China*

Perylene diimide (PDI) derivatives as a kind of promising non-fullerene-based acceptor (NFA) have got rapid development. However, most of the relevant developmental work has focused on synthesizing novel PDI-based structures, and few paid attentions to the selection of the polymer donor in PDI-based solar cells. Wide bandgap polymer (PBDB-T) and narrow bandgap polymer (PBDTTT-EFT) are known as the most efficient polymer donors in polymer solar cells (PSCs). While PBDB-T is in favor with non-fullerene acceptors achieving power conversion efficiency (PCE) more than 12%, PBDTTT-EFT is one of the best electron donors with fullerene acceptors with PCE up to 10%. Despite the different absorption profiles, the working principle of these benchmark polymer donors with a same electron acceptor, specially PDI-based acceptors, was rarely compared. To this end, we used PBDB-T and PBDTTT-EFT as the electron donors, and 1,1′ -bis(2-methoxyethoxyl)-7,7′ -(2,5-thienyl) bis-PDI (Bis-PDI-T-EG) as the electron acceptor to fabricate PSCs, and systematically compared their differences in device performance, carrier mobility, recombination mechanism, and film morphology.

Keywords: polymer solar cells, perylene diimide, charge transport, charge recombination, non-fullerene, wide bandgap, narrow bandgap

#### INTRODUCTION

Polymer solar cells (PSCs) have been attracting more and more attention in both research and industrial applications due to their unique properties such as solution processability, light weight, low cost, and high mechanical flexibility. High-efficiency PSCs typically employ the bulk heterojunction structures composed of a p-type polymer material and an n-type small molecular material. The most studied combination was a polymer semiconductor donor and a fullerenebased small molecule acceptor, reaching a power conversion efficiency (PCE) of 11.7% (Zhao et al., 2016a). Due to the drawbacks of fullerene-based materials, such as weak absorption in visible and near infrared region, limited tunability of energy levels, and poor morphology stability, further improvement of the PCE in such traditional donor/acceptor combinations is hindered (Eftaiha et al., 2014; Cheng and Zhan, 2016). To overcome these limiting factors, efficient non-fullerene-based acceptors (NFA) have been developed for decades. Yuze Lin (Lin et al., 2015a,b) and co-authors reported a series of NFAs based on highly electron-deficient (3-ethylhexyl-4-oxothiazolidine-2-yl) dimalononitrile (RCN) groups, such as ITIC IEIC and SFBRCN, getting PCE comparable to their fullerene counterparts and opening a new era of PSCs. After that, the first single cell with PCE more than 11% based on polymer/NFA was reported by Wenchao Zhao etc. (Zhao et al., 2016b). The RCN-based NFAs has been also introduced into ternary PSCs and tandem PSCs. The best PCE of single PSCs based on these acceptors has already exceeded 14% up to date (Xiao et al., 2017a; Li et al., 2018) and the optimal efficiency of tandem organic solar cells is up to 17.3% (Cui et al., 2018; Meng et al., 2018). The perylene diimide (PDI)-based small molecules as another kind of promising NFAs have been intensively studied. However, the development of PDI-based NFAs in efficiency is lagging. The main factor preventing PDI-based NFAs to get higher efficiency is that PDI moleculars have the intrinsic tendency to aggregate in solid thin film, where excimers are formed and the process of exciton diffusion/separation is severely limited (Ye et al., 2013). To overcome the aggregation, the design of twist PDI dimer derivatives linked at the imide positions or bay positions were designed. Bis-PDI-T-EG is an example of PDI dimmer linked by a thiophene group at bay position with a bandgap of 1.81 eV. A PCE of 6.1% was achieved when it was blended with PBDTTT-C-T by finely tuning the active layer morphology (Zhang et al., 2013, 2015). Recently, various linkers in PDI–π-linker–PDI type systems were reported, giving PCEs up to 9.5% (Liu et al., 2016; McAfee et al., 2017; Welsh et al., 2018a,b). In addition, the introduction of annulation to PDI molecular is another strategy to significantly improve performance giving PCEs in the range of 7–8% (Sun et al., 2015; Hendsbee et al., 2016; Meng et al., 2016; Dayneko et al., 2018). Although a great amount of work has been done in developing new PDI derivatives to restrict their intrinsic aggregation tendency, little work paid attention to the effect of the polymer donor selection on PDI-based on PSCs. PBDB-T, also named as PCE12, worked modest with PCBM, but performed amazingly high PCE with NFAs (Zhao et al., 2016b). PBDTTT-EFT, also named as PCE10 was found to be efficient with both fullerene-based acceptors and NFAs (Chen et al., 2015; Zhang et al., 2017). Despite the different absorption profiles, the working principle of these benchmark polymer donors with one PDI acceptor was rarely compared. To this end, we compared the PSCs made with PCE12 and PCE10 as the electron donors, and 1,1′ -bis(2-methoxyethoxyl)-7,7′ -(2,5-thienyl) bis-PDI (Bis-PDI-T-EG) as the electron acceptor. If not otherwise mentioned, we will refer PCE10 to PBDTTT-EFT, PCE12 to PBDB-T, and PDI to Bis-PDI-T-EG. The chemical structures of the PCE10, PCE12 and Bis-PDI-T-EG were shown in **Figure 1A**. Herein, we compared their differences in device performance, charge carrier mobility, recombination mechanism, and film morphology. The PSCs of PCE12/PDI and PCE10/PDI can give a PCE of 3 and 5.3%, respectively, both with FF about ∼50–60%. The hole mobilities of both devices were similar, 3.4 × 10−<sup>4</sup> and 6.4 × 10−<sup>4</sup> cm<sup>2</sup> /V s for PCE12/PDI and PCE10/PDI, respectively. However, the electron mobilities were 2.3 × 10−<sup>6</sup> and 1.2×10−<sup>5</sup> cm<sup>2</sup> /V s, which were much lower than the hole mobilities. The low and unbalanced charge carrier mobilities should be one of the limiting factors for the low FF of these PSCs. By combining photoluminescence (PL) quenching efficiency, light intensity dependent J-V measurements, transient photocurrent and transient photovoltage, we systematically studied the recombination profiles in two systems. Both systems showed a similar extent of bimolecular recombination, while the PCE10/PDI device suffered a severe trap-assisted recombination. The high electron mobility should be the key factor for efficient charge extraction and thus the high performance in PCE10/PDI devices. The mobility was resulted from the morphological difference, i.e., the distinct aggregation and phase separation in the blends. The results indicate that it is important to examine the donor and acceptor aggregation nature when we make choice of donors for Bis-PDI-T-EG based PSCs.

### RESULTS AND DISCUSSION

## Optical Properties and Photovoltaic Device Performance

**Figure 1B** shows the normalized UV-Vis absorption spectra of PCE10, PCE12, Bis-PDI-T-EG and their blend films. In the spectra, there are two distinguished features: (1) the absorption spectra of PCE12 and PDI substantially overlap with each other in the visible range from 450 to 700 nm; (2) the absorption spectra of PCE10 and PDI are well complementary to each other, covering the wavelength range from 300 to 800 nm. Thus, the PCE10/PDI blend film has the potential to harvest more photon energy compared with that of PCE12/PDI, which may result in higher short circuit current (JSC) in the devices (Xiao et al., 2017b).

The lowest unoccupied molecular orbital (LUMO) and the highest occupied molecular orbital (HOMO) offsets between the donor and the acceptor are shown in **Figure 1C** (Zhang et al., 2014, 2015; Zhao et al., 2016b). It is noticeable that the existing energy offsets for electron transfer from donor to acceptor and the hole transfer from acceptor to donor are sufficient in both PCE10/PDI and PCE12/PDI systems. We also notice the HOMO energy difference in PCE10 and PCE12, i.e., the HOMO of PCE12 is about 0.1 eV deeper compared with that of PCE10, which may lead to a higher device open circuit voltage (VOC). Proper device performance is expected for both polymer/PDI blends.

We fabricated solar cells in inverted device architecture to check the photovoltaic properties of the two blends. The device performance of optimized PSCs are listed in **Table 1**, and the optimal current density-voltage (J–V) curves and the corresponding external quantum efficiency (EQE) spectra are displayed in **Figures 1D,E**. As shown in **Table 1**, the device made with PCE12/PDI can yield PCEs of 3% in average, with a modest JSC and FF value of 6.15 mA cm−<sup>2</sup> and 52%, respectively. As predicted from the energy alignment, the VOC is quite high with a value of 0.94 V. On the other hand, the device made with PCE10/PDI can yield markedly better PCEs of 5.3% in average,

diagrams, (D) J–V characteristics and (E) EQE spectra of the PSCs based on PCE10/PDI and PCE12/PDI.

with a JSC value of 10.61 mA cm−<sup>2</sup> , a FF value of 56% and a VOC value of 0.89 V. The apparent difference in the value of JSC is also observed in the EQE spectra plotted in **Figure 1E**. The EQE spectra of devices made with PCE10/PDI cover the range from 300 to 800 nm, while the PCE12/PDI devices only cover the spectra range from 300 to 750 nm. Furthermore, the maximum EQE in PCE10/PDI devices is about 10% higher than that of the PCE12/PDI devices. One should note that the integrated JSC from the EQE spectra was within 5% deviation compared to the one measured under solar simulator.

To further understand the differences in JSC, we performed optical simulation on the maximum JSC in the two blends by considering the internal quantum efficiency (IQE) to be 100% and only relating with absorption. As shown in **Figures S1A,B** (Margulis et al., 2013), we found that the values of JSC,max of TABLE 1 | Device performance of PSCs based on different donors.


*Values were averaged from 10 working devices.*

PCE12/PDI and PCE10/PDI were 12.80 and 15.70 mA cm−<sup>2</sup> at the optimized film thickness and the difference of the calculated Jsc (12.8/15.7 = 0.815) was much closer compared with the measured values (6.15/10.61 = 0.579). The averaged IQE can be estimated as a ratio of measured JSC to maximum theoretical JSC,max obtained as mentioned above by optical simulation. In addition, the IQE can be separated into two contributions which are charge generation efficiency (ηgen) and charge collection (and transport) efficiency (ηcoll) (Benten et al., 2016):

$$\text{IQR} = \frac{\text{J}\_{\text{sc}}}{\text{J}\_{\text{sc,max}}} = \frac{\eta\_{\text{gen}}}{\eta\_{\text{coll}}} \tag{1}$$

The average IQE with value of 48% (PCE12/PDI) and 68% (PCE10/PDI) were obtained. We attribute the larger current density deviation of measured JSC and the non-unity IQE to not only the different absorption but also great related with recombination, and charge transport.



#### Charge Transport and Recombination

Before examining the charge recombination happened in the devices, we first checked the charge transport by space charge limited current (SCLC) model (Blakesley et al., 2014). The measured dark current density was fitted with the following equation and the parameters are listed in **Table 2**:

$$J(V) = \frac{9}{8} \varepsilon\_0 \varepsilon\_r \mu\_0 \exp\left(0.89 \beta \sqrt{\frac{V}{L}}\right) \frac{(V)^2}{L^3} \tag{2}$$

The dark current density and fitting curve are shown in **Figures S2A,B**. The hole mobilities of 3.4 × 10−<sup>4</sup> and 6.4 × 10−<sup>4</sup> cm<sup>2</sup> /V s were obtained for the blends of PCE12/PDI and PCE10/PDI. The electron mobilities of PCE12/PDI and PCE10/PDI were 2.3 × 10−<sup>6</sup> and 1.2 × 10−<sup>5</sup> cm<sup>2</sup> /V s, respectively, which were much lower than the hole mobilities. It is worth noting that the imbalance of hole and electron mobilities is likely the origin of significant space charge build-up in the optimized polymer/PDI solar cells, which in turn limits the photovoltaic performance.

The photoluminescence (PL) quenching efficiency is one of the methods to check whether the donor/acceptor combination may work or not. Low quenching efficiency usually relates to large domain size of the donor and acceptor and can translate to severe geminate recombination and poor exciton dissociation efficiency, resulting in bad device performance (Ye et al., 2013; Liu

et al., 2016). When the donors and acceptor were photoexcited individually, the PL quenching were shown in **Figures 2A–D**. As depicted in the **Figures 2A,B**, PL quenching efficiencies of 95% and 92% were achieved when the donor materials were excited at 680 nm and 620 nm, respectively. The PL quenching efficiencies larger than 90% suggest that the energy loss of geminate recombination in donor materials is secondary. When the acceptor was excited at 550 nm, PL quenching efficiencies of 89% and 77% were obtained as shown in **Figures 2C,D**. The lower PL quenching efficiencies indicate that the energy loss of geminate recombination in the blends is severe when the acceptor was excited. As presented in **Figures 2B,D**, PCE12/PDI system behaves lower PL quenching efficiency and more severe geminate recombination when both the donor material and acceptor were excited (Ye et al., 2013).

Light intensity dependent JSC and VOC were reported as an easy way to probe the recombination patterns in PSCs. In general, the JSC was plotted against incident light intensity in a log-log scale, following a relationship: JSC ∞ I α , whereby 1) α = 1 indicates that all dissociated free carriers are swept out of the device prior to the bimolecular recombination and 2) α < 1 implies a dependence of JSC on bimolecular recombination. As shown in **Figure 3A** and **S3**, the α was fitted with a value of 0.95 (±0.01) and 0.97 (±0.01) for PCE12/PDI and PCE10/PDI devices, respectively, suggesting that there was a certain extent of bimolecular recombination in both devices. The comparable α value means that bimolecular recombination was similar in both devices (Cowan et al., 2010; Koster et al., 2011). Next, we fitted the VOC/incident light intensity in a natural-log scale to a relationship: VOC ∞ nkT/q ln(I), where k, T, and q are the Boltzmann constant, temperature in Kelvin, and the elementary charge, respectively. The parameter n (usually in the range of 1–2) reflects the presence/absence of carrier traps across the active layers or at interfaces with the electrodes. Any deviations from n = 1 (trap-free condition) point to the existence of the effect of trap-assisted recombination. As presented in **Figure 3B** and **S3**, the n-values for PCE12/PDI and PCE10/PDI are 1.12 (±0.02) and 1.22 (±0.10), respectively. The n value larger than 1 indicates that trap-assisted recombination is in both devices and the relatively larger n value of PCE10/PDI implies that trapassisted recombination is more severe in the PCE10/PDI devices (Koster et al., 2005).

To have a deeper understand on the recombination profiles and quantitatively recombination characteristics, we then performed transient photocurrent (TPC) and transient photovoltage (TPV) characteristics (Li et al., 2011). From **Figures S4A,B**, we noticed that in both systems, the current reached to steady state current within 2 µs without current spike, in both systems, resulting from proper charge generation. By integrating the current after the pulse lights off, the total

generated charges can be estimated as discussed later in this part. We further compared the normalized TPC of the two systems at 1 sun condition, and it was found that the current decay in the system of PCE10/PDI was faster than that of the PCE12/PDI system shown in **Figure S6**, implying that the charge extraction in PCE10/PDI was better, in agreement with the results we made on the mobility measurement. While TPC provides the information on charge generation and extraction at short circuit condition, TPV gives the information on charge carrier lifetime and recombination at open circuit condition. The carrier life time was 1.7 and 0.9 µs for PCE12/PDI and PCE10/PDI systems at VOC condition, respectively. The longer carrier life time of PCE12/PDI indicating reduced recombination loss. As fitted in **Figure 3C** and **S5**, the charge carrier lifetime changes with the charge carrier density following the relationship of τ = τ0n −λ , where τ is the carrier lifetime, n is the carrier density, λ is the recombination order. The recombination order λ = 2 implies pure bimolecular recombination in the deceives, and other λ value implies both bimolecular recombination and trap-assisted recombination. We found that λ = 1.61 (±0.04) and λ = 2.16 (±0.10) were obtained from PCE12/PDI and PCE10/PDI systems. Based on the recombination order, the recombination rate then can be calculated as krec = 1 (1+λ)nτ (shown in **Figure 3D**) (Guo et al., 2013; Li et al., 2016). We can find that the krec of PCE12/PDI is lower than that of PCE10/PDI when charge carrier density is larger than 3.3 × 10<sup>15</sup> cm−<sup>3</sup> . While close to the 1 sun condition, the charge carrier density is far larger than 3.3 × 10<sup>15</sup> cm−<sup>3</sup> , and then krec of PCE12/PDI is far lower than that of PCE10/PDI. Overall, light intensity dependent JSC and TPC indicate the bimolecular recombination was comparable and the charge extraction in PCE10/PDI was better; the light intensity dependent Voc and TPV study suggest that the dominating recombination route in PCE10/PDI device was trap-assisted recombination which leads to a severe recombination rate. One may argue that the recombination rate of PCE12/PDI system was lower than that of the PCE10/PDI thus, it appears to contradict with the better device performance of PCE10/PDI. Here we remind our readers that the electron mobility of PCE12/PDI device is about 5 times slower than that of PCE10/PDI device, so we conclude that the high electron mobility should be a very important factor to give an efficient charge extraction and then the device performance in PCE10/PDI devices.

#### Morphology Characterization

Finally, we studied the morphological properties of the PCE12/PDI and PCE10/PDI blend films using tapping-mode atomic force microscopy (AFM). The PCE12/PDI film shows a rougher surface (Rq = 1.25 nm) than that of PCE10/PDI (Rq = 1.25 nm, see **Figures 4A,B**). What's more, larger extend of the phase separation with granular aggregate sizes was observed for PCE12/PDI blend films, as shown in **Figures 4C,D** and **Figure S7**. Due to the limited exciton dissociation length (10– 20 nm), smaller extent of phase-separation is beneficial for realizing efficient exciton dissociations in the device, suggesting that PCE10/PDI film has a more favourable morphology than that of the PCE12/PDI film. The morphology results well explained the low PL quenching efficiency of PCE12/PDI owing to the strong geminate recombination. The low carrier

mobility of PCE12/PDI can also be attributed to the large phase-separation, breaking the continuous pathway for electron transport. Therefore, we can conclude that the PCE12 is tending to form large size aggregations with Bis-PDI-T-EG, which is unfavorable to device performance. While PCE10 presents better compatibility with Bis-PDI-T-EG, favorable phase separation can be expected in the blend. As reported, PCE12 tends to aggregate in films and PCE10 has the tendency of forming amorphous films (Zhao et al., 2016b; Baran et al., 2017, 2018). The differences in the chemical structure of PCE12 and PCE10 should be responsible for the significant difference in thin film morphology, and it is important to check the donor materials' aggregation nature when those were chosen for PSCs with Bis-PDI-T-EG as the acceptor.

## CONCLUSION

In this work, we systematically compared two polymer/PDI blends on their optical properties, photovoltaic performance, charge carrier transport and recombination, and the thin film morphology. We found that the PCE10/PDI and PCE12/PDI can give a PCE of 5.3 and 3%, respectively, both with FF about ∼50– 60%. The hole mobilities of both devices are comparable, 3.4 × 10−<sup>4</sup> and 6.4 × 10−<sup>4</sup> cm<sup>2</sup> /V s were obtained for PCE12/PDI and PCE10/PDI, respectively. However, the electron mobilities behave 5 times difference, 2.3 × 10−<sup>6</sup> and 1.2 × 10−<sup>5</sup> cm<sup>2</sup> /V s were obtained for PCE12/PDI and PCE10/PDI, respectively. The obvious unbalanced charge carrier mobility resulted in low FF of these PSCs. By combining PL quenching efficiency, light intensity dependent J-V measurements, transient photocurrent and transient photovoltage, we noticed that both systems showed a similar extend of bimolecular recombination and PCE10/PDI behaved a severe trap-assisted recombination. Although the

## REFERENCES


recombination rate of PCE10/PDI system was stronger than that of the PCE12/PDI, the high electron mobility and the wide absorption spectrum of PCE10/PDI film result in better device performance in PCE10/PDI devices. The mobility was determined by the distinct aggregation and phase-distribution in the blend. Our findings suggest that it is important to check the aggregation nature of donor materials for Bis-PDI-T-EG based PSCs, and a proper choice is that donor material doesn't tend to aggregate, leading to favorable phase separation.

## AUTHOR CONTRIBUTIONS

TY, ZK, CZ, and SL proposed the idea of this paper and contributed to analize the experiment results and wring the paper. ZK, TY, and XZ contributed to the fabrication of the solar cells and characterization. SJ, CK, and CT contributed to the synthesis of the Bis-PDI-T-EG acceptor.

## FUNDING

This work was supported by the National Science Foundation of China (Grant No. 51502058, 61504041), CAS Pioneer Hundred Talents Program (Y82A060Q10), the China Postdoctoral Science Foundation (Grant No. 2015M570284), the Postdoctoral Foundation of Heilongjiang Province (LBH-TZ0604), and the Special Fund of Technological Innovation Talents in Harbin City (Grant No. 2017RAQXJ085).

## SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem. 2018.00613/full#supplementary-material


with Perylene Diimide to yield novel non-fullerene acceptors for organic solar cells. Molecules 23:931. doi: 10.3390/molecules23040931


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Ye, Jin, Kang, Tian, Zhang, Zhan, Lu and Kan. This is an openaccess article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

## Pyrene-Imidazole Based Aggregation Modifier Leads to Enhancement in Efficiency and Environmental Stability for Ternary Organic Solar Cells

#### Hui Lin† , Xiaoyang Du† , Lijuan Li, Caijun Zheng and Silu Tao\*

*School of Optoelectronic Science and Engineering, University of Electronic Science and Technology of China, Chengdu, China*

#### Edited by:

*Chuanlang Zhan, Institute of Chemistry (CAS), China*

#### Reviewed by:

*Zhan'Ao Tan, North China Electric Power University, China In Hwan Jung, Kookmin University, South Korea*

> \*Correspondence: *Silu Tao silutao@uestc.edu.cn*

*†These authors have contributed equally to this work*

#### Specialty section:

*This article was submitted to Organic Chemistry, a section of the journal Frontiers in Chemistry*

Received: *14 September 2018* Accepted: *06 November 2018* Published: *28 November 2018*

#### Citation:

*Lin H, Du X, Li L, Zheng C and Tao S (2018) Pyrene-Imidazole Based Aggregation Modifier Leads to Enhancement in Efficiency and Environmental Stability for Ternary Organic Solar Cells. Front. Chem. 6:578. doi: 10.3389/fchem.2018.00578* A novel pyrene-imidazole derivative (PyPI), which can form effcient π-π stacking in solid film, has been utilized in organic solar cells (OSCs). The stacking of small a molecule PyPI can facilitate a charge transfer and suppress fullerene aggregation. As a result, PTB7-Th: PyPI: PC71BM based ternary OSC exhibits a high power conversion efficiency (PCE) of 10.36%, which presents a 15.88% increase from the binary device (8.94%). Concurrently, the ternary OSC shows a much better thermal and light illumination stability. Under continuous 60◦C annealing for 3 h, in atmosphere, the device still remains at 94.13% efficiency more than the pristine state, while the control device remains at 52.47% PCE. Constant illumination under Air Mass (AM) 1.5G irradiation (100 mW cm−<sup>2</sup> ) in atmosphere, the PCE of OSC remains at 72.50%. The high conversion efficiency and excellent environmental stability of the PyPI based ternary OSC, has narrowed the gap between laboratory investigation and industrial production.

Keywords: pyrene-imidazole, π-π stacking, organic solar cells, fullerene aggregation, environmental stability

### INTRODUCTION

Solvent processed organic solar cells (OSCs) have attracted extensive attention for their superiority in achieving high power conversion efficiency (PCE), low fabrication cost and fascinating potential application in flexible electronics (Kaltenbrunner et al., 2012; Chen et al., 2014; Cui et al., 2017; Bergqvist et al., 2018; Cheng et al., 2018; Zhang H. et al., 2018). Generally, existing OSCs can be classified as binary, tandem, and ternary structures. In a traditional bulk heterojunction binary system, although the matched donor and acceptor can form a bicontinuous network interpenetrating structure to accelerate exciton dissociation and charge collection, the narrowed absorption bonds limited the further optimization on the device efficiency (Huang et al., 2017; Xu and Gao, 2018) (Li et al., 2018; Liu et al., 2018). To compensate for the shortcomings of the absoption spectrum, tandem OSCs composed of two or more subcells were fabricated to capture more photons and yielded more photon-generated carriers than conventional single-junction OSCs (Chen S. et al., 2017; Che et al., 2018; Zhang Y. et al., 2018). However, the complexity of the device structure and intricate interface,caused an increase in the device fabrication cost (Ameri et al., 2009; Kumari et al., 2017).

Compared with tandem OSCs, ternary strategy, which adds a third component to a binary system to broaden the absorption spectrum and promotes an interaction between the donor and acceptor, is an emerging and promising candidate for high performance OSCs with a simple device structure (Lu et al., 2015; Liu et al., 2016; Nian et al., 2016; Chen Y. et al., 2017; Xu et al., 2017). For ternary OSCs, the screening of the third component (either polymers or small molecules) is crucial. Challenges still remain in polymer materials as their purification and reproducibility are poor, further more, the chemical structure of polymers difficult hard to confirm. Other than polymer materials, small molecule materials have a simpler synthetic route and it is easy to obtain high purity. Additionally, the small molecules always have a mono-dispersed structure, with controlled energy levels and negligible batch to-batch variations (Chen et al., 2013; Roncali et al., 2014). Therefore, ternary OSCs using a small molecule as the third component have attracted increasing attention and have a great potential for achieving high-performance OSCs (Park et al., 2016; Chen Y. et al., 2017; Kumari et al., 2017; Zhang et al., 2017). As for the current ternary OSCs, the short-circuit current density's (JSCs) are still limited by the narrowed absorption strength, which is because of the thickness of active layers are confined to about 100 nm (Yang et al., 2015; Zhang J. et al., 2015; Zhang Y. et al., 2015; Zhang et al., 2017; Gasparini et al., 2016). That is to say, finding a way to enhance the JSCs of current OSCs is essential.

In this work, a novel small molecule PyPI (9,10-diphenyl-9H-pyreno[4, 5-d]imidazole) has been utilized to construct ternary organic solar cells. The small molecule PyPI can form efficient π-π stacking in a solid film, which is beneficial to accelerate a charge transfer in an active layer. Furthermore, the addition of PyPI can suppress fullerene aggregation and enhance device stability. For device fabrication, polymer PTB7-Th (poly(4,8-bis(5-(2-ethylhexyl)thiophen-2-yl)benzo[1,2-b;4,5-b0] dithiophene-2,6-diylalt-(4-(2-ethylhexyl)-3-fluorothieno[3,4 b]thiophene-)-2-carboxylate-2-6-diyl), and fullerene PC71BM ([6,6]-phenyl-C71-butyric acid methyl ester) was respectively used as a donor and acceptor. This polymer-fullerene system has great compatibility and is widely used in the field of organic solar cells. After precise modulation, 10% PyPI doped ternary OSC showed a high PCE of 10.36% with an optimized JSC of 19.26 mA/cm<sup>2</sup> , which exhibited a 15.88% enhancement from the control device. In addition, after continuous thermal annealling at 60 or 80◦C in atmosphere for 180 min, the PCE of the device can also be kept above 89.01%, while the control device remained at 36.96% PCE. The remarkable thermal stability is among the best of fullerene based OSCs. After continuous light illumination under 100 mW/cm<sup>2</sup> , the PyPI-containing ternary device also reveals optimized stability with a small efficiency roll-off of 27.23%, while the control device exhibits a huge PCE roll-off of 58.79%. The improvement in environmental stability demonstrates that the addition of the small molecule PyPI indeed supprsses the aggregation of fullerene.

## EXPERIMENTAL SECTION

### General Information

The materials and solvents utilized in the device fabrication and measurements were received from commercial suppliers without further purification. PTB7-Th (wt. 145,000) and PC71BM was purchased from 1-Material and American Dye Source. PEDOT:PSS was purchased from Xi'an p-OLED Technology Corp. MoO<sup>3</sup> and LiF received from Luminescence Technology Corp. All solvents used in the device fabrication process originated from Sigma-Aldrich or Alfa Chemical.

## Device Fabrication and Measurement

Conventional inverted device structures were used for the binary and ternary OSCs. ITO covered glasses with a sheet resistance of 15 Ohm per square was utilized as the substrates for these devices. The substrates were ultrasonically cleaned, in the order of deionized water, ethyl alcohol, acetone and ethyl alcohol. Before transport layers were deposited, the substrates were dried by a nitrogen blow. For the transport layer, ZnO precursor solution was formed by dissolving 110 mg of zinc acetate (Zn(CH3COO)2•2H2O) and 31 mg of ethanolamine (NH2CH2CH2OH) in 1 ml of 2-methoxyethanol (CH3OCH2CH2OH) and stirring at room temperature over night. As for the active layers, Donor and acceptor were blended with the ratio of 1:1.5, where the donor contained PTB7-Th and the small molecule PyPI, with varying proportions and maintained at a total concentration of 10 mg/ml. Chlorobenzene (CB) was used as the solvent for each of the devices. 3.0 vol% 1,8-diiodooctane (DIO) was added in the mixture as an additive. The active layer precursor solution was stirred in a nitrogen filled glove box for 24 h. For OSC fabrication, ZnO nano-particles were formed by spin-coating the precursor solution with 5,000 rpm for 30 s, after which the substrates were transferred to a heating stage and annealed immediately at 200◦C for 1 h in atmosphere. The substrates were the transferred to the glove box and the active layer precursor solutions were spin-coated onto the ZnO buffer layer to yield an uniform film (∼120 nm). After that, the substrates were transferred to a vacuum deposition chamber and when the pressure of the chamber reached 5 × 10−<sup>4</sup> Pa, 10 nm MoO<sup>3</sup> and 150 nm, Ag were evaporated at a rate of 0.5 and 3 Å/s, subsequently. An active area of 2.3 mm<sup>2</sup> was formed by a shadow mask.

## Experimental Measurements

UV absorption spectra and photoluminescence spectra of monomeric and blend films were recorded by a Hitachi U-3010 UV-VS spectrophotometer and a Perkin-Elmer LS50B Luminescence spectrophotometer, respectively. The HOMO/LUMO energy level was determined by cyclic voltammetry with a CHI600E electrochemical analyzer. Nitrogen saturated DCM was used as a solvent with 0.1 mol/L tetrabutylammonium hexa?uorophosphate as the supporting electrolyte.

The performance of the solar cells were measured by AM 1.5G simulated sunlight (Newport Oriel Sol3A Simulator, 100 mW/cm<sup>2</sup> ) with a Keithley 2,400 source meter instrument. EQE properties were determined by a QEX10 Quantum Efficiency Measurement System (PV Measurements, Inc.). The thicknesses were calibrated by a AMBIOS-XP2 step profilometer. The surface morphologies of the binary and ternary blend films were determained by an atomic force microscope (AFM) under ambient conditions. All the films were formed on ZnO coved substrates. The molecule aggregation and formed domain size were observed by a transmission electron microscopy (TEM) scanning-probe SPM system (Hitachi TEM system) under 100 kV in "Ceshigo Research Service, www.ceshigo.com."

## RESULTS AND DISCUSSIONS

#### Characterization and Optical Properties

The chemical structures of used materials in the device as well as device structure are shown in **Figure 1**. It is well-known that the fullerene aggregates in the interspace between the stacked clearances of polymers, while the small molecule PyPI can form efficient π-π stacking in the clearance in a polymer, which can suppress this inferior phenomenon as indicated in the diagram of **Figure 1A** (Liu et al., 2017).We tested through cyclic voltammetry, that the LOMO and HOMO levels of the small molecule PyPI was −2.37 and −5.42. **Figure 2A** reveals the absorption spectra of PTB7-Th, PC71BM and the small molecule PyPI, the maximum absorption peak for PTB7-Th centered at 704 nm and that of PyPI film was located at 355 and 391 nm. As in the blend films, the increase of small molecule contents along with the absorption intensity, gradually increased in the range of short wavelength, while the intensity declined in turn around long wavelengths of 600–800 nm as shown in **Figure 2B**. This is because the contents of PTB7-Th have been decreased along with the increase of PyPI. Furthermore, PyPI contained films shows a small red-shift in long wavelength absorption, suggesting an enhanced stacking order in polymers, which may be caused by the positive interaction between the small molecule and PTB7-Th.

### Photovoltaic Performance

To evaluate the contribution of the π-π stacking effect on the device performance, the small molecule PyPI was utilized as the third component to fabricated ternary OSCs. Device structure was performed as follows: ITO/ ZnO (20 nm) /active layer (120 nm) /MoO<sup>3</sup> (10 nm)/Ag (150 nm) and is shown in **Figure 1A**, a ZnO transport layer was formed with 20 nm thickness and the active layer was prepared in a glove box with

components of PyPI.

120 nm thickness. The total donor concentration was kept at 10 mg/ml in CB solvent, where the doping ratio of PyPI was tuned from 0 to 20% in donors. The weight ratio of donor: acceptor was maintained at 1:1.5. DIO was used as a solution additive with a volume ratio of 3%. To estimate the average OSC performance, about 20 samples were fabricated for each parameter.

The current density-voltage (J-V) characteristics of the binary and ternary OSCs are shown in **Figure 3A**, the corresponding

FIGURE 3 | (A) J-V curves of the binary and ternary OSCs under AM 1.5G irradiation at 100 mW cm−<sup>2</sup> ; (B) The EQE curves the binary and ternary OSCs.


*<sup>a</sup>Jcalc is calculated from EQE spectra.*

*<sup>b</sup>Statistical data obtained from 20 devices.*

device parameters are listed in **Table 1**. After rigorously optimizing the fabrication conditions, the binary OSC obtained a maximum PCE of 9.11%, along with a short-circuit current density (JSC) of 17.46 mA/cm<sup>2</sup> , an open-circuit voltage (VOC) of 0.78 V and a fill factor (FF) of 65.38%. Upon adding 5% to 20% PyPI into the binary systems, the device performances were observably improved. The highest PCE of 10.36% was achieved for PTB7-Th: 10% PyPI: PC71BM, which exhibited a maximum enhancement of 15.88% more than the control device. Concurrently, the JSC value was enhanced from 17.46 to 19.26 mA/cm<sup>2</sup> , suggesting that the charge transfer was greatly improved when PyPI was doped in the binary system. The enhancement in current density is attributed to the stacking effect of PyPI that accelerates the charge transfer in the active layer. The FF improvement of ternary OSCs can be attributed to the optimizing film morphology as discussed below. When the concentration of the small molecule increased to 20%, the PCE decreased along with the degradation on JSC and FF, which was mainly dominated by the overlarge domain size of the ternary OSC. EQE properties of the binary and ternary OSCs were measured to calibrate the high JSCs, and more detailed data are listed in **Table 1**. **Figure 3B** shows the EQE curves with a different PyPI component, and all devices exhibit a prominent photo-generated current response in the whole visible absorption region from 300 to 800 nm. Along with the doping ratio of PyPI increasing from 0 to 10%, EQEs present a remarkable enhancement both in short wavelength (300–400 nm) and long wavelength (500–800 nm). The improvement in a 300–400 nm absorption bond is attributed to the contribution of PyPI, while the enhancement in the long wave range of polymer (500– 800 nm) may be attributed to the optimization of polymer crystallinity. Further added the doping ratio of PyPI to 20%, the device EQE showed huge degeneration, and device performance also decreased, as the excess added PyPI reduced the content of the polymer and decreased the donor/acceptor connections. As a result, the ternary OSC shows a reduced JSC and FF.

#### Recombination Dynamics

The charge generation and recombination dynamics' behavior in these solar cells were studied. JSC and VOC vs. light intensity (PL) plotted on logarithm coordinate with the linear fittings are shown in **Figures 4A,B**, respectively. In JSC-P<sup>L</sup> measurements,

the variation in JSC as a function of P<sup>L</sup> can be concluded as JSC ∝ P α 0 , where α = 1 is indicative of the inexistence of bimolecular recombination in the film under short-circuit conditions. The

FIGURE 6 | TEM images for (a) PTB7-Th:PC71BM; (b) PTB7-Th: 5%PyPI: PC71BM; (c) PTB7-Th:10%PyPI: PC71BM; (d) PTB7-Th:20%PyPI: PC71BM based film.

α values of the solar cells with 0, 5, 10, 15, and 20 % PyPI were 0.93, 0.95, 0.98, 0.97, and 0.95, respectively, indicating very weak bimolecular recombination in these devices. Additionally, as shown in **Figure 4B**, the Voc of optimal device based on PTB7-Th:10 % PyPI: PC71BM shows a logarithmic dependence on P<sup>L</sup> with a slope of 1.16 kT/q compared to that of 1.62 kT/q in the binary system. The slope value for the cells containing 5, 15, and 20 % PyPI were 1.43, 1.31, and 1.51 kT/q, respectively. The results suggested that the trap-assisted recombination was effectively alleviated by adding 10% PyPI. The characteristics of the photocurrent density (Jph) vs. the effective applied voltage (Veff ) were then measured to further understand the charge generation and dissociation process in the ternary OSCs with different contents of PyPI. In theory, Jph is defined as Jph = J<sup>L</sup> − JD, where J<sup>L</sup> and J<sup>D</sup> represent photocurrent density and dark current density, respectively. Veff is defined as Veff = V<sup>0</sup> − Vbias, where V<sup>0</sup> is the voltage at which J<sup>L</sup> is equal to J<sup>D</sup> and Vbias is the applied bias voltage. As shown in **Figure 4C**, the exciton dissociation probabilities (Pdiss), which are determined by JSC/Jsat, were calculated as 92.2, 95.4, and 93.1% for the PTB7- Th:PC71BM binary device and ternary devices with 10 % and 20% PyPI under the short-circuit condition, respectively. The higher Pdiss demonstrates that the corresponding ternary device has more efficient exciton dissociation and charge extraction. These results indicated that the ternary blend system can restrain charge recombination and facilitate exciton dissociation when compared to the binary system, which corresponds to the device performance.

#### Film Morphology and Molecule Distribution

To explore the reasons why the small molecule PyPI can improve the performance of conventional machines, the surface topography and potential of the active layers of the ternary devices were examined by an atomic force microscope (AFM). As shown in **Figure 5**, after addition of 5 and 10% PyPI, the ternary blend films remained homogeneous and smooth with similar root-mean-square (RMS) values of 1.53 and 1.29 nm, respectively, which is much better than that of the control film (1.78). Obviously, the RMS value decreases as the PyPI content increases, which means that a suitable amount of PyPI can be well-incorporated into the PTB7-Th:PC71BM based control device and the surface morphology of the membrane can be optimized. When the PyPI content exceeded 10 wt% and increased to 20 wt%, the RMS increased to 1.81 due to the lower solubility of the small molecules at such a high doping ratio, resulting in a larger domain size. The formation of large areas reduced the contact interface between the donor and acceptor, which is the main reason for the lower JSC and PCE of these ternary OSCs.

Since the AFM image only reflects the surface information of the film layer, the phase distribution is studied by TEM, and the images are shown in **Figure 6**. It is well-known that bright and dark areas correspond to rich PTB7-Th and PC71BM regions, depending on the electron density. After the addition of 5% PyPI (**Figure 6b**) to prepare a ternary blend film, a better phase separation was observed compared to the binary film. The ternary device with a 10% PyPI content showed a fairly uniform membrane morphology with a rich donor/acceptor interface (**Figure 6c**), resulting in higher JSC and FF. The nanofiber network of PTB7-Th can be observed in the PTB7- Th:PC71BM blend membrane. In addition, a large number of randomly distributed large regions are obtained in the active layer. For ternary blend membranes, the nanoscale network becomes more pronounced as the PyPI content increases. In the ternary blend membrane, PyPI can modulate the PTB7-Th molecular alignment and optimize phase separation and enhance photon collection of the active layer.

## Environmental Stability

It is well-known that the aggregation of fullerene is the dominant reason of OSC performance degeneration in atmosphere. As small molecule PyPI can form a positive π-π stacking to suppress fullerene aggregation and we speculate that the intruding PyPI in the fullerene based OSC can improve the devices environmental stability. To verify this deduction, device stability of the non-encapsulated OSCs was investigated through thermal and light irradiation annealing in atmosphere. Thermal annealing temperatures were set at 60 to 80◦C, which is a practical operation temperature range for OSCs under AM 1.5G irradiation (100 mW/cm<sup>2</sup> ). **Figure 7A** shows the thermal stability curves of control OSCs and 10% PyPI contained ternary OSCs that continuously annealed at 60 and 80◦C in atmosphere for 180 min. After the annealing process, the PCE of ternary OSC remains at 94.13% of it pristine state, which is much higher than the control device (52.47%). Even more surprising, by further annealing the devices at 80◦C for 180 min, PyPI based ternary OSCs also revealed a mild decay, as shown in **Figure 7B**, PyPI contained OSC still remained at 89.01% PCE, whereas the control binary device only processed 36.96%. The superior thermal stability of the ternary OSC is mainly ascribed to the addition of the small molecule which suppressed fullerene aggregation. **Figure 7C** shows the light illumination stability curves that were measured by continuously illuminating the PyPI based ternary device under 100 mW cm−<sup>2</sup> AM 1.5G irradiation in atmosphere. The PyPI-containing ternary device revealed optimized stability with a small efficiency roll-off of 27.23%, while the control device exhibited a large PCE rolloff of 58.79%. The improvement in light illumination stability is mainly attributed to the introduction of the small molecule PyPI, restraining the aggregation of fullerene. The control device exhibited more severe efficiency degradation than PyPI based ternary OSC, which is attributed to the fact that fullerene can form dimers when exposed in light illumination (Fortunato et al., 2013; Wang et al., 2014; Heumueller et al., 2016). The dimeric fullerenes could significantly suppress charge transfer and deteriorate film morphology, resulting in declined device performance.

## CONCLUSION

A novel ternary organic solar cell system containing PTB7- Th: PyPI as the donor and PC71BM as the acceptor has been fabricated to enhance device efficiency and environmental stability. The small molecule PyPI can form efficient π-π stacking in a solid film, which is beneficial to accelerate charge transfer and suppress fullerene aggregation in the active layer. After rigorous modulation, 10% PyPI contained ternary OSC exhibited a high PCE of 10.36%, which presented a 15.88% enhancement from the control device. Furthermore, the ternary OSC showed excellent thermal and light illumination stability. Under thermal annealing for 3 h in atmosphere, the device remain at 94.13% efficiency, over pristine state, while the control device only remained at 52.47% PCE. In the condition of constant illumination under AM 1.5G irradiation (100 mW cm−<sup>2</sup> ) in atmosphere, the PCE of OSC can remain at 72.50% PCE. The excellent performance of PyPI based OSC will stimulate the development of solar cells in practical production.

## AUTHOR CONTRIBUTIONS

All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.

## ACKNOWLEDGMENTS

The work was supported by the National Natural Science Foundation of China (NSFC Grant Nos. 61775029, 61604035 and 51533005), the Fundamental Research Funds for the Central Universities (ZYGX2016Z010), International Cooperation and Exchange Project of Science and Technology Department of Sichuan Province (Grant 18GJHZ).

## REFERENCES


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Lin, Du, Li, Zheng and Tao. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

## Development of n-Type Porphyrin Acceptors for Panchromatic Light-Harvesting Fullerene-Free Organic Solar Cells

Un-Hak Lee1†, Wisnu Tantyo Hadmojo2†, Junho Kim<sup>2</sup> , Seung Hun Eom<sup>1</sup> , Sung Cheol Yoon<sup>1</sup> \*, Sung-Yeon Jang<sup>2</sup> \* and In Hwan Jung<sup>2</sup> \*

*<sup>1</sup> Division of Advanced Materials, Korea Research Institute of Chemical Technology, Daejeon, South Korea, <sup>2</sup> Department of Chemistry, Kookmin University, Seoul, South Korea*

#### Edited by:

*Chuanlang Zhan, Institute of Chemistry (CAS), China*

#### Reviewed by:

*Francesca Di Maria, Istituto di Nanotecnologia (NANOTEC), Italy Yongsheng Liu, Nankai University, China*

#### \*Correspondence:

*Sung Cheol Yoon yoonsch@krict.re.kr Sung-Yeon Jang syjang@kookmin.ac.kr In Hwan Jung ihjung@kookmin.ac.kr*

*†These authors have contributed equally to this work*

#### Specialty section:

*This article was submitted to Organic Chemistry, a section of the journal Frontiers in Chemistry*

Received: *16 June 2018* Accepted: *18 September 2018* Published: *09 October 2018*

#### Citation:

*Lee U-H, Hadmojo WT, Kim J, Eom SH, Yoon SC, Jang S-Y and Jung IH (2018) Development of n-Type Porphyrin Acceptors for Panchromatic Light-Harvesting Fullerene-Free Organic Solar Cells. Front. Chem. 6:473. doi: 10.3389/fchem.2018.00473* The development of n-type porphyrin acceptors is challenging in organic solar cells. In this work, we synthesized a novel n-type porphyrin acceptor, PZn-TNI, via the introduction of the electron withdrawing naphthalene imide (NI) moiety at the meso position of zinc porphyrin (PZn). PZn-TNI has excellent thermal stability and unique bimodal absorption with a strong Soret band (300–600 nm) and weak Q-band (600–800 nm). The weak long-wavelength absorption of PZn-TNI was completely covered by combining the low bandgap polymer donor, PTB7-Th, which realized the well-balanced panchromatic photon-to-current conversion in the range of 300–800 nm. Notably, the one-step reaction of the NI moiety from a commercially available source leads to the cheap and simple n-type porphyrin synthesis. The substitution of four NIs in PZn ring induced sufficient n-type characteristics with proper HOMO and LUMO energy levels for efficient charge transport with PTB7-Th. Fullerene-free organic solar cells based-on PTB7-Th:PZn-TNI were investigated and showed a promising PCE of 5.07% without any additive treatment. To the best of our knowledge, this is the highest PCE in the porphyrin-based acceptors without utilization of the perylene diimide accepting unit.

Keywords: porphyrin acceptors, n-type porphyrins, organic solar cells, non-fullerene acceptors, panchromatic absorption

## INTRODUCTION

For decades, organic solar cells (OSCs) have been studied as a portable and low-cost power generator due to their unique advantages: light-weight, use of earth-abundant organic materials, solution-processability and flexibility. The power conversion efficiency (PCE) of OSCs, an important parameter to determine the performance of the OSCs, has been continuously improved by optimization of light-harvesting in active layers and hole/electron charge transport pathways (Zhan et al., 2015; Zhao J. et al., 2016; Lin et al., 2017; Che et al., 2018; Hou et al., 2018). In the 2000s, low-bandgap polymer donors and n-type fullerene acceptors blended OSCs were developed to make broad absorption in the active layer. Polymer donors showing a strong absorption coefficient were responsible for light-harvesting and exciton generation, while the n-type fullerenes effectively separate and transfer the electrons to the electrode (Kim et al., 2006; Liang et al., 2010; Li et al., 2012; Liao et al., 2013). However, this was not enough to cover all the visible band by blending polymer donors and fullerene acceptors. In the 2010s, new types of OSCs replaced the fullerenes with strong light-harvesting organic nonfullerene acceptors (NFAs). The active layer was composed of organic donors and NFAs, called fullerene-free OSCs (Cheng et al., 2017; Lin et al., 2017; Tang et al., 2018; Yan et al., 2018). Due to the strong absorption of both NFAs and organic donors in the visible area, their complementary absorption is important to achieve panchromatic photon-to-current conversion in the active layer. There are several successful strategies for panchromatic absorption in the visible and near infra-red area. The most common approach is mixing wide-bandgap donors and lowbandgap small-molecule NFAs such as ITIC (Zhao W. et al., 2016, 2017; Lin et al., 2017; Yao et al., 2017; Zhao F. et al., 2017). Another approach is mixing low-bandgap donors and wide-bandgap NFAs (e.g., blending of PTB7-Th and perylenediimide (PDI)-based acceptor) (Wu et al., 2016; Duan et al., 2017; Liang et al., 2017; Eom et al., 2018). The red-dye PDIs enable a strong absorption in the short-wavelength area of 400–600 nm, while showing effective n-type characteristics.

As a new approach for panchromatic absorption in an active layer, utilization of a nature-inspired porphyrin dye has recently emerged in OSCs.(Gao et al., 2015; Li et al., 2016; Hadmojo et al., 2018) The porphyrin dyes have peculiar bimodal absorption characteristics composed of Soret and Q bands; strong transition from ground state (S0) to second excited state (S2) yields the Soret band, while the weak transition from S0 to first excited state (S1) provides the Q-band. Thus, the strong Soret absorption of porphyrin dyes enables efficient short-wavelength absorption in the 400–600 nm, which can be blended with low-bandgap donors having a dominant absorption in the 600–800 nm for panchromatic absorption. In addition, the long-wavelength absorption of the Q-band intensifies the light harvesting in the low bandgap area where abundant solar flux exists. However, most of the developed artificial porphyrin derivatives are p-type materials and only a few porphyrin derivatives currently show ntype characteristics with a promising PCEs over 5% (Hadmojo et al., 2017; Zhang et al., 2017). Exploring new structures for n-type porphyrin materials is challenging in fullerene-free OSCs.

In this study, we synthesized a novel porphyrin acceptor, PZn-TNI, via Sonogashira coupling of 5,10,15,20-tetrakis-ethynyl porphyrin Zinc (II) (PZn) and 4-bromo-N-(2-ethylhexyl)- 1,8-naphthalimide (NI). NI is easily synthesized from the commercially available 4-bromo-1,8-naphthalic anhydride. This one-step reaction is beneficial in terms of time and cost for synthesizing the star-shape molecules that require an excess of NIs. Since the NI has n-type characteristics, the substitution of NIs to the four meso positions of PZn enables the excellent ntype properties as an electron acceptor. The ethyne π-bridge unit is incorporated between NI and PZn to increase the backbone planarity. The synthesized PZn-TNI showed excellent thermal stability with 5% weight loss temperature of 412◦C and showed unique bimodal absorption behavior with maximum peaks at 479 nm and 719 nm. The uncovered UV-Vis absorption spectrum from PZn-TNI is completely covered by the blending of a polymer donor, PTB7-Th, which resulted in the panchromatic photon-to-current conversion from 300 to 800 nm in OSCs. The planar backbone structure of PZn-TNI assists the sizable face-on orientation in the PTB7-Th:PZn-TNI blend film without additive treatment, which resulted in the highest PCE of 5.07% (VOC = 0.72 V, JSC = 13.84 mA cm−<sup>2</sup> , and fill factor = 0.51) in the additive-free OSCs. The excessive ordering of PTB7-Th:PZn-TNI film via pyridine additive rather reduced the photovoltaic performances. Our successful utilization of NI moiety in the PZn core will broaden and diversify the synthetic approaches for developing high-efficiency porphyrin acceptors.

### EXPERIMENTAL

#### Synthesis

5,10,15,20-tetrakis-ethynyl porphyrin Zinc (II) (3): Macrocyclic porphyrin compound 1 was synthesized according to the reported general procedure (Yen et al., 2006). Compound 1 (0.90 g, 1.29 mmol) and zinc acetate (2.4 g, 13.1 mmol) were dissolved in the co-solvent (200 ml) of dichloromethane:methanol = 9:1 v/v%. The resulting mixture was refluxed at 65◦C for 24 h. After removing the solvents, the remaining solid was rinsed with dichloromethane (300 mL) to give a purple solid compound 2 (0.65 g, yield: 62%). Without further purification, compound 2 was directly used to make compound 3. Compound 2 (0.60 g, 0.79 mmol) was dissolved in 100 mL anhydrous tetrahydrofuran (THF). Tetran-butylammonium fluoride solution 1.0 M in THF (3.6 ml, 3.6 mmol) was slowly added to the reaction mixture. The resulting mixture was stirred at room temperature for 4 h. After removing the solvents, the crude solid product was rinsed sequentially with methanol, dichloromethane, water and acetone. After drying the dark purple solid product 3 (180 mg, yield: 49%), it was immediately used in the next step to prevent the coupling reaction between the two terminal alkynes. <sup>1</sup>H NMR (THF-d6, 400 MHz, ppm): δ 9.60 (s, 2H), 5.37 (s, 1H),

N,N'-(2-hexyldecyl) 4-bromo naphthalene imide (4): 2-Hexyldecyl amine (2.4 g, 9.93 mmol) was added to the suspension of 4-bromo-1,8-naphthalic anhydride (2.5 g, 9.03 mmol) in dry ethanol (50 mL). The reaction mixture was refluxed overnight at 110◦C, and then cooled down to room temperature. After evaporating the solvents, the remaining crude solid was purified using column chromatography on silica gel with an eluent of CH2Cl2:n-hexane = 4:1 to give a yellow solid compound 4 (3.0 g, 66%). <sup>1</sup>H NMR (CDCl3, 400 MHz, ppm): δ 8.65 (d, J = 6.4 Hz, 1H), 8.56 (d, J = 7.6 Hz, 1H), 8.40 (d, J = 8.0 Hz, 1H), 8.03 (d, J = 7.6 Hz, 1H), 7.82 (m, 1H), 4.14 (t, J = 8.0 Hz, 2H), 1.70 (m, 1H), 1.24 (br, 27H), 0.87 (t, J = 7.2 Hz, 3H).

PZn-TNI: Compound 3 (100 mg, 0.212 mmol), compound 4 (1.06 g, 2.12 mmol), Pd2(dba)<sup>3</sup> (40 mg, 0.044 mmol) and AsPh<sup>3</sup> (100 mg, 0.33 mmol) were dissolved in dry THF (15 mL) and triethylamine (15 mL). The reaction mixture was stirred at 65◦C for 4 days under N<sup>2</sup> atmosphere, and then quenched by distilled water. The organic layer was extracted using dichloromethane and water, and then the moisture in the organic solution was removed by Na2SO4. After evaporating the solvents, the solid residue was purified by column chromatography using an eluent of CH2Cl2:n-hexane = 4:1. Then it was further purified using recycling size exclusion chromatography to give a deep green solid PZn-TNI (290 mg, 75%). <sup>1</sup>H NMR (THF-d6, 400 MHz,

ppm): δ 9.02 (br, 2H), 8.85 (br, 1H), 8.40 (d, J = 4.8 Hz, 1H), 8.28 (d, J = 5.6 Hz, 1H), 8.19 (br, 1H), 7.88 (br, 1H), 4.24 (m, 2H), 1.92 (m, 1H), 1.39 (br, 27H), 0.96 (m, 3H). <sup>13</sup>C NMR (THFd6, 100 MHz, ppm): δ 163.58, 163.32, 149.75, 131.61, 131.49, 130.46, 130.18, 128.58, 128.35, 127.90, 124.02, 123.05, 102.08, 101.88, 96.30, 41.51, 33.14, 31.09, 31.05, 31.02, 30.99, 30.92, 30.87, 30.61, 29.22, 28.66, 23.82, 14.71. MALDI-TOF-MS: m/z: calcd. for C140H160N8O8Zn: 2145.17 [M] <sup>+</sup>; found 2145.994.

### RESULTS AND DISCUSSION

The synthetic procedure of PZn-TNI was recorded in **Scheme 1** and in the Supporting Information (SI) in detail. The porphyrin ring 1 was synthesized from pyrrole and 3-(trimethylsilyl)propiolaldehyde in the presence of BF3·Et2O followed by oxidation with 2,3-dichloro-5,6-dicyano-1,4 benzoquinone (DDQ). The zinc porphyrin (PZn) compound 2 was obtained using Zn(OAc)2. The deprotection of TMS group by tetra-n-butylammonium fluoride (TBAF) was performed immediately before synthesizing the final acceptor, PZn-TNI. 4-Bromo-1,8-naphthalic anhydride purchased from Sigma-Aldrich was alkylated with 2-hexyldexylamine to give a compound 4. This one-step reaction to prepare the electron withdrawing NI unit is highly beneficial in terms of time and cost for achieving the n-type porphyrins. The final porphyrin acceptor, PZn-TNI, was achieved via Sonogashira coupling with PZn compound 3 and excess of NI compound 4, which was identified by <sup>1</sup>H-NMR and matrix assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS) (**Figures S1**–**S4**). The synthesized PZn-TNI showed excellent solubility in common organic solvents such as tetrahydrofuran (THF), dichloromethane (DCM), and chloroform (CF). In addition, it has excellent thermal stability, which was determined by thermal gravimetric analysis (TGA), with a 5% weight loss temperature (T5d) of 412◦C under an N<sup>2</sup> atmosphere (**Figure S5**).

Absorption spectra of PZn-TNI were measured in solution and the film state, as shown in **Figure 1**. PZn-TNI exhibited clear bimodal absorption composed of the Soret band (300 – 600 nm) and Q-bands (600–800 nm); the maximum absorption peaks of PZn-TNI were 479 and 713 nm in solution, and 478 and 719 nm in the film. The absorption of PZn-TNI was complementary to that of the low-bandgap donor polymer, PTB7-Th, which induced a well-balanced short- and longwavelength absorption in the entire wavelength of 300–800 nm (**Figure 2B**). Notably, the film of PZn-TNI showed broadened

and red-shifted Q-band absorption spectra compared to that in solution, indicating the enhanced intermolecular π-π stacking in the film state. Since the planar NIs and PZn are connected by an sp-hybridized ethyne π-bridge, PZn-TNI possesses a highly planar conjugated backbone for efficient intermolecular stacking. The optical bandgap (E ◦pt <sup>g</sup> ) of PZn-TNI was 1.63 eV, which was calculated from the absorption onset wavelength of 761 nm in the film. We previously reported the pyridine additive effect on the molecular ordering of porphyrin derivatives; pyridine enhances the intermolecular ordering of porphyrin derivatives via coordination to the zinc (Hadmojo et al., 2017). As shown in **Figure 1B**, the absorption of PZn-TNI was significantly broadened and red-shifted compared to that without pyridine treatment, resulting in the bathochromic shift of 42 nm in the Soret band and 15 nm in the Q-band. This supports our previous hypothesis for the pyridine effect on the molecular ordering and confirms that pyridine enhances the intermolecular ordering of PZn-TNI in the film states.

To evaluate the energy levels of PZn-TNI, the highest occupied molecular orbital (HOMO) energy levels and the lowest unoccupied molecular orbital (LUMO) energy levels were measured using cyclic voltammetry. The oxidation and reduction onset potentials of PZn-TNI were 1.07 and −0.80 V, respectively, which corresponds to HOMO levels (EHOMO,CV) and LUMO levels (ELUMO,CV) of −5.50 and −3.62 eV, respectively. The optical LUMO energy level (ELUMO,UV) was calculated to be −3.87 eV from EHOMO,CV and E ◦pt <sup>g</sup> of PZn-TNI. The energy diagram of the PTB7-Th and PZn-TNI was shown in **Figure 1D**, and the EHOMO,CV, ELUMO,CV, and ELUMO,UV of polymer donor (PTB7-Th) were taken from our previous measurement (Zhang et al., 2015; Hadmojo et al., 2016). The LUMO energy level of PZn-TNI is suitable for electron transport from PTB7-Th to PZn-TNI, while the HOMO of PZn-TNI is appropriate for hole transport from PZn-TNI to PTB7-Th (Marcus, 1963; Clarke and Durrant, 2010). The optical and electrochemical properties of PZn-TNI are summarized in **Table 1**.

Porphyrin acceptor-based fullerene-free OSCs were fabricated by blending PTB7-Th and PZn-TNI (**Figure 2A**). The current density-voltage (J-V) characteristic of PTB7-Th:PZn-TNI devices was investigated via changing the weight ratio between PTB7-Th and PZn-TNI, and the photovoltaic performance was optimized at the weight ratio of 1:1.5 w/w. The photovoltaic properties are summarized in **Figure 2C**, **Figure S6**, and **Table 2**. The best PCE of 5.07% was achieved with a VOC of 0.72 V, a JSC of 13.84 mA cm−<sup>2</sup> , and FF of 0.51 (**Figure 2C**). As shown in **Figure 2D**, the external quantum efficiency (EQE) spectra of PTB7-Th:PZn-TNI devices cover the entire visible area of 300–800 nm and


*<sup>a</sup>Chloroform solution.*

*<sup>b</sup>Film on a quartz plate.*

*<sup>c</sup>Bandgap calculated from the film-state absorption onset wavelength (*λ*onset).*

*<sup>d</sup>HOMO levels determined from Eox of the first oxidation potential of PZn-TNI.*

*<sup>e</sup>LUMO levels calculated from EHOMO, CV and E*◦*pt g .*

w/w. (D) the EQE spectra of PTB7-Th:PZn-TNI devices.

showed the panchromatic photon-to-current conversion due to the complementary solar flux absorption between PTB7-Th and PZn-TNI. Notably, the additive-free film-formation process provided superior photovoltaic performance compared to the pyridine-assisted one as shown in **Figure 2C**. As shown in **Figure 2D**, the EQE was increased in the entire wavelength, which indicates that hole/electron transport properties of both PTB7-Th and PZn-TNI are improved in the additive-free devices.

To understand the charge recombination mechanisms of PTB7-Th:PZn-TNI devices in the presence and absence of pyridine additive, the J-V characteristics were investigated as a function of the illumination intensity. The power law dependence of JSC on the illumination intensity is generally expressed as JSC ∞ I α , where I is the light intensity and α is an exponential factor (**Figure 3A**) (Blom et al., 2007; Azmi et al., 2016). The α value of the PTB7-Th:PZn-TNI devices was close to unity regardless of additive treatment, indicating the negligible bimolecular recombination in PTB7-Th:PZn-TNI devices. However, the VOC vs. illumination intensity was highly affected by the pyridine treatment. Assuming there is no trap-assisted recombination under an open-circuit condition, the slope of VOC vs. the illumination intensity produces 1.00 kT/q (Mihailetchi et al., 2006). The higher value of kT/q indicates the more probability of trap-assisted recombination under an open-circuit condition (Mandoc et al., 2007; Azmi et al., 2016). As shown in **Figure 3B**,


TABLE 2 | Photovoltaic properties of PTB7-Th:PZn-TNI devices.

*<sup>a</sup>Average PCEs more than 10 devices.*

PTB7-Th:PZn-TNI devices with and without additives show the slope of 2.63 and 1.93 kT/q, respectively. This implies that PTB7-Th:PZn-TNI devices in the absence of additives have the lowest trap-assisted recombination in anopen-circuit condition. The hole and electron transport properties of PTB7- Th:PZn-TNI devices were measured by a space-charge-limitedcurrent (SCLC) analysis (**Figures 3C,D**) (Mihailetchi et al., 2005). The electron- and hole-only devices were fabricated with a structure of ITO/ZnO/PTB7-Th:PZn-TNI/ZnO/Al and ITO/PEDOT:PSS/PTB7-Th:PZn-TNI/MoOx/Ag, respectively. In the presence of pyridine additive, the hole and electron mobilities of PTB7-Th:PZn-TNI were 2.4 × 10−<sup>4</sup> and 1.3 × 10−<sup>6</sup> cm<sup>2</sup> V −1 ·s −1 , respectively, whereas, in the absence of additives, the hole and electron mobilities were increased to 2.9 × 10−<sup>4</sup> and 2.5 × 10−<sup>6</sup> cm<sup>2</sup> V −1 ·s −1 , respectively. Thus, it is expected that the pyridine additive worsens the nanomorphology of PTB7-Th:PZn-TNI devices via excessive intermolecular aggregation.

The morphology of the PTB7-Th:PZn-TNI active layer was investigated by atomic force microscopy (AFM) (**Figure 4**) and two-dimensional grazing incidence X-ray diffraction (2D-GIXD) analyses (**Figure 5**). In AFM images, PTB7-Th:PZn-TNI blended film possesses bicontinuous crystalline domains in the absence of additives (**Figures 4A,B**), whereas the addition of pyridine additive intensifies the intermolecular ordering of PZn-TNI domains, leading to severe phase segregation between PTB7-Th and PZn-TNI (**Figures 4C,D**). The 2D-GIXD results also support the AFM analysis. Additive-free PTB7-Th:PZn-TNI film showed clear π-π stacking orientation (010) peak at ∼1.6 Å−<sup>1</sup> along the q<sup>z</sup> axis, which indicates the face-on orientation with a dspacing of ∼3.9 Å (**Figures 5B,C**). However, the pyridine-treated PTB7-Th:PZn-TNI film showed the increased π-π stacking

interaction and induced the phase aggregation (**Figure 5C**). The (010) peak in the blend film is assigned to the orientation of PTB7-Th domains (**Figure 5A**), which implies that the decrease in the photovoltaic performances in presence of pyridine is probably due to the aggregation of the PTB7-Th domains in the PTB7-Th:PZn-TNI blend film. As a result, the PZn-TNI having highly planar molecular structure possesses the sizable π-π intermolecular stacking and crystalline nanomorphology in the additive-free solvent system, which means that no more post-treatment is required in PTB7-Th:PZn-TNI blend system. In addition, the additive-free system can prevent undesirable morphological change and photo-oxidation degradation by additives in the active layer (Li et al., 2017).

### CONCLUSIONS

We have synthesized a novel porphyrin acceptor, PZn-TNI, by incorporating four naphthalene imide (NI) units at the meso position of the PZn core. PZn-TNI showed unique bimodal absorption with a strong Soret band and a weak Q-band. The insufficient long-wavelength absorption of PZn-TNI was covered by a low-bandgap donor, PTB7-Th. As a result, bulk heterojunction fullerene-free OSCs composed of PZn-TNI and PTB7-Th showed panchromatic photon-to-current conversion covering entire area of 300–800 nm. The PTB7-Th:PZn-TNI devices exhibited a promising PCE of 5.07%, which is the highest and the first promising PCE in the porphyrin-based acceptors except for those utilizing the PDI units. Notably, the additivefree solution process provided the best photovoltaic performance, whereas the pyridine additive had a negative effect on the nanomorphology by the excessive molecular aggregation of the PTB7-Th:PZn-TNI film. The planar backbone structure of PZn-TNI assists the sizable molecular ordering in the PTB7-Th:PZn-TNI film without additive treatment, which is favorable for practical applications.

## AUTHOR CONTRIBUTIONS

IJ and SY conceived the ideas and designed the PZn-TNI. U-HL synthesized all the materials and JK assisted the characterization of all the materials. S-YJ supervised all the device fabrication and optimization. WH fabricated all the OSC devices and SE assisted the device characterization.

#### ACKNOWLEDGMENTS

The authors gratefully acknowledge support from the New and Renewable Energy Core Technology Program of the Korea Institute of Energy Technology Evaluation and Planning (KETEP), granted financial resources from the Ministry of Trade, Industry and Energy, Republic of Korea (No. 20163030013960), the National Research Foundation (NRF) Grant funded by the Korean Government (MSIP, No. 2016R1A5A1012966 and No. 2017R1C1B2010694), and the Global Scholarship Program for Foreign Graduate Students at Kookmin University in Korea.

#### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem. 2018.00473/full#supplementary-material

## REFERENCES


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Lee, Hadmojo, Kim, Eom, Yoon, Jang and Jung. This is an openaccess article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

## Recent Progress in Fused-Ring Based Nonfullerene Acceptors for Polymer Solar Cells

#### Chaohua Cui\*

*Laboratory of Advanced Optoelectronic Materials, College of Chemistry, Chemical Engineering and Materials Science, Soochow University, Suzhou, China*

The progress of bulk-heterojunction (BHJ) polymer solar cells (PSCs) is closely related to the innovation of photoactive materials (donor and acceptor materials), interface engineering, and device optimization. Especially, the development of the photoactive materials dominates the research filed in the past decades. Photoactive materials are basically classified as p-type organic semiconductor donor (D) and an n-type organic semiconductor acceptor (A). In the past two decades, fullerene derivatives are the dominant acceptors for high efficiency PSCs. Nevertheless, the limited absorption and challenging structural tunability of fullerenes hinder further improve the efficiency of PSCs. Encouragingly, the recent progresses of fused-ring based A-D-A type nonfullerene acceptors exhibit great potential in enhancing the photovoltaic performance of devices, driving the power conversion efficiency to over 13%. Such kind of nonfullerene acceptors is usually based on indacenodithiophene (IDT) or its extending backbone core and end-caped with strong electron-withdrawing group. Owing to the strong push-pulling effects, the acceptors possess strong absorption in the visible-NIR region and low-lying HOMO (highest occupied molecular orbital) level, which can realize both high open-circuit voltage and short-circuit current density of the devices. Moreover, the photo-electronic and aggregative properties of the acceptors can be flexibly manipulated via structural design. Many strategies have been successfully employed to tune the energy levels, absorption features, and aggregation properties of the fused-ring based acceptors. In this review, we will summarize the recent progress in developing highly efficient fused-ring based nonfullerene acceptors. We will mainly focus our discussion on the correlating factors of molecular structures to their absorption, molecular energy levels, and photovoltaic performance. It is envisioned that an analysis of the relationship between molecular structures and photovoltaic properties would contribute to a better understanding of this kind of acceptors for high-efficiency PSCs.

Keywords: polymer solar cells, nonfullerene acceptor, molecular design, power conversion efficiency, energy levels

#### Edited by:

*Chuanlang Zhan, Institute of Chemistry (CAS), China*

#### Reviewed by:

*Renqiang Yang, Qingdao Institute of Bioenergy and Bioprocess Technology (CAS), China Weiwei Li, Institute of Chemistry (CAS), China*

> \*Correspondence: *Chaohua Cui cuichaohua@suda.edu.cn*

#### Specialty section:

*This article was submitted to Organic Chemistry, a section of the journal Frontiers in Chemistry*

Received: *12 June 2018* Accepted: *20 August 2018* Published: *25 September 2018*

#### Citation:

*Cui C (2018) Recent Progress in Fused-Ring Based Nonfullerene Acceptors for Polymer Solar Cells. Front. Chem. 6:404. doi: 10.3389/fchem.2018.00404*

## INTRODUCTION

Typically, bulk-heterojunction (BHJ) polymer solar cells (PSCs) are composed of a photoactive layer sandwiched between a transparent anode and a low work function metal cathode (Li, 2012; Li et al., 2012; Nielsen et al., 2012; Chen et al., 2013; Heeger, 2013; Janssen and Nelson, 2013; Zhan et al., 2015; Elumalai and Uddin, 2016; Zhan and Yao, 2016; Zhang et al., 2017). The PCE of PSC is proportional to open-circuit voltage (Voc), short-circuit current density (Jsc), and fill factor (FF). The progress of PSCs is closely related to the innovation of photoactive materials (donor and acceptor materials) (He and Li, 2011; Li, 2013; Cui et al., 2014, 2016; Ye et al., 2014; Lu et al., 2015; Cui and Wong, 2016; Cui Y. et al., 2017; Hu et al., 2017; Lopez et al., 2017; Osaka and Takimiya, 2017; Zou et al., 2017; Gupta et al., 2018; Liu et al., 2018; Sun et al., 2018), interface engineering (He et al., 2012; Duan et al., 2013; Chueh et al., 2015; Wang et al., 2015; Chen et al., 2016; Street, 2016), and device optimization (Ameri et al., 2009, 2013; Meillaud et al., 2015; Cui C. et al., 2017; Li W. et al., 2017; Zhao et al., 2018). Especially, the development of PSCs is always accompanied by photoactive material innovations. As the key component, photoactive materials are basically classified as a p-type organic semiconductor donor (D) and an n-type organic semiconductor acceptor (A). Due to the unique advantages of strong electron-accepting and high electrontransport capabilities, fullerene derivatives were predominately used as the acceptor in PSCs in the past two decades, driving the power conversion efficiency (PCE) of PSCs to 11–12% (Liu et al., 2014; Zhao J. et al., 2016). Nevertheless, fullerenes based acceptors show critical shortcomings of weak absorption and limited structural modification, hindering further improve photovoltaic performance of devices. To overcome these obstacles of fullerenes based acceptors, many efforts have been devoted to developing new kind of nonfullerene acceptor materials (Hendriks et al., 2014; Cheng et al., 2018; Hou et al., 2018; Yan et al., 2018). Very recently, A-D-A conjugated fusedring molecules based on indacenodithiophene (IDT, **Figure 1**) or DTIDT unit (**Figure 4**) were reported as excellent nonfullerene acceptors for high performance PSCs, leading the PCE of device to over 13% (Wang et al., 2016; Li S. et al., 2018). Very recently, the PCEs of nonfullerene based PSCs have been driven to a milestone of over 14% (Li S. et al., 2018; Zhang et al., 2018). Unlike fullerene derivatives, fused-ring based nonfullerene acceptor materials offer many molecular design strategies to tune their optoelectronic properties and thus photovoltaic performance. In this review, we will provide some representative cases of molecular manipulation on IDT and DTIDT based nonfullerene acceptors to fine-tune the physicochemical and photovoltaic properties. We hope that this review article would contribute to a better understanding of the design strategies of high performance fused-ring based acceptors for efficient nonfullerene PSCs.

#### IDT BASED FUSED-RING ACCEPTORS

IDT unit (**Figure 1**) which features phenylene ring fused to thiophene was firstly reported by Wong in 2006 (Wong et al., 2006) Such fused rings structure is beneficial to forming effective interchain π-π overlap and enhance the rigidity of the molecular backbone as well as the degree of conjugation. Zhan et al. innovatively used IDT as central core to develop an A-D-A (A = acceptor, D = donor) type acceptor material (NA1) with 2-(3-oxo-2,3-dihydroinden-1-ylidene)-malononitrile as terminal acceptor unit (Lin et al., 2016a). NA1 showed promising energy levels and absorption spectrum as acceptor material for PSCs. By using NA1 as acceptor and PDBT-T1 as donor to fabricate PSC device, a high PCE of 7.39% was obtained, with Voc = 0.92 V, Jsc = 13.39 mA cm−<sup>2</sup> , and FF = 0.60 (**Table 2**) (Lin et al., 2016a). Molecular optimization based on NA1 greatly affects the photovoltaic performance. In the following, we will discuss the molecular design strategies including extension of conjugated backbone, substituted side chains, and end-capped group were conducted on NA1.

### Extension of Conjugated Backbone With Donor Unit

The Voc of PSCs device is tightly correlated with the energy level difference between the HOMO of the donor and the LUMO of the acceptor. Therefore, high LUMO level of acceptor material is essential for achieving high Voc value. In D-A conjugated molecular system, the donor unit mainly determines the HOMO level. In other words, the optical bandgap (Eg) of D-A based molecules can be tuned by incorporating donor unit as conjugated extension while maintaining the similar LUMO level. For example, Zhan et al. employed one and two IDT units as conjugated extension block to develop two molecules NA2 and NA3 (Lin et al., 2016a). Due to the longer conjugated backbone, the absorption profiles of NA2 and NA3 are effectively red-shifted compared to NA1, and NA3 possesses the lowest E<sup>g</sup> of 1.53 eV (**Table 1**). On the other hand, NA2 and NA3 showed up-shifted HOMO levels while similar LUMO levels. Due to the weaker molecular π-π stacking compared to NA1, the NA2 based device exhibited a low PCE of 2.58%, while no photovoltaic response was observed from the NA3 based device. In comparison with NA1, NA4 with thiophene units as π-bridge for conjugated extension showed slightly red-shifted absorption spectrum (E<sup>g</sup> = 1.55 eV), up-shifted HOMO level of −5.42 eV, and similar LUMO level of −3.85 eV (Bai et al., 2015a). The device based on PBDTTT-C-T:NA4 exhibited a PCE of 3.93%, with a Voc of 0.90 V, Jsc of 8.33 mA cm−<sup>2</sup> , and FF of 0.523 (**Table 2**). Presumably owing to the conjugated twists, the E<sup>g</sup> was increased to 1.57 eV when attaching 2-ethylhexyl chains in thiophene π-bridge of NA1 (IEIC, **Figure 1**) (Lin et al., 2015b). By using IEIC as acceptor and PTB7-Th as donor to fabricate device, a promising PCE of 6.31% was obtained, with Voc = 0.97 V, Jsc = 13.55 mA cm−<sup>2</sup> , and FF = 0.48 (**Table 2**). To reduce the E<sup>g</sup> of IEIC, Hou et al. replaced the 2-ethylhexyl chains of IEIC with alkoxy chains to develop a new acceptor IEICO (Yao et al., 2016). Attributing to the strong electrondonating ability of alkoxy chains, IEICO exhibited a lower E<sup>g</sup> of 1.34 eV than IEIC. Relative to IEIC, the density functional theory calculation result suggests that the introduction of alkoxy chains effectively up-shifted the HOMO level (∼0.19 eV) while

maintaining the similar LUMO level (0.01 eV lower than IEIC). The PSCs using IEICO as acceptor yielded a high PCE of 8.4%, with a Voc of 0.82 V and a Jsc of 17.7 mA cm−<sup>2</sup> , while the control device with IEIC as acceptor exhibited a much lower PCE of 4.9% (**Table 2**) (Yao et al., 2016). In comparison with IEIC-based PSCs, the higher Jsc value of IEICO-based device should be resulted from the much broader photo-response spectrum with higher external quantum efficiency. Bo et al. used bis(alkoxy)-substituted or dialkyl-substituted benzene ring as π bridge for conjugated extension to develop two molecules NA5 and NA6 (**Figure 1**) (Liu Y. et al., 2017). Benefiting from the noncovalent S···O interaction locks, NA6 exhibited better planarity and broader absorption spectrum than NA5, with a lower E<sup>g</sup> of 1.63 eV (**Table 1**). In addition, NA6 with locked conformation exhibited a higher quantum yield, which can effectively suppress the non-radiative energy loss and afford higher Voc for devices. TABLE 1 | Summary of absorption properties and energy levels of IDT core based nonfullerene acceptors shown in Figure 1.


TABLE 2 | Summary of photovoltaic properties of the nonfullerene acceptors shown in Figure 1.


TABLE 3 | Summary of absorption properties and energy levels of IDT core based nonfullerene acceptors shown in Figure 2.


The device based on NA6 realized a promising PCE of 9.60%, with Voc = 1.01 V, Jsc = 17.52 mA cm−<sup>2</sup> , and FF = 0.54, while the NA5 based device showed a much lower PCE of 2.3% (**Table 2**).

As demonstrated above, using electron rich unit as conjugated backbone extension for NA1 is an effective strategy to broaden absorption spectrum, rise up HOMO levels while maintain similar LUMO levels of resulting molecules.

## Extension of Conjugated Backbone With Acceptor Unit

Since the LUMO distribution is mainly located at the acceptor unit in D-A conjugated molecular system, using acceptor unit as building block can simultaneously manipulate the LUMO level and E<sup>g</sup> of the molecules. Zhan et al. develop a nonfullerene acceptor NA7 (**Figure 1**) using benzothiadiazole as π-bridge (Wu et al., 2015). NA7 shows flat backbone configuration which is beneficial for charge transport, and a large dihedral angle between the hexylphenyl group and backbone plane which can prevent the over self-aggregation when blending with P3HT. Due to the relatively high LUMO level of NA7, the device based on P3HT:NA7 exhibited a high Voc of 0.84 V, with a high PCE of 5.12% (**Table 2**) (Wu et al., 2015). Bazan et al. developed two NA7 analogs, NA8 and NA9 (**Figure 1**), with different positions of the fluorine atom in benzothiadiazole unit (Zhong et al., 2017). Relative to NA7 (E<sup>g</sup> = 1.68 eV), NA8 exhibited a similar E<sup>g</sup> of 1.67 eV, while NA9 showed a slightly larger E<sup>g</sup> of 1.71 eV. The orientations of the fluorine atoms show little influence in the HOMO and LUMO levels but affect the calculated conformational diversity and the electrostatic potential of the molecules. The device based on PTzBI:NA8 exhibited a PCE of 7.44%, higher than that of PTzBI:NA9 based device (PCE = 5.28%). The photovoltaic performance of NA9 based device is poorer than that of NA8 based device, which should be resulted from the less optimal BHJ morphology. Zhou et al. replaced the benzothiadiazole units of NA7 by benzotriazole units to develop NA10 (**Figure 1**) (Tang et al., 2018). Due to the weaker electron-accepting ability of benzotriazole than benzothiadiazole, NA10 showed a higher LUMO level than NA7. Therefore, the J61:NA10 based device achieved an encouraging Voc of 1.24 V, with a PCE of 3.02% (**Table 2**).

In short, the extension of conjugated backbone with donor or acceptor units will generally broaden absorption spectra and reduce E<sup>g</sup> of the resulting molecules, and the introduction of donor units as π-bridge is more effective to reduce the E<sup>g</sup> than acceptor units. On the other hand, the LUMO levels will be up-shifted when using donor units as π-bridge, while the incorporation of acceptor units will lead to higher LUMO levels.

#### Side Chains Engineering

The conjugated side chains substituents on the IDT unit will increase steric hindrance, reduce intermolecular interactions, and prevent over self-aggregation and large phase separation in blend film. Herein, the physicochemical and photovoltaic properties of IDT based acceptors can easily tune via side chains engineering in IDT unit. Zhan et al. developed an acceptor with non-conjugated alkyl chains in IDT unit (NA11, **Figure 2**) (Jia et al., 2017). NA11 exhibited a nearly flat molecular backbone configuration, with a lower Eg, higher HOMO level, and higher electron mobility than NA7 (**Table 3**). The device based on PTB7- Th:NA11 yielded a higher PCE of 8.7% than that of PTB7- Th:NA7 based device (**Table 4**). Furthermore, the NA11 based device exhibited better thermal stability and photo stability in comparison with NA7 based device. McCulloch et al. reported two alkyl chains substituted IDT based nonfullerene acceptors

TABLE 4 | Summary of photovoltaic properties of the nonfullerene acceptors shown in Figure 2.


NA12 and NA 13 (**Figure 2**) (Holliday et al., 2016). NA12 with linear alkyl chains showed a stronger crystallinity and a narrower E<sup>g</sup> relative to NA13 with branched chains, resulting in higher Jsc and PCE values (**Tables 3**, **4**). In addition, the oxidative stability of these devices is superior to the benchmark P3HT:PC60BM device.

#### Effects of End-Capped Groups

The end-capped groups also affect the optical and electrochemical properties of this kind of nonfullerene acceptors. Hou et al. extended the π-conjugation area of the end group of NA1 to develop a new acceptor IDTN (as shown in **Figure 3**) (Li S. et al., 2017). The enlarged π-conjugation by phenyl unit in the end group effectively leads to red-shifted absorption and slightly lower LUMO and HOMO levels compared with NA1 (**Table 5**). Due to the enhanced intermolecular interactions and molecular ordering, IDTN shows a better molecular planarity and higher electron mobility than NA1. Therefore, an outstanding PCE of 12.2% was achieved from the PBDB-TF:IDTN based device, which is significantly higher than that of the NA1 based device (PCE = 7.4%). Zhan et al. used 2-(benzo[c][1,2,5] thiadiazol-4-ylmethylene)-malononitrile as end-capped groups to develop a new acceptor NA14 (**Figure 3**) (Bai et al., 2015b). Relative to NA7, the stronger electron-withdrawing ability of end-capped units of NA14 leads to a lower E<sup>g</sup> of 1.60 eV, lower LUMO of −3.8 eV, and lower HOMO of −5.6 eV (**Table 5**). The PBDTTT-C-T:NA14 based device afforded a relatively high PCE of 4.26%. Zhou et al. systematically engineered the end-capped units of three nonfullerene acceptors to carefully tune the driving force for high Voc and Jsc values (Tang et al., 2018). With the increase of the electron-withdrawing ability of the end-capped units from NA15, NA10, to NA16, the LUMO levels and E<sup>g</sup> simultaneously decrease (**Table 5**). By fine-tune the LUMO level of acceptor via end-capped unit, NA16 exhibited sufficient energy offset with J61 for efficient charge generation. The device based on J61:NA16 obtained a high PCE of 8.25%, with a high Voc of 1.15 V (**Table 6**). Zhu et al. developed two thieno[3,4-b]thiophene-based acceptor, NA17 and NA18, with different end-capped groups (**Figure 3**) (Liu et al., 2016; Liu F. et al., 2017) NA17 exhibited an E<sup>g</sup> of 1.54 eV, HOMO level of −5.50 eV, and LUMO level of −3.63 eV. Relative to NA 17, NA18 with stronger electron-withdrawing terminal group possesses lower E<sup>g</sup> of 1.32 eV and deeper LUMO of −3.90 (**Table 5**). Device based on PTB7-Th:NA17 yielded a high PCE of 10.07%, with Voc = 0.87 V, Jsc = 16.48 mA cm−<sup>2</sup> , and FF = 0.70 (**Table 6**). Attributing to the deeper LUMO and broader absorption spectrum of NA18, PTB7-Th:NA18 based device exhibited a lower Voc of 0.73 V and a higher Jsc of 16.48 mA cm−<sup>2</sup> , leading to a promising PCE of 9.58% (**Table 6**). Yang et al. changed the end-capped unit of IEIC and IEICO to develop two analogs IDTC and IDTO (**Figure 3**) (Luo et al., 2017). Unexpectedly, IDTO showed slightly blue-shifted absorption range relative to IDTC. Nevertheless, the introduction of alkoxy groups effectively improved the intermolecular interactions and up-shifted the LUMO level of IDTO. The device based on PBDB-T:IDTC exhibited a PCE of 9.35%, with a Voc of 0.917 V, while the PBDB-T:IDTO based device showed a higher PCE of 10.02% and a higher Voc of 0.943 V (**Table 6**). Hou et al. introduced fluorine atoms onto the end group of IEICO (IEICO-4F, **Figure 3**) to

nonfullerene acceptors shown in Figure 3.


shown in Figure 3.


enhance the intramolecular charge transfer effect. IEICO-4F showed lower Eg of 1.24 eV and higher LUMO level of −4.19 eV than IEICO. Using IEICO-4F as acceptor, PBDTTT-EFT or J52 as donor, high Jsc values over 20 mA cm−<sup>2</sup> were both recorded in the corresponding devices (**Table 6**).

### IDTT BASED FUSED-RING ACCEPTORS

IDTT unit with two additional extended thiophene rings than IDT exhibits excellent planarity (Wong et al., 2006). The pioneering work for IDTT-based nonfullerene acceptor is the development of ITIC which was firstly reported by Lin et al. (2015a) and Zhan et al. (2015). ITIC possesses strong absorption, suitable energy level, good electron transport ability, and good miscibility with various polymer donors. A promising PCE of 6.80% was achieved from the PTB7-Th:ITIC based device (**Table 8**) (Lin et al., 2015a). Later on, many state-of-the-art ITICbased PSCs with excellent photovoltaic performance have been reported (Bin et al., 2016; Gao et al., 2016; Qin et al., 2016; Zhao W. et al., 2016; Yang et al., 2017a,b; Xu et al., 2018), and ITIC has been regarded as excellent acceptor for high performance

TABLE 7 | Summary of absorption properties and energy levels of IDT core based nonfullerene acceptors shown in Figure 3.


PSCs. To further improve the photovoltaic performance, plenty of molecular design strategies have been carried out to optimize the physicochemical of ITIC. Zhan et al. extended the fusedring core of ITIC and end-capped with different acceptor unit to develop a series of ITIC derivatives INIC, INIC1, INIC2, and INIC3 (**Figure 4**) (Dai et al., 2017). Relative to ITIC, INIC with extended conjugated core showed red-shifted absorption spectrum (E<sup>g</sup> = 1.57 eV), slightly higher HOMO level, and slightly lower LUMO level (**Table 7**). The introduction of fluorine atom onto the end-capped group of INIC effectively red-shifted the absorption and down-shifted the HOMO and LUMO levels (**Table 7**). The device based on INIC:FTAZ showed a high Voc of 0.957 V, with moderate PCE of 7.7%. The fluorination in INIC leads to lower Voc values, while significantly improves the FF and Jsc of the corresponding devices, delivering a higher PCE values (**Table 7**). The end-capped groups of ITIC show great impact in its photovoltaic performance. Variations of endcapped groups have been conducted onto ITIC to optimize its photovoltaic properties. To up-shift the LUMO level of ITIC without causing too much steric hindrance for intermolecular packing, Hou et al. modulated the LUMO levels of ITIC by incorporating one and two methyls in the end-capped groups (IT-M and IT-DM, **Figure 3**). Benefited from the weak electrondonating property of methyl, the LUMO levels of IT-M and IT-DM were elevated by 0.04 and 0.09 eV relative to ITIC, respectively (**Table 7**) (Li et al., 2016). Therefore, higher Voc values of 0.94 V and 0.97 V were achieved from the PBDB-T:IT-M and PBDB-T:IT-DM based device, respectively. Encouragingly, a remarkable PCE of 12.05% was realized from the PBDB-T:IT-M based device, with Jsc = 17.44 mA cm−<sup>2</sup> , and FF = 0.735 (**Table 8**). One successful molecular optimization on ITIC is the incorporated F-atoms into the end-capping groups to develop IT-4F (**Figure 4**) (Zhao et al., 2017). Due to the electron-pulling

TABLE 8 | Summary of photovoltaic properties of the nonfullerene acceptors shown in Figure 4.


effect of the fluorine atom, IT-4F showed reduced LUMO level, red-shifted absorption spectrum, and enhanced intramolecular charge transfer effects than ITIC. By rational selection of polymer donor with matching energy level, over 13% PCEs have been achieved from IT-4F based PSCs (Cui Y. et al., 2017; Li S. et al., 2018; Zhang et al., 2018). Other end groups engineering such as the replacement of phenyl-fused indanone of ITIC by thienyl-fused indanone as end-groups (ITCC) also affects the electronic properties and enhances intermolecular interactions (Yao et al., 2017b). ITCC possesses larger E<sup>g</sup> of 1.67 eV and up-shifted HOMO and LUMO levels (**Table 7**) than ITIC. In combination with the improved electron-transport properties and high-lying LUMO level of ITCC, an impressive Voc of 1.01 V and a high PCE of 11.4% was achieved from the ITCC based PSC device (**Table 8**). Changing the orientation of end-capped thiophene of ITCC (ITCPTC, **Figure 4**) leads to reduced E<sup>g</sup> of 1.58 eV and deeper energy levels (**Table 7**) (Dongjun et al., 2017). Furthermore, such thiophene-fused ending group can promote the molecular interactions and crystallization compared to ITIC with a benzene-fused end-capped group. The PSCs device using ITCPTC as acceptor and PBT1-EH as donor demonstrated a high PCE of 11.8%, with a remarkable FF of 0.751 (**Table 8**). Further molecular optimization of ITCPTC is the introduction of methyl onto the thiophene-fused end groups (MeIC, **Figure 4**) (Luo et al., 2018). Due to the weak electron-donating ability

of methyl group, MeIC showed slightly up-shifted LUMO level than ITCPTC and maintained the intramolecular interaction and crystallization. The MeIC-based PSC achieved a high PCE of 12.54%, with a Voc of 0.918, a Jsc of 18.41 mA cm−<sup>2</sup> , and FF of 0.742% (**Table 8**). Li et al. introduced double bond πbridges into ITIC to develop three acceptor materials (NA19, NA20, and NA21, as shown in **Figure 4**) (Li X et al., 2017). The insertion of vinylene π-bridge reduces the Eg, and the fluorine substitution down-shifts the HOMO and LUMO levels of the molecules (**Table 7**). The PSC device based on J71:NA19 showed a moderate PCE of 7.34%. Significantly enhanced Jsc of 19.73 mA cm−<sup>2</sup> was obtained from J71:NA20 based device, with a high PCE of 9.72%. In comparison with the NA19 and NA20 based devices, the devices based on J71:NA21 exhibited the highest PCE of 10.54%, with a notable Jsc of 20.60 mA cm−<sup>2</sup> .

Similar to IDT unit, the steric effect of tetrahexylphenyl substituents on the IDTT unit also can reduce intermolecular interactions and prevent the acceptor from forming excessively large crystalline domains when blending with donor material. Thus, the electronic and intramolecular properties can be fine-tuned via the side chains manipulation on IDTT unit. Li et al. developed an analog (m-ITIC, **Figure 4**) by side chain isomerism engineering on the alkyl-phenyl substituents of ITIC (Yang et al., 2016). m-ITIC exhibited slightly reduced E<sup>g</sup> and up-shifted LUMO and HOMO levels, while more crystalline and stronger film absorption coefficient than ITIC. In comparison with J61:ITIC based device, overall better photovoltaic performance was realized in J71:m-ITIC based device (**Table 8**). The replacement of phenyl side chains on ITIC by thienyl side chains leads to lower energy levels and increased intermolecular interactions of resulting molecule (ITIC-Th) (Lin et al., 2016b). The enhanced intermolecular interaction of ITIC-Th relative to ITIC should be attributed to the easy polarization of sulfur atom and sulfur-sulfur interaction. A high PCE of 9.6% was obtained from the PDBT-T1:ITIC-Th based device (**Table 8**). Consider the fact that linear alkyl chains could potentially improve the packing ability and the charge transport mobility of resulting molecules over bulky side chains, Heeney et al. developed an IDTT-based acceptor with linear alkyl side chains (C8-ITIC) (Fei et al., 2018). C8-ITIC showed reduced Eg, higher absorptivity, and increased propensity to crystallize than ITIC. The device based on C8-ITIC recorded an impressive PCE of 13.2%, which is higher than that of ITIC based device (PCE = 11.71%, **Table 8**).

#### REFERENCES


#### Summary and Perspective

In summary, we have reviewed the recent progress of IDT and IDTT based nonfullerene acceptors for PSCs. Compared with fullerene acceptors, IDT and IDTT based nonfullerene acceptors offer plenty of molecular design possibilities to tune the physicochemical properties. With the purpose to well-match with the specific donor material, the absorption feature and energy levels of IDT and IDTT based acceptors can be easily and effectively manipulated by rational selection of π-bridge and end-capped groups. Moreover, the intermolecular packing, molecular orientation, as well as crystallinity can be optimized by side-chains engineering to form good morphology with donor materials. Benefiting from the diversification of chemical modification on acceptors and donors, significant progress has been achieved from the nonfullerene acceptors based PSCs. Obviously, the emerging of nonfullerene acceptors brings a bright future for PSCs field. Nevertheless, nonfullerene acceptors still confront challenges. Firstly, although it is straightforward to manipulate the optical absorption and energy level of the IDT and IDTT based molecules, the anisotropic conjugated structures of nonfullerene acceptors make it more complicate to tune the miscibility between donor and acceptor for welldeveloped morphology. The deep insight into the relationship between molecular structure and photovoltaic should be further exploited. In particularly, various nonfullerene acceptors with different photo-electronic and molecular packing properties have been developed, the rational selection of acceptor material to well-match with polymer donor is essential. Secondly, to further improve the photovoltaic performance, much effort should be devoted to manipulate the energy levels of donor and acceptors for minimizing the energy loss of the devices. Finally, the stability of nonfulleren based device should also be fully investigated.

#### AUTHOR CONTRIBUTIONS

The author confirms being the sole contributor of this work and approved it for publication.

#### ACKNOWLEDGMENTS

This work is supported by National Natural Science Foundation of China (51603136), Jiangsu Provincial Natural Science Foundation (BK20150327), and China Postdoctoral Science Foundation (2015M581855 and 2017T100395).


solar cells with a new non-fullerene acceptor. Nat. Commun. 7:11585. doi: 10.1038/ncomms11585


and non-fullerene organic solar cells. Mater. Chem. Front. 1, 1389–1395. doi: 10.1039/C7QM00025A


**Conflict of Interest Statement:** The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The reviewer WL and handling Editor declared their shared affiliation.

Copyright © 2018 Cui. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

## Dithienonaphthalene-Based Non-fullerene Acceptors With Different Bandgaps for Organic Solar Cells

Meiqi Zhang1,2, Yunlong Ma<sup>1</sup> and Qingdong Zheng<sup>1</sup> \*

*<sup>1</sup> State Key Laboratory of Structural Chemistry, Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, Fuzhou, China, <sup>2</sup> University of Chinese Academy of Sciences, Beijing, China*

Compared to the traditional fullerene derivatives, non-fullerene acceptors show more tunable absorption bands as well as adjustable energy levels which are favorable for further PCE enhancement of organic solar cells. In order to enhance light-harvesting property of dithienonaphthalene (DTN)-based acceptors, we designed and synthesized two novel non-fullerene acceptors (DTNIF and DTNSF) based on a ladder-type DTN donor core flanked with two different acceptor units. In combination with a benchmark wide bandgap copolymer (PBDB-T), the best performance device based on DTNIF displayed a high PCE of 8.73% with a short-circuit current (*J*sc) of 13.26 mA cm−<sup>2</sup> and a large fill factor (FF) of 72.77%. With a reduced bandgap of DTNSF, the corresponding best performance device showed an increased *J*sc of 14.49 mA cm−<sup>2</sup> although only a moderate PCE of 7.15% was achieved. These findings offer a molecular design strategy to control the bandgap of DTN-based non-fullerene acceptors with improved light-harvesting.

#### Edited by:

*Chuanlang Zhan, Institute of Chemistry (CAS), China*

#### Reviewed by:

*Chaohua Cui, Soochow University, China Jianhua Huang, Huaqiao University, China*

\*Correspondence: *Qingdong Zheng qingdongzheng@fjirsm.ac.cn*

#### Specialty section:

*This article was submitted to Organic Chemistry, a section of the journal Frontiers in Chemistry*

Received: *07 July 2018* Accepted: *30 August 2018* Published: *24 September 2018*

#### Citation:

*Zhang M, Ma Y and Zheng Q (2018) Dithienonaphthalene-Based Non-fullerene Acceptors With Different Bandgaps for Organic Solar Cells. Front. Chem. 6:427. doi: 10.3389/fchem.2018.00427* Keywords: organic solar cell, non-fullerene, ladder-type structure, power conversion efficiency, bandgap

## INTRODUCTION

Organic solar cells (OSCs) have attracted increasing attention over the past decade due to their light-weight, mechanical flexibility, and potential low-cost (Facchetti, 2011; Liu et al., 2014; Rong et al., 2015). Bulk heterojunction (BHJ) OSCs featuring with an active layer of an electron acceptor material blended with an electron donor material, are widely used (Wu et al., 2011; Chen et al., 2013; Wang et al., 2014; Xu et al., 2015). In the early years' research on OSCs, fullerene derivatives, such as [6,6]-phenyl-C71-butyric acid methyl ester (PC71BM) and [6,6]-phenyl-C61-butyric acid methyl ester (PC61BM) have been used as the dominant electron acceptors due to their high electron mobilities and unique phase separation property when blended with rod-like donor materials (You et al., 2013; Ma et al., 2016). Although power conversion efficiencies (PCEs) of fullerene-based OSCs have surpassed 10% in single-junction OSCs (Liu et al., 2014; Chen et al., 2015; He et al., 2015; Zhang et al., 2015), the poor absorption in the visible region and the limited tunability in energy levels of the fullerene derivatives prevent a further PCE improvement of fullerene-based OSCs (Zhan et al., 2011). To break these limitations, emerging efforts have thus been devoted to designing non-fullerene acceptors which could have broader absorption, more adjustable energy levels and structural flexibility in comparison with the fullerene derivatives (Li et al., 2015; Lin et al., 2015; Nielsen et al., 2015; Liu et al., 2016, 2018; Qin et al., 2016; Zhang et al., 2016; Tang et al., 2017; Shen et al., 2018). Among the non-fullerene acceptors, small molecules with acceptor-donor-acceptor (A-D-A) configuration are popular in organic photovoltaic field because the HOMO and LUMO energy levels of A-D-A type molecules can be separately tuned by selecting suitable donor cores and acceptor terminals (Lin et al., 2015; Wu et al., 2015; Zhao et al., 2017). Using ladder-type angular-shaped dithienonaphthalene (DTN) as the donor unit and 2-(3-oxo-2,3-dihydro-1H-inden-1-ylidene)malononitrile (INCN) as the strong electron-withdrawing unit, we reported an A-D-A type non-fullerene acceptor (DTNIC8), recently, which exhibited a bandgap of 1.73 eV and a decent PCE of 9.03% (Ma et al., 2017b). In order to obtain DTN-based nonfullerene acceptors with an up-shifted LUMO energy level, we further used 5-(benzo[c][1,2,5]thiadiazol-4-ylmethylene)-3 ethyl-2-thioxothiazolidin-4-one as the weak electron-deficient group (Ma et al., 2017a). The resulting non-fullerene acceptor (DTNR) exhibited a similar bandgap of 1.72 eV but a much high-lying LUMO energy level of −3.75 eV which is beneficial for achieving a large Voc for the corresponding PSC. Both the DTN-based acceptors showed relatively wide bandgaps with intense absorption bands in the range of 500–750 nm (Ma et al., 2017a,b). In order to improve the PCEs of OSCs based on wide bandgap donor materials, the bandgaps of non-fullerene acceptors based on DTN should be reduced further. For the A-D-A type non-fullerene acceptors, their bandgaps can be reduced by using stronger electron withdrawing groups as terminals (Zhao et al., 2017) and by extending π-conjugation length of the molecular backbone (Dai et al., 2017).

In this context, two novel DTN-based non-fullerene acceptors, DTNIF and DTNSF, were designed and synthesized by using a stronger electron withdrawing group of 2-(6-fluoro-3-oxo-2,3-dihydro-1H-inden-1-ylidene)malononitrile (INCNF), or by inserting two additional thiophene bridges in molecular backbone (shown in **Figure 1**). Inverted OSCs were fabricated by blending a typical wide bandgap copolymer (PBDB-T in **Figure 1**) with the DTN-based non-fullerene acceptors. The PBDB-T:DTNIF-based devices exhibited a PCE of 8.73% with a high FF of 72.77%, and a short circuit current (Jsc) of 13.26 mA cm−<sup>2</sup> . However, the PBDB-T:DTNSF-based devices showed a moderate PCE of 7.15% with an increased Jsc of 14.49 mA cm−<sup>2</sup> and a low FF of 54.62%. Moreover, we also studied effects of the terminal units on the bandgap, energy level, and charge transporting property of the DTN-based non-fullerene acceptors.

## RESULTS AND DISCUSSION

#### Synthesis and Characterization

The synthetic routes of DTNIF and DTNSF are shown in **Scheme 1** and the synthetic details are described in the Experimental section. Compounds **1** and **2** were synthesized according to our earlier published methods (Ma et al., 2013, 2017b).

Compound **3** was obtained in 96% yield by the Stille coupling reaction between Compound **2** and 5-bromothiophene-2-carbaldehyde using Pd(PPh3)<sup>4</sup> as the catalyst. A Knoevenagel condensation reaction between Compound **3** and INCNF afforded DTNSF in 46% yield. DTNIF was synthesized in 60% yield by using the same condensation reaction between Compound **1** and INCNF. The chemical structures of DTNIF and DTNSF were determined by using <sup>1</sup>H NMR and high-resolution mass spectrometry. The purity of the acceptors was verified further by elemental analysis. All non-fullerene materials are soluble at room temperature in the traditional organic solvents, such as CH2Cl2, chlorobenzene, and CHCl<sup>3</sup> etc.

#### Optical and Electrochemical Properties

The absorption properties of DTNIF and DTNSF were investigated in CHCl<sup>3</sup> solution as well as in thin film. The obtained linear absorption spectra are shown in **Figure 2** and the specific optical data are shown in **Table 1**.

In chloroform solution, DTNIF displayed a strong absorption band in the wavelength region of 500–700 nm with a clear shoulder peak at 600 nm which can be attributed to the intramolecular charge transfer from the electron donating core to the electron withdrawing terminals. Compared to DTNIF, DTNSF showed a red-shifted absorption band in the wavelength region of 520–780 nm which can be ascribed to its extended conjugation with two additional thiophene bridges. The maximum extinction coefficient of DTNIF (1.8 × 10<sup>5</sup> M−<sup>1</sup> cm−<sup>1</sup> at 637 nm) was higher than that of DTNSF (1.4 × 10<sup>5</sup> M−<sup>1</sup> cm−<sup>1</sup> at 682 nm). From solution to thin film, both the nonfullerene acceptors showed broader and red-shifted absorptions. The optical bandgaps estimated from their absorption edges were 1.63 eV and 1.47 eV for DTNIF and DTNSF, respectively, both of which are lower than the bandgap of DTNIC8 (1.73 eV in **Table 1**). With the standard A-D-A configuration in the molecular backbone, DTNIF exhibits a deep HOMO energy level of −5.82 eV. However, with the insertion of two additional thiophene units in the A-D-A backbone, the HOMO energy level of DTNSF increases to −5.52 eV together with a significant reduced bandgap of 1.47 eV which is mainly attributed to its extended π-conjugation in comparison with DTNIF. As shown in **Figure 2B**, DTNSF in thin film exhibits a more complementary absorption spectrum with PBDB-T in comparison that with DTNIF, suggesting a possible enhanced Jsc value for the DTNSFbased OSC.

The electrochemical properties of the non-fullerene acceptors were tested by electrochemical cyclic voltammetry (CV). Here, ferrocene was used as an internal reference, which has a HOMO level of −4.80 eV. The cyclic voltammograms are shown in **Figure 3A** and the corresponding data are listed in **Table 1**. According to their onset potentials, the HOMO/LUMO energy levels of DTNIF and DTNSF were calculated to be −5.82/−3.92 and −5.52/−4.00 eV, respectively. PBDB-T has HOMO and LUMO energy levels of −5.33 and −2.92 eV, respectively, which energetically matched with those of the acceptors (**Figure 3B**).

#### Photoluminescence

In order to know the exciton dissociation as well as the charge transfer behaviors of donor/acceptor blends, photoluminescence (PL) quenching experiments were carried out and the results were shown in **Figure 4**. We selected 665 nm and 750 nm as the excitation wavelengths to respectively excite DTNIF and DTNSF in either pure or blend films. The donor/acceptor blend ratios were fixed at 1:1 by mol. As shown in **Figures 4A,C**, the


TABLE 1 | Optical and electrochemical properties of DTNIF and DTNSF.

*<sup>a</sup>Estimated from the onset of the absorption spectra of thin films; <sup>b</sup>EHOMO* = *–(*ϕ*ox* + *4.82) eV; <sup>c</sup>ELUMO* = *–(*ϕ*red* + *4.82) eV.*

strong emission of DTNIF at 700 nm and 760 nm (DINSF at 825 nm) in the blend film apparently quenched when compared to that in the pure film, demonstrating the efficient hole transfer from both the acceptors to PBDB-T (donor). As for the PL emission of PBDB-T (shown in **Figures 4B,D**), the PL intensities of PBDB-T:DTNIF and PBDB-T:DTNSF blend films decreased significantly in comparison with those of the pure PBDB-T film when excited at 570 and 580 nm, respectively. It suggested that there is efficient electron transfer from the PBDB-T donor to both the acceptors. These results demonstrated that both the non-fullerene acceptors and the polymer donor contribute to the photocurrent generation of the OSCs.

#### Photovoltaic Performance

PBDB-T is a wide-bandgap polymer donor which has a strong absorption band in the wavelength range from 500 to 700 nm. The absorption bands of our non-fullerene acceptors generally match the absorption band of PBDB-T. Thus, we chose PBDB-T as donor to fabricate OSCs with an inverted device structure: indium tin oxide (ITO)/ZnO/donor:acceptor/MoO3/Ag. The active layers were spin-coated by using PBDB-T:acceptor (w/w, 1:1) blend solution in chlorobenzene (18 mg/mL) without any additives and post-treatments. The J–V curves of the best performance devices are shown in **Figure 5A** and detailed device parameters are summarized in **Table 2**.

Under simulated AM 1.5 G, 100 mW cm−<sup>2</sup> illumination and the optimal device fabrication condition, the best performance OSC based on PBDB-T:DTNIF showed a PCE of 8.73% with a Voc of 0.90 V, a Jsc of 13.26 mA cm−<sup>2</sup> and a FF of 72.77%. Nevertheless, the best performance DTNSF-based device exhibited a PCE of 7.15% with a Voc of 0.92 V and a lower FF of 54.62%. The lower FF was mainly resulted from the lower and less balanced hole and electron mobilities for the PBDB-T:DTNSF active layer. However, the Jsc of 14.49 mA cm−<sup>2</sup> for the DTNSF-based device is larger than the PBDB-T:DTNIFbased counterpart owing to the red-shifted absorption of DTNSF. We noticed that the DTNSF-based device showed a slightly higher Voc than the DTNIF-based device despite the fact that DTNIF possesses a higher LUMO level than DTNSF. Besides the energy gap (between HOMO of the donor and LUMO of the acceptor) which can affect the Voc of corresponding device, other factors, such as recombination rate, reverse saturation current, carrier density, defect states and crystallinity, and charge-transfer states could also play an important role in influencing the Voc (Elumalai and Uddin, 2016). Therefore, it is reasonable that the DTNIF-based device exhibited a relatively lower Voc of 0.90 V.

As shown in **Figure 5B**, external quantum efficiency (EQE) spectra of the best performance devices were measured to ensure the accuracy of the PCE measurements. The device based on DTNIF has higher EQE values in the wavelength range of 300– 750 nm with a maximum value of 78% at 660 nm. In contrast, the EQE spectrum edge of the best performance DTNSF-based device extended to 850 nm, which agrees with the absorption spectrum of the DTNSF blend film. The Jsc values obtained by integrating the EQE data with the solar spectrum (AM 1.5 G) were 13.26 and 14.21 mA cm−<sup>2</sup> for DTNIF and DTNSF,

respectively. The integrated values are in consistent with those from the J-V measurement within 2% error.

### Film Morphology Analysis

Tapping-mode atomic force microscopy (AFM) was used to characterize the morphology of active layer that has an important influence on the performance of OSCs. The film samples for AFM analysis were prepared in identical fashion to those prepared for device fabrication in which the donor/acceptor blend ratios were fixed at 1:1 by mol. The obtained AFM images were presented in **Figure 6**. The AFM height images of the DTNIF and DTNSF-based blend films showed similar and apparently fibrillar structures (**Figures 6A,B**). However, the DTNSF-based blend shows smoother root-mean-square (RMS) roughness (Rq) than the DTNIF-based blend. Compared to PBDB-T:DTNIF film with a R<sup>q</sup> of 3.25 nm, the RMS roughness of PBDB-T:DTNSF film decreased to 2.09 nm which could be attributed to the smaller intramolecular twisted angel and greater coplanarity of DTNSF. As shown in the phase images (**Figures 6C,D**), fibrillar structure can also be observed in both the blend films. In comparison with PBDB-T:DTNSF blend film, PBDB-T:DTNIF film revealed fibrillar structures with larger sizes which will be favorable for efficient charge transport in the DTNIF-based devices as confirmed by their higher hole and electron mobilities.

As mentioned above, the optimal morphology can enhance charge transport efficiency that will further affect the Jsc and FF of OSCs. We measured the electron (µe) and hole (µh) mobilities using the space charge limited current (SCLC) method with the device structures of ITO/ZnO/PBDB-T:acceptor/Ca/Al and ITO/PEDOT:PSS/PBDB-T:acceptor/Au, respectively. For both the hole- and electron-only devices, the donor/acceptor ratios for are fixed at 1:1 by mol. The J-V characteristics of the hole-only and electron-only devices are shown in **Figure 7** and the mobility data are shown in **Table 3**. The µ<sup>e</sup> and µ<sup>h</sup> for the PBDB-T:DTNIF blend film were calculated to be 1.79 × 10−<sup>5</sup> and 1.87 × 10−<sup>5</sup> cm<sup>2</sup>


V −1 s −1 , respectively, which far exceeded those for the PBDB-T:DTNSF film (µ<sup>e</sup> = 6.70 × 10−<sup>6</sup> and µ<sup>h</sup> = 1.35 × 10−<sup>5</sup> cm<sup>2</sup> V −1 s −1 ). More balanced µh/µ<sup>e</sup> ratio of 1.04 was observed for the PBDB-T:DTNIF blend film when compared to a larger µh/µ<sup>e</sup> ratio of 2.01 for the PBDB-T:DTNSF blend. Thus, the higher and more balanced hole and electron mobilities of the PBDB-T:DTNIF blend can explain the higher FF of the resulting solar cell.

#### CONCLUSIONS

In summary, we have developed two novel non-fullerene acceptors, DTNIF and DTNSF, with different bandgaps. The introduction of F atom into the terminal group leads to a slightly narrow bandgap and a red-shifted absorption in the 500–750 nm region. To achieve a more complementary spectrum of nonfullerene acceptor with wide bandgap donor materials, such as PBDB-T, we further introduced two thiophenes as bridge units which lead to a more planar molecular configuration with an extended conjugation. Without any additive and post-treatment, the best performance DTNIF-based device exhibited a PCE of 8.73% with Voc of 0.90 V, FF of 72.77% and Jsc of 13.26 mA cm−<sup>2</sup> . The best performance device based on DTNSF afforded an enhanced Jsc of 14.49 mA cm−<sup>2</sup> although only a moderate PCE of 7.15% was obtained due to the decreased and unbalanced hole and electron mobilities of the DTNSF-based active layer. It should be noted that the device performance of DTNIF-based

TABLE 3 | Hole and electron mobilities of the SCLC devices based on two different active layers.


OSCs might be improved by selecting other donor polymers of more complementary absorption spectra.

### EXPERIMENTAL SECTION

#### Materials and Characterization

All the solvents were purified and dried according to standard procedures. The donor polymer PBDB-T (99.9%) was bought from Solarmer Materials, Inc. Compounds **1** and **2** were prepared by using the reported procedure (Ma et al., 2013, 2017b).

Synthesis of **3**: Compound **2** (0.6 g, 0.56 mmol), 5 bromothiophene-2-carbaldehyde (0.32 g, 1.6 mmol) and Pd(PPh3)<sup>4</sup> (30 mg, 0.03 mmol) were dissolved in 30 mL of degassed toluene in a two neck flask. After refluxing for 24 h under nitrogen, the mixture was cooled down to room temperature. Then the solvent was removed by evaporation, and the remaining residue was purified by column chromatography (silica gel) using petroleum ether/CH2Cl<sup>2</sup> (3:1) as eluent. Finally, a dark brown solid (0.47 g, 96%) was obtained. <sup>1</sup>H NMR (CDCl3, 400 MHz, ppm): 9.91 (s, 2H), 8.04 (d, J = 8.0 Hz, 2H), 7.75 (d, J = 8.0 Hz, 2H), 7.67 (d, J = 8.0 Hz, 2H), 7.43 (s, 2H), 7.37 (s, 2H), 2.17–2.04 (m, 8H), 1.01–0.52 (m, 60H). HRMS (MALDI) m/z: calc. for C62H80O2S4: 984.5017; found: 984.5027. Elemental analysis (%) calc. for C62H80O2S4: C, 75.56; H, 8.18; found: C, 75.79; H, 8.09.

Synthesis of **DTNSF**: To a solution of Compound **3** (200 mg, 0.2 mmol) in dry CHCl<sup>3</sup> (30 mL), 2-(6-fluoro-3-oxo-2,3-dihydro-1H-inden-1-ylidene)malononitrile (340 mg, 1.6 mmol) were added. After degassing with nitrogen for 30 min, 0.15 mL of pyridine was added. The mixture was stirred at reflux for 24 h under nitrogen atmosphere. After the mixture was cooled to room temperature, it was poured into 100 mL of methanol. A precipitate was formed and filtered off which was further purified by using column chromatography (silica gel) with petroleum ether/CH2Cl<sup>2</sup> (1:1) as the eluent. A dark green solid (130 mg, 46%) was obtained. <sup>1</sup>H NMR (CDCl3, 400 MHz, ppm): 8.91 (d, J = 8.0 Hz, 2H), 8.42 (d, J = 8.0 Hz, 2H), 8.10 (d, J = 8.0 Hz, 2H), 8.01–7.97 (m, 2H), 7.86 (d, J = 8.0 Hz, 2H), 7.73–7.70 (m, 2H), 7.68 (d, J = 8.0 Hz, 2H), 7.50–7.45 (m, 4H), 2.20–2.09 (m, 8H), 1.05–0.54 (m, 60H). HRMS (MALDI) m/z: calc. for C86H86F2N4O2S4: 1,373.5689; found: 1,373.5674. Elemental analysis (%) calc. for C86H86F2N4O2S4: C, 75.18; H, 6.31; N, 4.08; found: C, 75.47; H, 6.20; N, 3.77.

Synthesis of **DTNIF**: To a solution of Compound **1** (174 mg, 0.2 mmol) in dry CHCl<sup>3</sup> (30 mL), 2-(6-fluoro-3-oxo-2,3-dihydro-1H-inden-1-ylidene)malononitrile (337 mg, 1.6 mmol) were added. After degassing with nitrogen for 30 min, 1 mL of pyridine was added into the mixture which was further stirred at reflux for 24 h under nitrogen atmosphere. Then the mixture was cooled down to room temperature. The reaction mixture was poured into 100 mL of methanol. A precipitate was formed and filtered off which was further purified by using column chromatography (silica gel) with petroleum ether/CH2Cl<sup>2</sup> (1:1) as the eluent. A dark metallic luster solid (126 mg, 60%) was obtained. <sup>1</sup>H NMR (CDCl3, 400 MHz, ppm): 9.07 (d, J = 8.0 Hz, 2H), 8.79–8.76 (m, 0.8H), 8.46–8.43 (m, 3.2H), 8.04– 8.00 (m, 1.3H), 7.93–7.88 (m, 2H), 7.85–7.80 (m, 2H), 7.67– 7.64 (m, 0.7H), 7.51–7.46 (m, 2H), 2.23–2.11 (m, 8H), 1.00– 0.52 (m, 60H). HRMS (MALDI) m/z: calc. for C78H82F2N4O2S2: 1,209.5891; found: 1,209.5920. Elemental analysis (%) calc. for C78H82F2N4O2S2: C, 77.45; H, 6.83; N, 4.63; found: C, 77.50; H, 7.26; N, 4.24.

#### OSC Fabrication and Characterization

PSCs were fabricated by using a device configuration of indium tin oxide (ITO)/ZnO/active layer/MoO3/Ag. The ITO glass was cleaned by sequentially in detergent, deionized water, acetone, and isopropanol for half an hour each and dried for more than 12 h in an oven. Then, the ITO glass was subjected to ultraviolet/ozone treatment for 15 min. Later on, a ZnO precursor solution (0.23 M in 2-methoxyethanol) was spincoated on the ITO glass at 3,000 rpm for 50 s. On a hot plate (130◦C), the obtained films were heated for 10 min first, then they were annealed by an oven (200◦C) for 1 h. The blend ratios of the active layer (PBDB-T:acceptor) were fixed at 1:1 by weight. The donor/acceptor blends were dissolved in chlorobenzene with a total concentration of 18 mg/mL and spin-coated on top of the ZnO film (ca. 30 nm) in the glovebox. Finally, on the active layer, 10 nm of MoO<sup>3</sup> film was deposited followed by a further deposition of Ag film (100 nm). The active area of OSC devices was 6 mm<sup>2</sup> .

Solar cell characterization was tested under AM 1.5 G irradiation (100 mW cm−<sup>2</sup> ) from an Oriel Sol3A simulator (Newport) with a NREL-certified silicon reference cell. J–V measurements were carried out in air using a Keithley 2440 source measurement unit. External quantum efficiency (EQE) data were collected by a Newport EQE measuring system.

#### Instruments and Measurements

<sup>1</sup>H NMR was measured in CDCl<sup>3</sup> on a Bruker AVANCEspectrometer. Elemental analysis of the non-fullerene acceptors was obtained on an Elementar Vario EL Cube analyzer. UV– Vis absorption spectra for all the samples were performed on a Perkin-Elmer Lambda 365 spectrophotometer. Linear emission spectra for pure films or blended films were obtained by using a FLS920 spectrophotometer. Atomic force microscopy (AFM) was conducted in a tapping mode with a Bruker Nanoscale V station. A Bruker Dektak XT surface profilometer was used to test the thickness of thin films in this work. A threeelectrode CHI 604E electrochemical workstation was used to run the cyclic voltammetry (CV) using Bu4NPF<sup>6</sup> solution (0.1 M in acetonitrile) and a scan rate of 100 mV s−<sup>1</sup> . The solid films were precipitated on a Pt plate through dipping the Pt plate into the corresponding chloroform solutions and then took out for drying. Ag/AgNO<sup>3</sup> and a Pt wire were chosen as the reference electrode and the counter electrode, respectively. The LUMO and HOMO energy levels of films made by the small molecule were calculated by using the following equations:

$$E\_{\rm HOMO} = -\left(\varphi\_{\rm ox} + 4.82\right)\left(\rm eV\right)$$

$$E\_{\rm LUMO} = -\left(\varphi\_{\rm RED} + 4.82\right)\left(\rm eV\right)$$

Agilent 4155C semiconductor parameter analyzer was used to conduct the mobility measurements. Electron and hole mobilities were determined by using the space charge limited current model (SCLC) with an electron-only diode configuration of ITO/ZnO/active layer/Ca/Al and an hole-only diode configuration of ITO/PEDOT:PSS/active layer/Au, using current-voltage measurements in the range of –(3–10) V in the dark. The SCLC mobility was calculated by fitting the J-V curves to the Mott-Gurney relationship:

$$J = \frac{9}{8} \varepsilon\_{r\varepsilon 0} \mu \frac{V^2}{L^3}$$

Where ε<sup>0</sup> is the permittivity of free space (8.85 × 10−<sup>12</sup> F m−<sup>1</sup> ), εr is the dielectric constant of the active layer material (assumed to be 3), µ is the electron or hole mobility, L is the active layer thickness, V is the voltage drop across the electronor hole-only device (Vappl – Vbi, where Vappl is the applied voltage, and Vbi is the built-in voltage induced by the work function difference of the two electrodes). The electron/hole mobilities can be determined according to the slope of the J <sup>1</sup>/2–V curves.

#### AUTHOR CONTRIBUTIONS

QZ conceived the experiments. MZ and YM were primarily responsible for the experiments. MZ and QZ wrote the manuscript. All authors discussed the results.

### FUNDING

This work was supported by the National Natural Science Foundation of China (Nos. U1605241, 51703226, 51561165011), the Key Research Program of Frontier Sciences, CAS (No. QYZDB-SSW-SLH032), the CAS/SAFEA International Partnership Program for Creative Research Teams, and the Strategic Priority Research Program of the Chinese Academy of Sciences, Grant No. XDB20000000.

#### REFERENCES


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Zhang, Ma and Zheng. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

## Small-Molecule Electron Acceptors for Efficient Non-fullerene Organic Solar Cells

#### Zhenzhen Zhang, Jun Yuan, Qingya Wei and Yingping Zou\*

*College of Chemistry and Chemical Engineering, Central South University, Changsha, China*

The development of organic electron acceptor materials is one of the key factors for realizing high performance organic solar cells. Compared to traditional fullerene acceptor materials, non-fullerene electron acceptors have attracted much attention due to their better optoelectronic tunabilities and lower cost as well as higher stability. Non-fullerene organic solar cells have recently experienced a rapid increase with power conversion efficiency of single-junction devices over 14% and a bit higher than 15% for tandem solar cells. In this review, two types of promising small-molecule electron acceptors are discussed: perylene diimide based acceptors and acceptor(A)-donor(D)-acceptor(A) fused-ring electron acceptors, focusing on the effects of structural modification on absorption, energy levels, aggregation and performances. We strongly believe that further development of non-fullerene electron acceptors will hold bright future for organic solar cells.

#### Edited by:

*Chuanlang Zhan, Institute of Chemistry (CAS), China*

#### Reviewed by:

*Qingdong Zheng, Fujian Institute of Research on the Structure of Matter (CAS), China Francesca Di Maria, Istituto di Nanotecnologia (NANOTEC), Italy Lei Ying, South China University of Technology, China*

> \*Correspondence: *Yingping Zou yingpingzou@csu.edu.cn*

#### Specialty section:

*This article was submitted to Organic Chemistry, a section of the journal Frontiers in Chemistry*

Received: *17 June 2018* Accepted: *23 August 2018* Published: *18 September 2018*

#### Citation:

*Zhang Z, Yuan J, Wei Q and Zou Y (2018) Small-Molecule Electron Acceptors for Efficient Non-fullerene Organic Solar Cells. Front. Chem. 6:414. doi: 10.3389/fchem.2018.00414* Keywords: organic solar cells, efficiency, small molecule, fused ring, perylene diimide

## INTRODUCTION

Energy is the important foundation of human survival and economic development. With the rapid development of the global economy, the traditional non-renewable fossil energy such as coal, petroleum, and natural gas appears to be decreasing, and the burning of fossil fuels brings about greenhouse gases such as carbon dioxide and other chemical pollutants. At the background of energy crisis and environmental pollution, the development of clean and renewable energy has become the world's urgent requirements (Zhan et al., 2015). The emerging new energy sources include nuclear, solar, wind, hydro, and tidal energy. Among them, solar energy has the advantages of being clean, non-polluting, widely distributed, and non-exhaustive. It meets the requirements of sustainable development in the world today. There are three main ways to use solar energy: solar to thermal energy conversion, photoelectric conversion and photochemical conversion. Presently, the photoelectric conversion is one of the world focuses. The development of photovoltaic cells has become a promising solution for transforming solar energy into electricity. The first photovoltaic cell based on monocrystalline silicon materials was invented by Bell Laboratories (Chapin et al., 1954). Since then, the performance based on inorganic semiconductor solar cells began to get improved. However, the shortcomings of the complicated preparation process, high production cost, inflexibility in processing limited the preparation and application of large-area inorganic solar cells. On the contrary, organic solar cells (OSCs) have some merits of light weight, low cost, mechanical flexibility (Sariciftci et al., 1992; Li and Zou, 2008; Krebs, 2009; Li, 2011; Li et al., 2012; Heeger, 2014). More importantly, organic raw materials are abundant and the photoelectric properties can be modified by simple and feasible organic synthesis.

Zhang et al. Small-Molecule Electron Acceptors

Nowadays, the typical OSCs active layers are bulk heterojunction (BHJ) structures, which are based on percolate and continuous electron donor (D) and electron acceptor (A) blend films. The working mechanism of OSCs is generally divided into four steps: (1) The active layer absorbs photons and then forms excitons (electron-hole pairs); (2) Exciton diffuses to D/A interface; (3) Exciton dissociates into free holes and electrons; (4) Free holes and electrons transport to the corresponding electrodes through the donor and acceptor channels, and subsequently are collected by electrodes. Finally, the photocurrent is formed in the external circuit (Lin and Zhan, 2014). To achieve high efficiency, an amount of small molecule/polymer donor materials have been developed, the power conversion efficiency (PCE) of fullerene OSCs had made a dramatic progress with values over 10% after decades of the tireless efforts by scientific community (Zhao et al., 2016b). However, the further development of fullerene-based OSCs encounters bottlenecks due to the inherent defects of fullerene derivatives, such as limited tunability of absorption and energy level, costly preparation and purification as well as poor stability.

In contrast to fullerene derivatives, non-fullerene acceptors (NFAs) can be modified by classical synthesis strategies, for example donor (D)-acceptor (A) conjugation, conformation locked and incorporation of functional groups, which is beneficial to adjusting crystallinity, electrical and optical properties. Although the first bilayered OSC is based on nonfullerene acceptor, the development of the electron acceptor lagged far behind of the donor materials in early studies (Kallmann and Pope, 1959). Early stage, rylene diimides derivatives, including perylene diimide (PDI) and naphthalene diimide (NDI), occupied the forefront of the non-fullerene materials. Before 2013, the PCEs were only about 1–3% (Bloking et al., 2011). After decades of mediocrity, Yao's group reported a novel acceptor (bis-PDI-T-EG), the performance achieved first breakthrough with PCE of 4.03% (Zhang et al., 2013). The second progress was the discovery of the ITIC, when blended with PTB7-Th, the device delivered a PCE of 6.8%, which is higher than 6.05% efficiency of PTB7-Th: PC61BM based devices (Lin et al., 2015b). This inspiring study showed that the performance of non-fullerene solar cells is expected to catch up or even be superior to fullerene based solar cells. In recent years, nonfullerene solar cells have once again revived and become a hot topic in photovoltaic researches (Liang et al., 2017; Zhang et al., 2018a). Currently, the highest efficiency has exceeded 14% for single-junction NF-OSCs and 15% for tandem NF-OSCs (Che et al., 2018; Zhang et al., 2018b).

To achieve high performance NF-OSCs, the primary factor to consider is the design and synthesis of acceptor materials. Generally, a promising acceptor should meet the following criteria:

a) The acceptor should have complementary absorption with the donor as much as possible to increase photon utilization, which is beneficial for achieving high external quantum efficiency (EQE) and short-circuit current density (Jsc). For example, to better match high performance narrow bandgap donors, wide or ultra-narrow bandgap acceptors should be designed. Narrow bandgap acceptors are likely to work well with wide or medium bandgap donors. In addition, the photocurrent can be formed by generation of excitons from both donor (channel 1) and acceptor (channel 2). Thus, apart from the complementary absorption with donors, the optical absorptivity of the NFA is also important (Nielsen et al., 2015; Cheng et al., 2018; Wadsworth et al., 2018).


Except for the above mentioned prerequisites, simple synthesis and low cost NFAs are beneficial for practical applications.

NFAs are classified into two major classes of polymers and small molecules. Small molecule NFAs have been intensively investigated by blending with polymer and small molecule donor materials (**Figure 1**) owing to their features over their polymeric counterparts, which include clear molecular structures, high purity and batch-to-batch stability (Roncali, 2009). In this review, we will focus on discussing the small molecule NFAs developed for high efficiency OSCs in recent years. **Figure 1** listed chemical structures of the polymer donors referred herein.

## PDI BASED SMALL MOLECULE ELECTRON ACCEPTORS

PDI derivatives have attracted considerable attention as NFAs since they possess excellent optical absorption, similar energy levels to fullerenes, high electron mobility as well as good stability. Moreover, these properties can be easily tailored through the substituent groups on the bay region or on the nitrogen atoms (Zhao et al., 2013). The major problem is that PDI

units tend to form large aggregates domains, which is more than the exciton diffusion length, led to less exciton separation and poor performance (Zhang et al., 2012). Therefore, it's essential to design and synthesize high performance PDI derivatives with moderate aggregations for effective exciton separation and charge transport.

Until now, several chemical modification methods have been used to reduce the self-aggregation of PDI and achieve good results (**Figure 2** and **Table 1**). The initial design strategy is to introduce alkyl side chains on the nitrogen position or the ortho position. A series of alkyl-substituted PDI acceptor (PDI-1, PDI-2, PDI-3) was reported and studied, with improved solubility during solution processing and weakening the crystallinity to some extent. When blending with P3HT, the performance was poor (Kamm et al., 2011). But mapping other donors and optimizing the conditions for device fabrications, the OSCs based on p-DTS(FBTTh2)<sup>2</sup> : PDI-2 blend film showed the PCE of 5.13% (Chen et al., 2015). TP-PDI was a bay-substituted tetraphenyl functionalized PDI derivative, which suppressed the strong aggregation tendency due to steric hindrance effects. While blended with PTB7-Th, a PCE of 4.1% was achieved (Cai et al., 2015).

Moreover, PDI dimers can also reduce their crystallization tendency. Two PDI units were brought together using hydrazine as a linker, giving H-di-PDI. The perylene units are oriented perpendicular to each other, alleviated the aggregation. A PCE of 2.78% has been achieved when PBDTTT-C-T was used as donor material (Rajaram et al., 2012). Wang designed three PDI dimers (s-diPBI, d-diPBI, and t-diPBI), with singly-linked, chiral doublylinked, and graphene like triply-linked between two PDI units, respectively. Blended with PBDTTT-C-T, s-diPBI delivered the best photovoltaic performance up to 3.63%, which is the result of a flexible structure with a twist angle of about 70◦ (Jiang et al., 2014). Subsequently, s-diPBI was modified by inserting thiophene and selenophene units in the bay positions, affording two new acceptors (SdiPBI-S and SdiPBI-Se). Both acceptors have a more twisted configuration and higher LUMO energy levels due to big and loose outmost electron clouds of sulfur and selenium. Moreover, the selenium is more polarized than sulfur, which is helpful to improving intramolecular interactions and carrier mobility. Thus, SdiPBI-Se exhibited a higher performance with PCE of 8.42 vs. 7.16% for SdiPBI-S when blended with same donor PDBT-T1 (Sun et al., 2015; Meng et al., 2016b). 7b was obtained by incorporating nitrogen heteroatom in the bay position of PDI to further study the potential of bay-linked PDI dimers. By modulating the N-R functional group, the selfassembly of acceptor would be changed. When the alkyl chain of the bay position is ethyl, the device demonstrated a best PCE of 7.55% with P3TEA as donor. More significantly, N-annulation of the PDI derivative can be synthesized in gram scale without the need for purification using column chromatography (Hendsbee et al., 2016).

Aside from direct linking two PDI units, twisted structure can be also achieved by using functional groups as the linkage. Bis-PDI-T-EG produced small phase domains with a size of ∼30 nm. A promising PCE of 4.03% was obtained due to significant reduction of the aggregation (Zhang et al., 2013). This is the first time the PCE more than 4% in non-fullerene OSCs, demonstrated that the introduction of the π linkage is an effective method to improve photovoltaic performance, the synthetic steps of Bis-PDI-T-EG were shown in **Scheme 1**. Almost at the same time, another acceptor (SF-PDI2) featuring spirobifluorene linker was developed. When P3HT was used as donor, the PCE of 2.35% was achieved. The results demonstrated that steric-demanding substituents on PDI units was able to suppress self-aggregation and crystallization (Yan et al., 2013). Moreover, donor material PffBT4T-2DT can match particularly well with SF-PDI<sup>2</sup> with complementary absorption and small driving force. The NF-OSCs possessed a high PCE of 6.3% (Zhao et al., 2015). After that, another NF-OSCs based on P3TEA: SF-PDI<sup>2</sup> were fabricated, exhibiting ultrafast and efficient charge separation despite of a negligible driving force, with an excellent PCE of 9.5% (Liu et al., 2016b). A twisted PDI dimers (IDT-2PDI) with bulky indacenodithiophene as a bridge is developed as an electron acceptor. The OSCs based on BDT-2DPP: IDT-2PDI blend film showed a PCE of 3.12% (Lin et al., 2014a).

Undoubtedly, both approaches to reduce the strong πstacking aggregation by connecting two PDI units with single bond or linker have been efficient and shown improved photovoltaic performance, compared to traditional PDI derivatives. However, It must be admitted that the twisting of the structure will hinder the effective π-π stacking and diminish the charge transport. Thus, the trade-off between high electron mobility and effective exciton dissociation need to be solved in order to achieve excellent performance (**Figure 3** and **Table 2**). Based on these considerations, two PDI dimers substituted at the α position (αPBDT) and β position (βPBDT) with benzodithiophene (BDT) unit were synthesized. The absorption revealed αPBDT have stronger intermolecular π-π stacking and higher packing order than βPBDT due to good planarity. The OSCs based on αPBDT as acceptor demonstrated a PCE of 4.92%, which is 39% higher than that of βPBDT counterparts, which is consequence of higher electron mobility and more efficient exciton dissociation in the αPBDT-based devices (Zhao et al., 2016a). A class of fused but helical PDI oligomers (hPDI, hPDI3, hPDI4) with ethylene group as bridges were designed and studied, which all possess strong light absorption, weak aggregation trendency and both hole and electron can be generated in both the donor and acceptor phases. The device based on PTB7-Th: hPDI4 reached a highest PCE of 8.3% (Zhong et al., 2014, 2015). A series of fused heterocycle PDI derivatives with different chalcogen atoms of O, S and Se (FPDI-F, FPDI-T, FPDI-Se,) were reported. Compared to unfused PDIs, fused PDIs increased effective conjugation and reduced reorganization energy helpful for high charge mobility, while maintaining nonplanar structure for suppress the strong aggregation. Moreover, the device based on FPDI-T showed a best photovoltaic performance with a PCE of 6.72% because of smallest twist angle leading to high packing order and close π-π stacking (Zhong et al., 2016). The first triplet tellurophene-PDI based acceptor (BFPTP) possessed long exciton lifetime and diffusion distances for efficient exciton dissociation rather than recombination. Thus, the PBDB-T: BFPTP blended films delivered a PCE of 7.52% (Yang et al., 2018a). A fused and twisted PDI derivative with twisted thieno[2,3-b]thiophene (TT) as linker (cis-PBI) was reported. When blended with PBDB-T, the OSCs demonstrated a high PCE of 7.6% as a result of high electron mobility and isotropic crystalline properties of electron acceptor (Jiang et al., 2017). A fused PDI derivative with indacenodithieno [3,2-b]thiophene (IDTT) as central core (FITP) maintaining rigid conjugated skeleton and hexylphenyl side chains of IDTT hindered the large crystallites. The devices

based PTB7-Th: FITP exhibited a high PCE of 7.33% due to elevated LUMO and superior electron mobility (Li et al., 2016b). Compound 3 with a planar conformation utilized weak electron acceptor (thieno-pyride-thieno-isoquinoline-dione) bridge for the lateral PDIs. When blended with PTB7-Th, the devices delivered a PCE of 5.03% (Carlotti et al., 2018).

In general, the three-dimensional (3D) or quasi-3D PDI derivatives have good compatibility with polymer donors and 3D charge-transporting channel. A star-shaped PDI acceptor (S(TPA-PDI)) with a triphenylamine (TPA) core displayed weak molecular aggregation and strong absorption as well as matched energy levels with PBDTTT-C-T. A PCE of 3.22% was achieved with 5% 1,8-diiodooctane (DIO) solvent additive (Lin et al., 2014b). A pyrene-fused PDI derivative (TPAPPDI) possessed upshifting LUMO energy level and low bandgap. The devices exhibited a PCE of 5.10% with ultra-high Voc of up to 1.21V (Zhan et al., 2017). Two twisted propeller configuration PDI derivatives (TPH and TPH-Se) were developed. The investigations indicated that TPH-Se possessed more compact 3D network assembly due to the Se. . . O interactions. The PDBT-T1: TPH-Se solar cell showed a relatively high PCE of 9.28% while 8.28% for TPH based polymer solar cells (PSCs) (Meng et al., 2016a). Three PDI tetramers (TPC-PDI4, TPE-PDI4, and TPPz-PDI4) with twisted 3D structure were systematically studied through the relationship between structure and performance of 3D acceptor. The results revealed that intramolecular twist angle changed as the sequence of TPPz-PDI<sup>4</sup> < TPE-PDI<sup>4</sup> < TPC-PDI4. Although TPPz-PDI<sup>4</sup> showed the strongest aggregation, it still had fine phase separation and effective charge transfer, and therefore, the highest PCE of 7.1% was obtained ascribed to high electron mobility (Lin et al., 2016a). Further transformed planar SF-PDI<sup>2</sup> into a 3D molecular conformation created SF-PDI4. SF-PDI<sup>4</sup> demonstrated a 3D interlocking geometry, which prevented excessive rotation and reinforcing conformational uniformity. The PCE of the PV4T2FBT: SF-PDI<sup>4</sup> based devices was 5.98% (Lee et al., 2016). A PDI acceptor (TPB) exhibited cross-like


TABLE 1 | Summary of the photophysical and photovoltaic properties of selected perylene diimide-based electron acceptors from PDI-1 to IDT-2PDI.

*S stands for the mobility measured by the space charge limited current (SCLC) method and O for the organic field effect transistor (OFET)method; N stands for the neat film and B for the blended film.*

molecular conformation but still partially conjugated with the BDTTh core. The PTB7-Th: TPB based solar cells achieved a PCE of 8.47% due to better conjugation and planarity (Wu et al., 2016). A star-shaped PDI derivative (PBI-Por) with porphyrin as central core was studied. Because porphyrin showed the large conjugated macrocycle and three characteristic absorption bands in the visible and NIR regions, the non-fullerene PSCs based on PBDB-T: PBI-Por blend films achieved a PCE of 7.4% (Zhang et al., 2017a).

## ACCEPTOR-DONOR-ACCEPTOR (A-D-A) FUSED-RING ELECTRON ACCEPTORS

In recent years, A-D-A conjugated structures seem to be the most promising class of NFAs. The conjugated push-pull structure containing electron-rich and electron-poor units induces strong intramolecular charge transfer, which is beneficial to reducing the optical band gap. Moreover, variation of the donor or acceptor units can be used to regulate the HOMO or LUMO energy levels. The fused ring backbone facilitates electron delocalization and broadens absorption, and it can prevent the torsion or conformational transition of the molecular skeleton and enhance carrier mobility. The presence of the side chains on the conjugated backbone firstly ensures solution processing, in addition, and reduces molecular stacking, inhibits strong selfassembly and large phase separations.

### Fused Tricyclic Small Molecule Acceptors

The three-membered ring is fused ring structure with the smallest size (**Figure 4** and **Table 3**). Dibenzosilole (DBS) unit was the firstly to be used in A-D-A type NFAs, due to good electron-transporting properties, in addition to low-lying LUMO energy levels of silole moiety deriving from effective interactions between σ ∗ -orbital of the silicon-carbon bond and π ∗ -orbital of the butadiene. Diketopyrrolopyrrole (DPP) exhibits excellent



*S stands for the mobility measured by the space charge limited current (SCLC) method and O for the organic field effect transistor (OFET)method; N stands for the neat film and B for the blended film.*

light absorption and strong electron-withdrawing properties. A novel linear NFA (DBS-2DPP) based on DBS as central core and DPP as end group was reported in 2013, which possesses strong and broad absorption and moderate electron mobility. When P3HT was used as donor, the blended film formed fibrous nano-interpenetrating network, leading to a PCE of 2.05% (Lin et al., 2013). Based on this strategy, another DPP derivative (F(DPP)2B2) was developed. Consisting of fluorene as the core and two benzene end-capped DPP as the terminal, F(DPP)2B<sup>2</sup> possessed excellent light-harvesting capability, moderate energy levels and good charge-transporting with the value of 2.8 × 10−<sup>4</sup> cm<sup>2</sup> ·V −1 ·s −1 . While P3HT was also used as a donor material, the devices delivered a PCE of 3.1% with an extremely high Voc (Shi et al., 2015). Because of concise synthesis and ready availability of fluorene, two isomeric acceptors (F8IDT and FEHIDT) using 2,3-dihydro-1H-indene-1,3-dione (ID) as end group were synthesized. The density functional theory calculations have shown that the LUMO energy of FxIDT were similar to that of fullerene derivatives, demonstrating FxIDT can be potentially used as acceptor materials. The devices based on P3HT: FxIDT blend films showed different performance (1.67% for F8IDT; 2.43% for FEHIDT). The main reason can be attributed to the difference of LUMO energy levels and the degree of electronic coupling between molecules, leading to various and low Voc (Winzenberg et al., 2013). To reach a higher Voc, 3-ethylrhodanine is a reasonable choice as end group relative to ID due to weaker electron-withdrawing nature. Two rhodanine-based acceptors (Cz-RH and Flu-RH) were obtained. Cz-RH and Flu-RH possessed high-lying LUMO energy levels of −3.50 and −3.53 eV, respectively, compared to F8IDT, resulting in an excellent Voc of 1.03 V. The devices exhibited a good photovoltaic performance with PCE of 3.08% for P3HT: Flu-RH and 2.56% for P3HT: Cz-RH. The difference was mostly attributed to the Jsc, which originated from the maximum EQE intensity of 40% and a more efficient charge transfer from donor to acceptor in P3HT: Flu-RH blend films with higher photoluminescence (PL) quenching efficiencies of 86.7% (Kim et al., 2014). Subsequently, another acceptor FBR, bearing fluorene core and 3-ethylrhodanine end group but flanked by electron-deficient benzothiadiazole (BT) rather than thiophene spacer, was reported. BT as linker extends the conjugation and enhances charge transport. FBR exhibited a nonplanar 3D molecular structure, which is helpful to suppressing large aggregation and achieve efficient exciton separation confirmed by PL quenching efficiencies of over 90%. When blended with P3HT, the device showed a PCE of 4.11% with high Voc of 0.82 V as a result of high LUMO energy level compared to PC60BM. It is a pity that the Jsc and FF of P3HT: FBR is inferior to those of P3HT: PC60BM, which can be caused by the difference of devices thickness and faster geminate

recombination. Moreover, large overlapping absorption in P3HT: FBR blend films limited the more photocurrent generation. To harvest more photons across the solar spectrum, a low bandgap polymer PffBT4T-2DT was used to replace wide bandgap P3HT as donor, the device achieved PCE up to 7.8% with improved Jsc. The increase of Voc is originated from deep HOMO energy levels of donor (Holliday et al., 2015; Baran et al., 2016). Except for the optimization of donor materials, modification of acceptor materials also play an important role in improving the light absorption properties. FRd<sup>2</sup> was developed based on FBR, but incorporation of furan spacer between BT and rhodanine end group, which help extending π-conjugation and reducing optical ban gaps. Employing PTB7-Th as donor, the device exhibited a PCE of 9.4% with Jsc of 15.7 mA cm−<sup>2</sup> , which is the reported highest performance for fluorene-based acceptors so far (Suman et al., 2017). Dicyanovinyl (DCV) unit was also an excellent electron-accepting motif to build A-D-A NFAs, because target molecules containing DCV can induce intramolecular charge transfer and promote planarity, which tends to achieve improved carrier mobilities. A set of acceptors (FBM, CBM and CDTBM) flanked by BT as spacer and DCV as end group were systematic studied. FBM and CBM possessed similar electronic properties, but CDTBM exhibited res-shifted absorption and deep LUMO level. Thus, CDTBM obtained higher Jsc and FF but lower Voc, leading to similar performance with PCE of ∼5% (Wang et al., 2016). Another stronger electron-withdrawing end-capping group, 2-(6-oxo-5,6-dihydro-4H-cyclopenta[c]thiophen-4-ylidene)malononitrile (IC), built upon the structure of DC, could lower the band gap of the acceptor. A easily synthesized and high yield acceptor DICTF, bearing fluorene central block and thiophene spacers as well as IC terminal group, was reported in 2016. DICTF has strong and complementary absorption in the visible region and matched energy levels with PTB7-Th. The devices delivered a PCE near 8% (Li et al., 2016a). Benzo[1,2-b:4,5-b ′ ]dithiophene (BDT) and its derivatives as electron-rich units in conjugated polymers have been well studied and have demonstrated outstanding results. Recently, BTCN-M, in which 4,8-bis-thiophene-substituted benzo[1,2-b:4,5-b ′ ]dithiophene (BDT-T) as central block and IC as end group linking with BDT-T by thiophene spacer, was synthesized. Due to high steric hindrance caused by alkyl side groups in the BDT unit, BTCN-M showed weak π-π stacking, tending to act as acceptor material. Therefore, the devices based on BTCN-M: PBDB-T blended films exhibited a outperforming PCE of 5.89% whereas only 0.29% for that of BTCN-M: PC71BM blended films (Liu et al., 2018b). Cross-conjugated small


TABLE 3 | Summary of the photophysical and photovoltaic properties of fused tricyclic small molecule acceptors.

*S stands for the mobility measured by the space charge limited current (SCLC) method and O for the organic field effect transistor (OFET)method; N stands for the neat film and B for the blended film*

molecular acceptor PDIBDT-IT, which combined the advantages of both the A-D-A and PDI type acceptors, exhibited broad absorption band ranging from 300 to 700 nm. The devices based on PDIBDT-IT: PTB7-Th blended films exhibited a PCE of 6.06% (Liu et al., 2018e). It is worth noting that these electron-rich central core were based on symmetrical units, however, NFAs based on asymmetrical cores were promising acceptor materials. Such as, ITDI and ITBR incorporating indenothiophene as core delivered PCEs of 8.00 and 7.49% when blended with PBDB-T and PTB7-Th, respectively (Kang et al., 2017; Tang et al., 2017). Although this type of fused-ring acceptor has made some progress, most of acceptors exhibited wide or medium band gaps absorption with poor spectral coverage and encountered suboptimal morphologies, therefore, relatively low Jsc and FF were obtained.

## Fused Pentacyclic Small Molecule Acceptors

Indacenodithiophene (IDT) is the most representative fused pentacyclic donor unit in A-D-A acceptors due to rigid and coplanar structure for good absorption and excellent charge mobility. Moreover, the side chain substituents of the conjugated block can ensure the solution processability and inhibit strong self-assembly of molecules (**Figure 5** and **Table 4**).

IEIC with IDT as the core flanked by thiophene spacers and IC end groups was studied (**Scheme 2**). IEIC showed strong absorption in the 500–750 nm region with an extinction coefficient of 1.1 × 10<sup>5</sup> M−<sup>1</sup> cm−<sup>1</sup> at 672 nm and relatively high electron mobility of 2.1 × 10−<sup>4</sup> cm<sup>2</sup> V −1 S −1 . The blend films of PTB7-Th as donor and IEIC as acceptor showed nanoscale interpenetrating morphology, thereby a PCE of 6.31% was achieved (Lin et al., 2015c). The limitation of the PCE was poor FF and Jsc, which mainly came from the big overlapped absorption profiles and imbalanced charge mobility of active layer. With this in mind, when IEIC was laterly blended with a large bandgap donor polymer PffT2-FTAZ-2DT, the PCE reached 7.30%. The improved PCE can be attributed to complementary absorption and balanced charge mobility as well as moderate phase domain size (Lin et al., 2015a). IEIC was the first acceptor material using IDT as central core, and exhibited good performance at that time, providing a good theoretical basis for the later fused pentacyclic small molecule acceptors. The synthetic steps of ITIC were shown in **Scheme 2**. Using larger and looser outermost electron cloud, selenium atoms to replace sulfur atoms afford IDSe-T-IC, which possessed decreased bandgap of 1.52 eV and improved LUMO energy level as well as increased carrier mobility. Thus a high PCE of 8.58% was obtained with a large bandgap polymer J51 as donor (Li et al., 2016d). Designing and synthesizing low bandgap acceptor materials can also make better use of solar spectrum to absorb larger fractions of photons. IEICO, replacing alkyl groups with alkoxy groups, was reported with E<sup>g</sup> of 1.34 eV. Introduction of alkoxyl chains increased the HOMO energy level but had little effect on the LUMO level. By employing PBDTTT-E-T as the donor, the IEICO-based devices delivered a high PCE of 8.4% with an increased Jsc of 17.70 mA cm−<sup>2</sup> (Yao et al., 2016). A further development of IEICO obtained IEICO-4F by introducing F atoms in the IC end groups. When blended with a narrow bandgap polymer PTB7- Th, a PCE of 12.8% was achieved (Wang et al., 2018a). i-IEICO-4F, an isomer of IEICO-4F by attaching the end groups in the 4-position instead of 5-position at the neighboring spacers, is a twisted configuration, resulting in blue shifts and complementary absorption with the wide-bandgap polymer J52. The devices based on i-IEICO-4F delivered an excellent PCE of 13.18%

(Song et al., 2018). Introducing alkoxyl side chains at the central core rather than thiophene spacer provided IDTT2F, which exhibited excellent solubility and ordered molecular packing, resulting in a PCE of 12.79% blended with PBDB-T (Liu et al., 2018d). IDT-BOC6 was also synthesized by using IDT as central core and IC as end groups, but bis(alkoxy) substituted benzene ring as spacer. Incorporation of alkoxyl groups not only increased the LUMO energy levels but also induced conformational control and enhanced the planarity. IDT-BOC6 locked by intramolecular noncovalent interactions displayed a broad absorption spectrum, high electron mobility and weak nonradiative recombination. The devices based IDT-BOC6 afforded a PCE of 9.6% with PBDB-T as donor (Liu et al., 2017b). ITOIC-2F was also included noncovalently



*S stands for the mobility measured by the space charge limited current (SCLC) method and O for the organic field effect transistor (OFET)method; N stands for the neat film and B for the blended film.*

conformational locking, the corresponding devices delivered a PCE of 12.17% when blended with PBDB-T (Liu et al., 2018f).

A planar electron acceptor IDT-2BR was synthesized, which IDT was as core flanked by BT as the first electron-withdrawing group and the second electron-deficient 3-ethylrhodanine units on the periphery. The P3HT: IDT-2BR blended films exhibited clear interpenetrating networks and moderate phase separation with the addition of 3% CN. The devices achieved a PCE of 5.12% with a high FF of 68% due to balanced charge mobilities (Wu et al., 2015). IFBR-p was synthesized by incorporating fluorine atoms on the BT unit of IBT-2BR.The OSCs based on PTzBI: IFBR-p blend film showed a PCE as high as 7.44% as result of intermolecular and intramolecular interactions induced by C–H. . . F non-covalent force (Zhong et al., 2017). BTA1, containing benzo[d][1,2,3]triazole (BTA) as spacer, is analogous to the IBT-2BR acceptor. BTA was a weaker electron-deficient unit than BT, which would make it possess a higher LUMO energy level. Also while blended with P3HT, the device gives the PCE of 5.24% with a high Voc of 1.02 V (Xiao et al., 2017a). Meanwhile, quinoxaline (Qx) is the other weak electron-drawing unit and has been copolymerized with different electron-rich building block to get high performance D-A polymers (Yuan et al., 2017). Qx1, using Qx as bridge, was synthesized and explored. The devices based on P3HT: Qx1 blend films achieved a PCE of 4.03% with a Voc of 1.00 V (Xiao et al., 2018). ATT-1 can be considered as a IDT-2BR derivative, which used ester-substituted thieno[3,4 b]thiophene as spacer and 2-(1,1-dicyanomethylene) rhodanine as end group. ATT-1, adopting quinoidal resonance to extend the π-conjugation and enhance the absorption, exhibits a broad absorption with a high absorption coefficient of 1.2 × 10<sup>5</sup> L mol−<sup>1</sup> cm−<sup>1</sup> and slightly high LUMO energy level. When blended with PTB7-Th, the devices achieved a PCE of 10.07% after the addition of DIO. It's worthy to note that the PCE was only 4.46% without any post-treatment,. The investigations indicated that the addition of DIO provided an ideal morphology for efficient charge transport (Liu et al., 2016a). The design of ATT-1 wasfurther developed by substituting the thieno[3,4-b]thiophene spacer with thiophene-fused benzothiadiazole (BTT) unit as pbridge to obtain A2. The BTT unit connecting on the IDT core not only extend the conjugation length, but also stabilize the quinoid conjugation system, which resulted in red-shift absorption and low bandgap of 1.36 eV. Ultimately, the PCE of 9.07% was reached with an excellent Jsc of 20.33 mA cm−<sup>2</sup> (Xu et al., 2018a). IDT-BR with IDT core was designed and synthesized to solve the issues of FBR, including the large spectra overlap and poor charge percolation pathway. IDTBR had significantly red-shift absorption and tended to crystallize on length scales, meanwhile, O-IDTBR with linear alkyl chains was a more crystalline acceptor and had a further red-shift absorption. The resulting OSCs based on O-IDTBR achieved a

PCE of 6.4% while 6.05% for EH-IDTBR (Holliday et al., 2016). Alkyl and alkylaryl groups have been widely used as side chains of IDT to ensure solubility and suppress strong aggregations. Compared with alkylaryl units, alkyl substituents enable π-π stacking. However, alkyl substituted acceptor usually formed large domains, leading to incomplete exciton separation. Thus a new acceptor IDT-OB with asymmetric side chains was reported, which reduced strong self-assembly but still had close packing in film due to the existence of more configurationally isomers. As a result, 10.12% was reached for IDT-OB based devices without any post-treatment, while the PCE of 9.68% for IDT-2O, the performance based on IDT-2B was the worst, with only 6.42% efficiency (Feng et al., 2017c). By the replacement of the C-bridge of IDT with the Si-bridge, SiIDT-IC was obtained. Introduction of Si atom can result in a high-lying LUMO energy level to achieve a high Voc. When blended with PBDB-T, the devices delivered a PCE of 8.16% with high Voc of 0.92 V, but the performance is lower than the corresponding C-bridge acceptor (8.83%) due to inferiorJsc (Nian et al., 2018). Most of the reported NFAs are trans-arranged side chains linked with the central core. However, IDIDT-C8 with cis-arranged alkyl side chains had weaker π-π stacking than that of trans-arranged one. The blend films with PBDB-T as donor exhibited moderate molecular packing and film morphology, especially, IDIDT-C8 showed a good crystallinity and face-on orientation, resulting in an excellent PCE of 10.10% (Hou et al., 2018). Naphthalene diimide (NDI) was broadly used as acceptor unit due to their strong electron affinity and excellent electron transport properties. Naphtho[2,3-b]thiophene diimide (NTI) was connected on the IDT core to give a new acceptor IDT-NTI-2EH, which had red-shifted absorption and strong π-π stacking due to planar conjugated structure. The corresponding devices showed a PCE of 9.07% with PBDB-T as donor (Hamonnet et al., 2017).

### Fused Heptacyclic Small Molecule Acceptors

Indacenodithieno[3,2-b]thiophene (IDTT) was a further development of the IDT, from a fused pentacyclic to a fused heptacyclic structure (**Figure 6** and **Table 5**). The first A-D-A acceptor based on IDTT was ITIC, which IDTT was used as core directly flanked by IC. ITIC possessed strong and broad absorption in the visible and even NIR region, matched energy levels and good miscibility with PTB7-Th. The resulting OSCs based on PTB7-Th: ITIC blend films exhibited a promising PCE of 6.8%, which was better than that of the devices based on PTB7-Th: PC61BM (Lin et al., 2015b). Since then, the ITIC-based OSCs have shown high photovoltaic performance with multiple polymer donors (Bin et al., 2016; Xia et al., 2016; Yuan et al., 2016; Yu et al., 2017; Hu et al., 2018; Liu et al., 2018a; Xu et al., 2018b). ITIC was the first and a successful fused heptacyclic small molecule acceptors. The synthetic steps of ITIC were shown in **Scheme 3**. After that, much effort has been devoted to the modifications of ITIC structure, for example, by manipulating the aromatic core and changing the side chains as well as substituting the electron deficient end-capping groups (Wei et al., 2017; Alamoudi et al., 2018; Yang et al., 2018b).

It is well-established that the length, type and branch position of side chains play an important role in electronic properties and intermolecular self-assembly. C8-ITIC with four linear octyl side chains was reported for comparison with ITIC. C8- ITIC possessed a lower optical band gap, higher absorption coefficient and increased crystallinity. Blending with PFBDB-T, the devices delivered a PCE up to 13.2% while the devices based on ITIC showed only 11.71% efficiency (Fei et al., 2018). A new acceptor m-ITIC with meta-alkyl-phenyl side groups was synthesized to investigate the effects of side-chain isomerism. This work showed that m-ITIC had higher absorption coefficient, more crystallinity, and increased electron mobilities in comparison with ITIC. The resulting OSCs based on m-ITIC demonstrated a higher PCE of 11.77% than 10.57% for ITIC with a medium bandgap polymer J61 as donor (Yang et al., 2016). Actually, alkoxyphenyl side chains seemed more easily synthesized via simple etherification, which is beneficial for large scale production. The m-ITIC-OR bearing IDTT core with meta-alkoxyphenyl side chains and IC as end groups was reported. The HFQx-T: m-ITIC-OR blend films possessed



*S stands for the mobility measured by the space charge limited current (SCLC) method and O for the organic field effect transistor (OFET)method; N stands for the neat film and B for the blended film.*

high and balanced charge transport, negligible bimolecular recombination resulting in a promising PCE of 9.3%, which was higher than 9.07% efficiency of ITIC based devices under the same conditions (Zhang et al., 2017c). ITIC-SC2C6 based on branched 4-(alkylthio)-phenyl side chains was systematically explored. The investigations indicated that this acceptor had improved solubility, which is helpful for polymer donor to form nanofibrils. Consequently, the OSCs exhibited a PCE of 9.16% with PBDB-ST as donor (Zhang et al., 2017b). ITIC-Th, replacing phenyl side chains of ITIC with thienyl side groups, exhibited low LUMO energy levels, which can match with low bandgap and wide bandgap polymer donor. Additionally, ITIC-Th possessed high electron mobility owing to enhanced intermolecular interactions induced by S. . . S interaction. The OSCs were fabricated by blending ITIC-Th with low bandgap polymer PTB7-Th and wide bandgap polymer PBDB-T1, the PCE reached 8.7 and 9.6%, respectively (Lin et al., 2016b). ITIC and its derivatives has inevitable steric isomers between donor units and end groups linking by C = C covalent bond. To solve this defect, a definite molecular conformation ITC6-IC, which long alkyl chains were introduced into the terminal of IDTT, was synthesized and discussed. ITC6-IC exhibited planar structure, good solubility, high-lying LUMO energy levels and enhanced compatibility with donor materials. The blend films with PBDB-T as polymer donor and ITC6-IC as acceptor showed a fibril crystallization with bicontinuous network morphology

after thermal annealing. Consequently, the OSCs revealed a promising PCE of 11.61% with a high Voc of 0.97 V (Zhang et al., 2018d).

Density functional theory calculations reveal that the LUMO mainly delocalizes at the end groups of the ITIC derivatives. Furthermore, the side chains on the central donor units would hinder the tight stacking of the molecules, thus, the stacking of the end groups are likely to provide the main electron transport pathway. Indeed, atomistic molecular dynamic simulations referred that local intermolecular π-π stacking between the acceptor units of the ITIC film led to 3D molecular packing (Han et al., 2017; Yan et al., 2018). As a consequence, the reasonable regulation of ITIC terminal units is excepted to obtain higher LUMO energy level and better isotropic electron transport characteristics. A methyl group was introduced onto the phenyl of the IC to give IT-M, which elevates the LUMO energy levels due to the weak electron-rich properties of methyl. The devices demonstrated a high PCE of 12.05% with Voc of 0.94 V when blended with PBDB-T donor, which is the highest value for single-junction OSCs at that time (Li et al., 2016c). Then, replacing the methyl groups with the most electronegative fluorine atoms provided the new acceptor IT-4F. Although the fluorine atom resulted in low LUMO level, it had good crystallinity and high electron transport properties from noncovalent interactions of F. . . H,S. . . F and so on. The resulting OSCs based on PBDB-T-SF: IT-4F achieved a high PCE of 13.1% (Zhao et al., 2017). A similar effect of good crystallinities and noncovalent interactions can be found in other halogenated non-fullerene small molecular Cl-ITIC, Br-ITIC, and I-ITIC. The devices based on these halogenated acceptors showed PCEs of 9.5, 9.4, and 8.9%, respectively, which are higher than that of F-ITIC (8.8%) under the same circumstances (Yang et al., 2017). There are similar phenomena in other systems (Li et al., 2017c; Wang et al., 2018b). Apart from changing the substituents of the end groups, the modification of aromatic structure also attracted attention. Replacing the benzene of the IC with thiophene units gave isomers ITTC, ITCPTC, ITCT, and ITCC, which all show good potential in OSCs due to enhanced intermolecular π-π interaction induced by S. . . S interactions. When ITTC (or ITCPTC) as acceptor, the OSCs based on HFQx-T donor achieved a PCE of 10.4% (Zhang et al., 2017d) and 11.8% with PBT1-EH as donor (Xie et al., 2017). When PBDB-T as donor, the device based on ITCC delivered a PCE of 11.4% (Yao et al., 2017) and 11.27% for ITCT based devices (Liu et al., 2018c). Introducing methyl group onto the thiophene unit can also increase the Voc (Cui et al., 2017; Luo et al., 2018).


TABLE 6 | Summary of the photophysical and photovoltaic properties of fused eptacyclic small molecule acceptors.

*S stands for the mobility measured by the space charge limited current (SCLC) method and O for the organic field effect transistor (OFET)method; N stands for the neat film and B for the blended film.*

NFBDT is an isomer of ITIC, which based on a heptacyclic benzodi(cyclopentadithiophene) (FBDT) unit as core and IC as end groups. Due to symmetric and planar conjugated structure of BDT unit, the NFBDT possessed low bandgap of 1.56 eV. When blended with PBDB-T, the devices showed a PCE of 10.42% (Kan et al., 2017). Introducing 2-ethylhexyloxy on the BDT unit obtained BT-IC to further reduce E<sup>g</sup> (1.43 eV) by elevating the HOMO energy levels. The OSC fabricated by blending J71 and BT-IC achieved a PCE of 10.5% (Li et al., 2017d). ITIC2 with 5-(2-ethylhexyl) thiophene as side chains was a further development of NFBDT. The conjugated side chains is helpful for enhancing the absorption, intermolecular interaction and π-π stacking. Thus a high PCE of 11.0% was obtained while blended with FTAZ donor (Wang et al., 2017a). Fusing the thiophene spacers to the fluorene of DICTF afforded ladder acceptor FDICTF, leading to narrow bandgap, higher extinction coefficient and slightly higher LUMO energy level. In addition to extending the central core to enhanced absorption and intermolecular overlaps, extending conjugation end groups was also an effective strategy. For example, FDNCTF was obtained by replacing the benzene units of FDICTF with naphthalene units. The devices based on FDNCTF blended with PBDB-T delivered a higher PCE of 11.2% compared to 10.0% for FDICTF (Feng et al., 2017a; Qiu et al., 2017). Two acceptors DTCC-IC and DTCCIC-C17 based on dithienocyclopentacarbazole (DTCC) with different side chains were reported. Due to strong electron-donating properties and coplanar conjugated skeleton, both of them possessed strong absorption and good performance with PCE of 6.0% for DTCC-IC and 9.48% for DTCCIC-C17 (Cao et al., 2017; Hsiao et al., 2017). Two narrow bandgap acceptors ITTIC (1.46 eV) and ITVffIC (1.35 eV) were synthesized by incorporating thiophene units and double-bond as spacers, respectively. The PSCs based on PBDB-T1: ITTIC showed a PCE of 9.12% without any additives, and 10.54% for J71: ITVffIC blend films (Li et al., 2017b; Zhang et al., 2017e).

## Other Fused-Ring Small Molecule Acceptors

NITI, bearing an indenoindene core which is a carbon-bridged Estilbene with a centrosymmetry, exhibited a low optical bandgap of 1.49 eV and high extinction coefficient of 1.90 × 10<sup>5</sup> cm−<sup>1</sup> (**Figure 7** and **Table 6**). The corresponding devices delivered an excellent PCE of 12.74% by blending a large bandgap polymer PBDB-T due to the good charge transport property and proper phase separation (Xu et al., 2017). A further development of NITI was by replacing carbon-bridge with silicon-bridge to give NSTI. Bis-silicon-bridged stilbene (BSS) has rigid and coplanar structure and four side chains can suppress the strong aggregations. When blended with PBDB-T, the OSCs obtained a PCE of 10.33% with CN as additive (Zhang and Zhu, 2018). IHIC was a fused hexacyclic small molecule acceptor and the central core consisted of thieno[3,2-b]thiophene ring and two terminal thiophene. The thiophene-rich cores possessed symmetrical, rigid and coplanar structure, IC was used as the end group to construct push-pull structure, which is beneficial to inducing ICT and shifting the absorption spectrum to the NIR region. The IHIC showed strong NIR absorption with a narrow bandgap of 1.38 eV and a high electron mobility of 2.4 × 10−<sup>3</sup> cm<sup>2</sup> V −1 s −1 . The semitransparent OSCs achieved a PCE of 9.77% with a visible transmittance of 36% when blended with PTB7- Th (Wang et al., 2017b). Other similar thiophene-rich acceptors had strong NIR absorption and showed excellent performance (Jia et al., 2017; Dai et al., 2018; Li et al., 2018b). TPTT-IC was synthesized with a asymmetric thiophene-phenylene-thieno[3,2 b]thiophene-fused central core. Dipole-dipole interactions of asymmetric molecules tend to form strong π-π stacking on the face on orientation and achieved high FF (Li et al., 2018a). The OSCs achieved a PCE of 10.5% while blended with wide bandgap polymer PBT1-C (Li et al., 2018a). DTNIC8 can be considered as a further development of the IHIC by replacing thieno[3,2-b]thiophene with naphthalene. The angular-shaped central core dithienonaphthalene (DTN) had a more extended

π-conjugation system in comparison with IDT (Ma et al., 2017a). The devices based on PBDB-T: DTNIC8 blend films delivered a PCE of 9.03% with high FF of 73% attributed to well-defined film morphology (Ma et al., 2017b). IOIC2 is a naphthodithiophenebased fused octacyclic acceptor. IOIC2 had larger π-conjugation and stronger electron-rich properties, leading to higher LUMO energy levels, lower bandgap and higher electron mobilities, compared to naphthalene-based fused hexacyclic acceptor. Thus, the devices based on FTAZ: IOIC2 exhibited an excellent PCE of up to 12.3% (Zhu et al., 2018a,b). Another representative fused octacyclic acceptor was COi8DFIC with carbon-oxygenbridge, which had lower bandgap, higher electron transport properties due to higher electron-donating ability and more planar conjugated structure. When PTB7-Th was used as donor polymer, a high PCE of 12.16% was obtained with high Jsc of 26.12 mA cm−<sup>2</sup> (Xiao et al., 2017b). Compared to carbon-bridge, the molecules with nitrogen-bridge usually exhibited stronger electron-donating properties and better solution processing. The INPIC-4F with nitrogen-bridge and fluorinated IC as end group possessed narrow bandgap of 1.39 eV and high electron mobility as well as strong crystallinity. A PCE of 13.13% was achieved with PBDB-T as donor (Sun et al., 2018). To improve electrondonating ability, in addition to increasing the conjugation length, dithienopicenocarbazole (DTPC) possessed strong electron-rich properties by broadening the central core to two-dimensional conjugation system. DTPC-DFIC possessed low band gap of 1.21 eV and exhibited a PCE of 10.21% with PTB7-Th as donor (Yao et al., 2018). A novel small molecular acceptor (BZIC) bearing a D-A-D type thieno [3,2-b] pyrrolo-fused pentacyclic benzotrizole core was reported. A broad absorption spectra with optical bandgap of 1.45 eV was achieved due to increased intramolecular electronic interactions from D-A-D conjugated structure. BZIC was the first acceptor material based on a weak electron-deficient unit flanking with electron rich ring as central core rather than electron-donating fused ring unit like IDTT as core. By using HFQx-T as polymer donor, the OSCs exhibited a PCE of 6.30% (Feng et al., 2017b). Most of the above-mentioned acceptor molecules exhibited only one strong absorption band due to strong intramolecular charge transfer. However, M-BNBP4P-1 was developed with two strong absorption bands due to its delocalized LUMO and localized HOMO. The devices showed a PCE of 7.06% with PTB7-Th as donor (Liu et al., 2017a). IID-IC was an A-D-A'-D-A type acceptor. Because of the partially suppressed intramolecular charge transfer effects with the introduction of additional electron-deficient isoindigo unit, IID-IC exhibited a full width at half maximum of 190 nm but only 95 nm for ITIC. The OSCs based on J61: IID-IC delivered a PCE of 2.82% with broad photoresponses from 320 to 780 nm (Miao et al., 2018).

### REFERENCES

Alamoudi, M. A., Khan, J. I., Firdaus, Y., Wang, K., Andrienko, D., Beaujuge, P. M., et al. (2018). Impact of nonfullerene acceptor core structure on the photophysics and efficiency of polymer solar cells. ACS Energy Lett. 3, 802–811. doi: 10.1021/acsenergylett.8b00045

## SUMMARY AND OUTLOOK

In this review, two types of promising small-molecule electron acceptors were discussed: PDI based acceptors and A-D-A fused-ring electron acceptors. Traditional PDI units tended to form large aggregate domains leading to low exciton separation and highly torsional PDI derivatives could decrease the charge transport. Thus, a series of strategies, such as: forming PDI dimers and 3D PDI derivatives etc., were used to find the balance toward a certain aggregations that exhibited efficient exciton separation without sacrificing charge transfer and mobility. A-D-A type acceptor materials have been developed rapidly and have made exciting progress with highest PCE over 14%. In general, from fused tricyclic to fused octacyclic system, larger central core possessed redshifted absorption and high electron mobilities. Side chains were used to ensure solubility in common solvents and to inhibit strong self-assembly as well as to regulate molecular orientation and morphology. Electron-deficient end groups were used to tuning the LUMO energy level and π-π stacking.

Although non-fullerene-based solar cells have made tremendous progress in recent few years, in order to meet practical applications, designing and synthesizing new acceptor materials, pairing with donor materials, together with technical progress in device fabrications are highly desirable. When designing new active layer materials, basic properties such as absorption, energy levels and charge transport should be carefully considered. Another needed to consider is the cost and stability. We believe that a bright future for realizing high-performance and practical non-fullerene OSCs can be expected.

## AUTHOR CONTRIBUTIONS

ZZ collected the references, drew the structures, wrote the first draft of the manuscript; JY and QW helped with the revision of the manuscript and answered questions the reviewers raised; YZ supervised this project, revised the manuscript and helped all the submissions and giving the answers.

## ACKNOWLEDGMENTS

This work has been financially supported by the National Natural Science Foundation of China (21875286, 51173206), National Key Research & Development Projects of China (2017YFA0206600), Science Fund for Distinguished Young Scholars of Hunan Province (2017JJ1029) and Project of Innovation-driven Plan in Central South University, China (2016CX035).

Bin, H., Zhang, Z. G., Gao, L., Chen, S., Zhong, L., Xue, L., et al. (2016). Non-Fullerene Polymer Solar Cells Based on Alkylthio and Fluorine Substituted

Baran, D., Kirchartz, T., Wheeler, S., Dimitrov, S., Abdelsamie, M., Gorman, J., et al. (2016). Reduced voltage losses yield 10% efficient fullerene free organic solar cells with >1 V open circuit voltages. Energy Environ. Sci. 9, 3783–3793. doi: 10.1039/c6ee02598f

2D-Conjugated Polymers Reach 9.5% Efficiency. J. Am. Chem. Soc. 138, 4657–4664. doi: 10.1021/jacs.6b01744


Hindrance of Side Chains in A-D-A Small-Molecule Photovoltaic Materials. Chem. Mater. 30, 619–628. doi: 10.1021/acs.chemmater.7b03142


for high-performance organic solar cells. Adv. Mater. 29:1604964. doi: 10.1002/adma.201604964


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Zhang, Yuan, Wei and Zou. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

## Effects of Alkoxy and Fluorine Atom Substitution of Donor Molecules on the Morphology and Photovoltaic Performance of All Small Molecule Organic Solar Cells

Beibei Qiu1,2, Shanshan Chen<sup>3</sup> , Lingwei Xue<sup>1</sup> , Chenkai Sun1,2, Xiaojun Li 1,2 , Zhi-Guo Zhang<sup>1</sup> , Changduk Yang<sup>3</sup> and Yongfang Li 1,2,4 \*

<sup>1</sup> CAS Key Laboratory of Organic Solids, CAS Research/Education Center for Excellence in Molecular Sciences, Institute of Chemistry, Chinese Academy of Sciences, Beijing, China, <sup>2</sup> School of Chemical Science, University of Chinese Academy of Sciences, Beijing, China, <sup>3</sup> Department of Energy Engineering, Low Dimensional Carbon Materials Center, School of Energy and Chemical Engineering, Ulsan National Institute of Science and Technology, Ulsan, South Korea, <sup>4</sup> Laboratory of Advanced Optoelectronic Materials, College of Chemistry, Chemical Engineering and Materials Science, Soochow University, Suzhou, China

#### Edited by:

Donghong Yu, Aalborg University, Denmark

#### Reviewed by:

Qiang Peng, Sichuan University, China Florenci Vicent González, Universitat Jaume I, Spain

> \*Correspondence: Yongfang Li liyf@iccas.ac.cn

#### Specialty section:

This article was submitted to Organic Chemistry, a section of the journal Frontiers in Chemistry

Received: 20 June 2018 Accepted: 23 August 2018 Published: 13 September 2018

#### Citation:

Qiu B, Chen S, Xue L, Sun C, Li X, Zhang Z-G, Yang C and Li Y (2018) Effects of Alkoxy and Fluorine Atom Substitution of Donor Molecules on the Morphology and Photovoltaic Performance of All Small Molecule Organic Solar Cells. Front. Chem. 6:413. doi: 10.3389/fchem.2018.00413 Two benzothiadiazole (BT)-based small-molecule donors, SM-BT-2OR with alkoxy side chain and SM-BT-2F with fluorine atom substitution, were designed and synthesized for investigating the effect of the substituents on the photovoltaic performance of the donor molecules in all small molecule organic solar cells (SM-OSCs). Compared to SM-BT-2OR, the film of SM-BT-2F exhibited red-shifted absorption and deeper HOMO level of −5.36 eV. When blending with n-type organic semiconductor (n-OS) acceptor IDIC, the as-cast devices displayed similar PCE values of 2.33 and 2.76% for the SM-BT-2OR and SM-BT-2F-based devices, respectively. The SM-BT-2OR-based devices with thermal annealing (TA) at 120◦C for 10 min showed optimized PCE of 7.20%, however, the SM-BT-2F-based device displayed lower PCE after the TA treatment, which should be ascribed to the undesirable morphology and molecular orientation. Our results reveal that for the SM-OSCs, the substituent groups of small molecule donors have great impact on the film morphology, as well as the photovoltaic performance.

Keywords: benzothiadiazole, organic small molecule donors, fluorine substitution, alkoxy side chain, all small molecule organic solar cells

## INTRODUCTION

Organic photovoltaics (OPV), as one of the most promising next generation technologies to utilize solar energy, have been extensively investigated during the past several decades, due to its attractive advantages of light-weight, low-cost and capability to be fabricated into flexible and semitransparent devices (Liang and Yu, 2010; Li, 2012; Li et al., 2012). In recent years, bulk-heterojunction (BHJ) organic solar cells (OSCs) that use small-molecule n-type organic semiconductor (n-OS) as acceptor have gained significant progress (Lin and Zhan, 2014; Nielsen et al., 2015). In particular, most recently, power conversion efficiencies (PCEs) over 14% for single-layer device and 15% for tandem device have been achieved, demonstrating great potential for the commercialization of OSCs (Che et al., 2018; Li et al., 2018; Zhang et al., 2018).

Qiu et al. All Small Molecule OSCs

The rapid development of OSCs is a combination of the innovation of donor and acceptor photovoltaic materials, the methodology of morphology tuning and the optimization of device structures (Lee et al., 2008; Hau et al., 2010; Huang et al., 2014; Wang and Kyaw, 2014; Ye et al., 2014; Gao et al., 2015; Lin and Zhan, 2016; Zhang et al., 2016; Yan et al., 2017). Especially, the progress of photovoltaic materials plays a critical role in promoting the development of OSCs (Nielsen et al., 2015; Zhang and Zhu, 2017; Wadsworth et al., 2018). Especially, great improvement of the OSCs has been achieved by the development of wide bandgap conjugated polymer donors and narrow bandgap n-OS small-molecule acceptors (Bin et al., 2016a; Sun et al., 2018). Although small-molecule p-type organic semiconductor (p-OS) donors, compared with polymer donors, possess the advantages of well-defined chemical structures and easy purification, the photovoltaic performance of the all small molecules OSCs (SM-OSCs) is relatively lag behind, because of its more difficult morphology tuning than polymer donor/n-OS acceptor system (Yang et al., 2017; Shi et al., 2018). In order to further improve the performance of SM-OSCs, it is crucial to deeply investigate the relationship between the molecular structures and device performance of the SM-OSCs.

In the donor-acceptor (D-A) structured p-type semiconductor donor materials, benzothiadiazole (BT) unit is a widely used acceptor (A) building block, due to its superior advantages of planar structure and low-lying energy level (Lin and Zhan, 2016). Yan et al. adjusted alkyl chain lengths of BT-based polymer donor (PffBT4T-C9C13) and the PffBT4T-C9C13-based OSCs fabricated with a hydrocarbon solvents demonstrate a superior performance (Zhao et al., 2016). Then they designed another BT-based polymer donor P3TEA which demonstrated an efficient OSC with a negligible driving force when blending with a non-fullerene acceptor, SF-PDI<sup>2</sup> (Liu et al., 2016). Beside polymer donors, BT unit has been widely used in constructing small molecule semiconductor p-OS donors. Bazan et al. have systematically investigated BT-based donor-acceptordonor-acceptor-donor (D-A-D-A-D) type small molecule donor materials (Coughlin et al., 2014). In addition, BT-unit also has been used to construct n-type organic semiconductor materials and desirable results have been obtained (Holliday et al., 2016; Baran et al., 2017).

Considering the effective role of side chain engineering in morphology tuning, in order to study the relationship between device performance of the SM-OSCs and molecular structures, herein two BT-based small-molecule donors, SM-BT-2OR and SM-BT-2F (see **Figure 1**), were designed and synthesized for comparison studies. By rationally introducing alkoxy substituent and fluorine atom (F) on the BT unit, the two small molecules exhibits different absorption characteristics, frontier molecular orbital energy levels, and energy bandgaps. When blending with n-type organic semiconductor (n-OS) acceptor IDIC, the as cast devices for both small molecules donors showed similar PCE values. The SM-BT-2OR-based devices after thermal annealing (TA) at 120◦C for 10 min showed optimized PCE of 7.20% but lower open circuit voltage (Voc), while the SM-BT-2F-based device displayed lower PCE but slightly higher Voc after the TA treatment. Such opposite trends in the PCE and Voc should be ascribed to the different effect of the thermal annealing on the morphology and crystallinity of the two small molecules with the different substituent groups. Our results reveal that for the SM-OSCs the substituent groups have great impact on the film morphology and the photovoltaic performance.

## RESULTS AND DISCUSSION

#### Synthesis and Thermal Properties

**Scheme 1** shows the synthetic routes of SM-BT-2OR and SM-BT-2F. Compounds **M1** and **M2** were synthesized according to the previously reported methods in high yields (Gu et al., 2012; Feng et al., 2013). Then, the two small molecules SM-BT-2OR and SM-BT-2F were synthesized through similar procedures by a Stille-coupling reaction between compound **M3** and compound **M1** or **M2** with toluene as the solvent and Pd(PPh3)<sup>4</sup> as catalyst. Then, the two molecules were obtained through column chromatography. The two molecules show onset temperatures with 5% weight-loss of 331 and 377◦C for SM-BT-2OR and SM-BT-2F, respectively, in the thermogravimetric analysis as shown in **Figure S1**, indicating that both molecules possess good thermal stability for application in OSCs.

## Photophysical Properties and Electronic Energy Levels

**Figure 2A** shows the UV-vis absorption spectra of the two molecules in dilute CHCl<sup>3</sup> solutions and as thin films, and IDIC film. The specific absorption characteristics of the two molecules are summarized in **Table 1**. The two molecules in solution exhibit similar absorption profiles with the same maximum absorption wavelength at 542 nm besides the slightly higher absorption at short wavelength for SM-BT-2F solution. Whereas, in solid films, the maximum absorption wavelength of SM-BT-2F (606 nm) is red-shifted by ∼22 nm compared to that of SM-BT-2OR (584 nm), indicating that the introduction of fluorine (F) substituents could influence the stacking behavior of the molecules (Umeyama et al., 2013; Liu et al., 2014; Do et al., 2016). Because of the stronger intermolecular interactions in film state, the maximum absorption peaks of the two molecules show obvious bathochromic shifts of 42 and 64 nm for SM-BT-2OR and SM-BT-2F, respectively. The optical bandgaps (E opt <sup>g</sup> ) estimated from the UV-vis absorption onsets (701 nm for SM-BT-2OR and 745 nm for SM-BT-2F) in the film state are determined to be 1.77 and 1.66 eV for SM-BT-2OR and SM-BT-2F, respectively.

The highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) levels of the two molecules were measured by electrochemical cyclic voltammetry with Ag/AgCl as reference electrode, as shown in **Figure 2B.** The HOMO/LUMO energy levels (EHOMO/ELUMO) were calculated from the onset oxidation / reduction potentials (φox/φred) according to the equations of EHOMO/ELUMO = –e(φox/φred + 4.8 – φFc/Fc+) (eV) (Bin et al., 2016a). φFc/Fc<sup>+</sup> was measured to be 0.44 V vs. Ag/AgCl in this measurement system, and then the calculation equations are EHOMO/ELUMO = –e(φox/φred + 4.36) (eV). As shown in **Figure 2B**, the onset oxidation


TABLE 2 | Photovoltaic performance parameters of the optimized SM-OSCs based on SM-BT-2OR: IDIC and SM-BT-2F:IDIC.

corresponding devices. (C) Jph vs. Veff of the devices. (D) Light intensity dependence of Jsc of the devices.


<sup>a</sup>Average values with standard deviations were obtained from 10 devices.

<sup>b</sup>The slope of the dependence of Jsc on light intensity (P) on logarithmic coordinates. <sup>c</sup>The dependence of Voc on light intensity on semilogarithmic coordinate.

potentials (φox) / onset reduction potentials (φred) for SM-BT-2OR and SM-BT-2F are 0.98/−1.24 V and 1.00/−1.10 V vs. Ag/AgCl, respectively. The ELUMO/EHOMO of SM-BT-2OR and SM-BT-2F were calculated to be −3.12/−5.34 and −3.26/−5.36 eV, respectively. In comparison with SM-BT-2OR, the EHOMO of SM-BT-2F with F atoms substitution is slightly down-shifted by 0.02 eV, which is beneficial for higher Voc of the SM-OSCs. Compared with the energy level of IDIC, the LUMO energy level offsets (1ELUMO) and HOMO energy level offsets (1EHOMO) of SM-BT-2OR/IDIC and SM-BT-2F/IDIC are 0.71/0.35 and 0.57/0.33 eV, respectively, which is sufficient for charge separation (Hendriks et al., 2016).

## Photovoltaic Performance

SM-OSCs were fabricated with a conventional device structure of ITO/PEDOT:PSS/p-OS:IDIC/PDINO/Al, and characterized to investigate the photovoltaic properties of the p-OS SM-BT-2OR and SM-BT-2F (Zhang et al., 2014). **Figure 3A** shows the typical current density-voltage (J-V) curves of the best devices based on SM-BT-2OR and SM-BT-2F under the illumination of AM 1.5G, 100 mW cm−<sup>2</sup> , and the corresponding device performances, including Voc, Jsc, FF and PCE, are summarized in **Table 2**. As shown in **Figure 3A**, for the as-cast devices, the device based on SM-BT-2OR: IDIC displayed an inferior PCE of 2.33%, with a high Voc of 0.962 V, but a low Jsc and low FF. The SM-BT-2F-based device showed slightly higher PCE of 2.76%, with a higher Voc of 0.983 V, which should be ascribed to the deeper HOMO level of SM-BT-2F. Thermal annealing (TA)


TABLE 3 | Hole and electron mobilities of the SM-BT-2OR: IDIC and SM-BT-2F: IDIC blend films with or without thermal annealing treatment at 120◦C for 10 min.

treatment has been proved to be an effective method to tune the morphology and phase separation of non-fullerene OSCs (Bin et al., 2016b; Chen et al., 2017). **Table S1** displays the photovoltaic parameters of the OSCs based on SM-BT-2OR: IDIC (w/w, 2:1) and SM-BT-2F: IDIC (w/w, 2:1) as cast or with TA treatment at different temperatures. Interestingly, for the devices based on SM-BT-2OR: IDIC, TA treatment leads to higher photovoltaic performance, however, for SM-BT-2F: IDIC, TA treatment leads to poorer efficiency. As shown in **Table 2**, when treated at 120◦C for 10 min, the device based on SM-BT-2OR: IDIC demonstrated preferable PCE of 7.20%, with a slightly lower Voc of 0.939 V, an enhanced Jsc of 13.57 mA cm−<sup>2</sup> and a higher FF of 56.5%. However, the device based on SM-BT-2F: IDIC showed worse PCE of 1.82%, with a slightly higher Voc of 0.987 V, but a lower Jsc of 5.36 mA cm−<sup>2</sup> and a poorer FF of 30.2%. The reverse trends of the TA treatment on the PCE and Voc of the devices could be related to the different substituents of the donor molecules.

The input photon to converted current efficiency (IPCE) spectra of the best devices are shown in **Figure 3B**. Both of the two small molecule-based devices demonstrate broad photoresponse from 300 to 800 nm, which indicates that both the small molecule donors and the IDIC acceptor make contributions to the photo current. Compared to the device based on SM-BT-2F, the SM-BT-2OR-based devices present higher photoresponse. Especially, the SM-BT-2OR-based device with the TA treatment present a broad plateau with IPCE values of around 50–60% in the wavelength range of ca. 420–740 nm, resulting in the relatively higher Jsc of 13.57 mA cm−<sup>2</sup> . The Jsc values of the OSCs based on SM-BT-2OR and SM-BT-2F calculated from integration of the EQE spectra with the AM 1.5G reference spectrum are 13.31 and 7.17 mA cm−<sup>2</sup> , respectively, which are in good agreement with Jsc values measured from J-V curves.

In order to investigate the charge dissociation of the SM-OSCs, the relationship between photocurrent density (Jph, Jph = J<sup>L</sup> – JD, where J<sup>L</sup> and J<sup>D</sup> represent the current densities under the illumination and in the dark, respectively) and effective applied voltage (Veff ) was analyzed (Mihailetchi et al., 2004; Wu et al., 2011). As can be seen from **Figure 3C**, when Veff arrives at ∼3 V, Jph value for the SM-BT-2OR-based device with the TA treatment reached saturation (Jsat). While for the SM-BT-2F-based device (with and without the TA treatment) and the SM-BT-2OR-based device without the TA treatment (**Figure S2**), the Jph values do not show a saturation trend within the whole measurement region, suggesting serious recombination process. The charge dissociation and charge collection probability [P(E, T)] in the devices could be estimated by calculating the value of Jph/Jsat. Under the short circuit and maximal power output conditions, for the SM-BT-2OR-based devices, the P(E, T) values are 63% and 36% for the as-cast device, and 86 and 63% for the TA treated devices, suggesting a higher charge collection efficiency after TA treatment. However, the corresponding P(E, T) values of the SM-BT-2F-based devices are 68 and 44% for the as-cast device, and 53 and 30% for the TA treated devices, respectively, which means that the SM-BT-2F-based devices possess serious recombination whether with or without TA treatment, which could be ascribed to poorer morphology of the blend films of SM-BT-2F:IDIC.

To further study the charge recombination behavior in the SM-OSCs system, the dependence of Jsc on light intensity (P) was evaluated (**Figure 3D**; **Figure S3**). Generally, Jsc is known to follow a power-law dependence with respect to Plight, which can be described as Jph ∞ P α (Schilinsky et al., 2002; Koster et al., 2005). For the SM-BT-2OR-based device, the exponential factor (α) is 0.926 for the as-cast device (**Figure S3A**) and 0.953 for the device with the TA treatment (**Figure 3D**), indicating that bimolecular recombination could be effectively suppressed by the TA treatment. However, for the SM-BT-2F-based device with the TA treatment (**Figure S3B**), the exponential factor (α) is only 0.914, which is slightly lower than that without the TA treatment (α = 0.924, **Figure 3D**), suggesting the existence of a certain amount of bimolecular recombination whether with or without the TA treatment. Compared to the SM-BT-2F-based device, the SM-BT-2OR-based device with the TA treatment shows higher α value, indicating better charge transport capacity, which agrees well with its better device performance. We also measured the dependence of Voc on the light intensity to investigate the recombination mechanisms. As shown in **Figure S4**, for the SM-BT-2OR-based devices, the slopes are 1.20 and 1.11 kT/q for the devices as cast and with the TA treatment respectively, which means the recombination at open circuit is a bimolecular dominated process. While for the SM-BT-2F-based devices with or without the TA treatment, the slopes could be separated into two regions, as shown in **Figure S5**. Even under higher light intensity, the slopes are 1.41 kT/q and 1.78 kT/q for the devices as cast and with TA treatment, respectively, indicating a combination of monomolecular and bimolecular recombination processes. Besides, the larger value of 1.78 kT/q suggests a more undesirable morphology could be formed when the TA treatment was applied for the SM-BT-2Fbased device.

The space-charge limited current models, using holeonly and electron-only devices with architectures of ITO/PEDOT:PSS/active layer/Au and ITO/ZnO/active

FIGURE 4 | AFM height images (size: 5 × 5 µm<sup>2</sup> ) of SM-BT-2OR: IDIC blends as cast (a) and TA treated (b), SM-BT-2F: IDIC blends as cast (c) and TA treated (d). Root-mean-square (RMS) roughness values are given to describe the smoothness of the morphology. TEM images of SM-BT-2OR: IDIC blends as cast (e) and TA treated (f), SM-BT-2F: IDIC blends as cast (g) and TA treated (h).

layer/PDINO/Al, respectively, were applied to measure the hole and electron mobilities of the blends before or after the TA treatment (**Figures S6**, **S7**). As shown in **Table 3**, for both blend films, after the thermal annealing treatment, the charge mobilities (electron and hole mobility) were higher than those of the as cast devices. The hole mobilities of the SM-BT-2ORbased devices were measured to be 2.79×10−<sup>5</sup> and 7.37×10−<sup>5</sup>

cm<sup>2</sup> V −1 s −1 for the devices as cast and with TA treatment, respectively. Compared with the SM-BT-2OR-based devices, the hole mobility of the SM-BT-2F-based blends were rather low whether before (0.37×10−<sup>5</sup> cm<sup>2</sup> V −1 s −1 ) or after the TA treatment (1.77×10−<sup>5</sup> cm<sup>2</sup> V −1 s −1 ), leading to inferior FF and photovoltaic performance. For the SM-BT-2OR-based devices, after the TA treatment, the ratio of hole mobility/electron mobility was slightly balanced, which might benefit for obtaining higher FF of the SM-BT-2OR-based devices. Though the device performance of the SM-BT-2OR-based device with the TA treatment is fairly good, compared to previous results of the SM-OSCs (Qiu et al., 2017; Yang et al., 2018), the hole mobilites of the SM-BT-2OR-based blend is slightly lower, which limits its further improvement of photovoltaic performance.

Considering that the device performance is closely related to the blend morphology, detailed investigations on the morphological characteristics of the blend films were carried out (Li et al., 2005; Rivnay et al., 2012). Atomic force microscopy (AFM) was utilized to probe the surface morphologies of the blend films with or without the TA treatment. As shown in **Figure 4**, all the blends display homogeneous surfaces with a moderate root-mean-square (RMS) roughness. The SM-BT-2F: IDIC blend without the TA treatment presents higher RMS value than that of the SM-BT-2OR: IDIC blend, which should be ascribed to the better crystallinity of SM-BT-2F. Compared to the blend without the TA treatment, the blends with the TA treatment show larger RMS, indicating that the TA treatment could effectively tune the morphology of the active layers of the SM-OSCs. In order to clearly understand aggregation situation of both small molecules in the blend films, transmission electron microscope (TEM) was applied to investigate the phase separation. As shown in **Figure 4**, the black and white regions in the TEM images of the ascast blend of SM-BT-2OR: IDIC is uniformly distributed, indicating that SM-BT-2OR and IDIC can be well-mixed. For the as cast blend of SM-BT-2F: IDIC, the TEM image shows relatively large aggregation, which explains the rather low short circuit current density and photovoltaic performance. After thermal annealing, the phase separation of the SM-BT-2F: IDIC blend become even larger, resulting to the more severe recombination, thus leading to the poorer Jsc and device performance. While for the SM-BT-2OR: IDIC blend with the TA treatment, the TEM image shows enhanced phase separation with small fiber like aggregation, which is beneficial for effective exciton dissociation and charge transport, therefore, better device performance could be achieved. In addition, it should be noticed that, although the AFM and TEM images of the SM-BT-2F-based films seem rather uniform, the image of optical microscope presented obvious striped aggregation, as shown in **Figure S8**, indicating the inferior blend morphology on micrometer scale.

To gain deep insight into the molecular stacking differences between SM-BT-2OR and SM-BT-2F, grazing incidence wideangle X-ray scattering (GIWAXS) measurements were employed (Lilliu et al., 2011; Rivnay et al., 2012; Huang et al., 2014). The 2D GIWAXS patterns and corresponding cut-line profiles in the in-plane and out-of-plane directions are shown in **Figure 5** (neat films) and **Figure 6** (blend films). The neat SM-BT-2OR films showed amorphous feature with π-π stacking peak at 1.74 Å−<sup>1</sup> (d-spacing: 3.61 Å) and coherence length of 25.4 Å. The neat SM-BT-2F film displayed edge-on orientations with π-π stacking peak at 1.76 Å−<sup>1</sup> (d-spacing: 3.56 Å) and coherence length of 36.0 Å. The smaller π-π stacking peak and longer coherence length of neat SM-BT-2F film should be ascribed to the better molecular planarity and crystallinity caused by the introduction of F atoms on the BT unit, leading to rather large phase separation as discussed in TEM section.

When blending with IDIC, both films exhibits random orientation and disordered microstructure features with relatively weak peak intensities. After the thermal annealing treatment, significantly stronger peaks with more narrow width accompany by the appearance of new peaks. For both blends of SM-BT-2OR: IDIC and SM-BT-2F: IDIC, the 2D GIWAXS patterns exhibit strong lamellar (100) and (200) diffraction peaks in the out-of-plane direction, indicating a high degree of molecular ordering. Besides, the (010) diffraction peaks demonstrated in-plane preferred orientation, with enhanced coherence length of 47.0 and 52.1 Å, for the SM-BT-2OR: IDIC and SM-BT-2F: IDIC blends, respectively, resulting in better charge transport. Compared to the TA treated blend films of SM-BT-2OR, it can be observed that the SM-BT-2Fbased blend film are more prone to adopt a predominant edge-on crystalline orientation, which suggests that it possesses less three-dimensional (3-D) charge pathways in the active layer (Bin et al., 2017; Kumari et al., 2017). Thus, although the molecular packing and phase separation was enhanced, carrier collection was hampered, resulting in lower FF, Jsc and photovoltaic performance. These results indicate that for both small molecules, TA treatment could effectively enhance the molecular packing and affect the molecular orientation in the blend film. In addition, we want to mention that the GIWAXS results show weak face-on stacking, which is also observed for other SM-OSCs based on small molecule donor and small molecule acceptor (Bin et al., 2017; Yang et al., 2017). In order to further improve photovoltaic performance of the SM-OSCs, we should do more work on the morphology optimization of the blend active layer of the p-type small molecule donor and n-type small molecule acceptor.

It should also be mentioned that differing from most other fluorine-containing donor materials SM-BT-2F based device

displayed inferior FF and photovoltaic performance, which could be ascribed to the too large phase separation and unsatisfactory molecular orientation of SM-BT-2F.

### CONCLUSIONS

In summary, two benzothiadiazole based small-molecule donors, SM-BT-2OR and SM-BT-2F, were designed and synthesized for investigating the effect of the substituents on the photovoltaic performance of the molecules. Compared to SM-BT-2OR, because of the substitution of fluorine atom (F), SM-BT-2F presented red-shifted absorption profile in film state and deeper HOMO level of 5.36 eV. When blending with n-type organic semiconductor (n-OS) acceptor IDIC, the as-cast devices displayed similar PCE values of 2.33 and 2.76% for the SM-BT-2OR and SM-BT-2F-based devices, respectively. When TA treatment at 120◦C for 10 min was applied, the SM-BT-2ORbased devices showed better performance of 7.20%, while the SM-BT-2F-based device displayed even lower PCE. The lower PCE of the SM-BT-2F-based device should be ascribed to the rather large phase separation and more in-plane preferred orientation of ππ stacking when the TA treatment was used, which decreased the exciton dissociation and charge transportation. Besides, the reduced Voc of the SM-BT-2OR-based devices with the TA treatment should be due to the enhanced phase separation. Our results reveal that for the SM-OSCs the substituent groups have

#### REFERENCES


great impact on the film morphology, as well as the photovoltaic performance.

#### AUTHOR CONTRIBUTIONS

BQ and YL designed the two small molecules. BQ carried out the materials synthesis and device fabrication and photovoltaic performance studies. Z-GZ, CS, and XL participated in the discussion of the material synthesis. LX and Z-GZ provided the cathode buffer layer material. SC and CY measured the GIWAXS diffraction patterns. YL supervised the project. BQ and YL write the manuscript.

#### ACKNOWLEDGMENTS

The work was supported by the Ministry of Science and Technology of China (973 project, No. 2014CB643501) and NSFC (Nos. 91633301, 21734008, and 51673200) and the Strategic Priority Research Program of the Chinese Academy of Sciences, Grant No. XDB12030200.

#### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem. 2018.00413/full#supplementary-material


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Qiu, Chen, Xue, Sun, Li, Zhang, Yang and Li. This is an openaccess article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

## Comparison of the Solution and Vacuum-Processed Squaraine:Fullerene Small-Molecule Bulk Heterojunction Solar Cells

Guo Chen<sup>1</sup> \*, Zhitian Ling<sup>1</sup> , Bin Wei <sup>1</sup> , Jianhua Zhang<sup>1</sup> , Ziruo Hong<sup>2</sup> , Hisahiro Sasabe<sup>2</sup> \* and Junji Kido<sup>2</sup>

*<sup>1</sup> Key Laboratory of Advanced Display and System Applications, Ministry of Education, Shanghai University, Shanghai, China, <sup>2</sup> Department of Organic Device Engineering, Graduate School of Engineering, Research Center for Organic Electronics, Yamagata University, Yonezawa, Japan*

#### Edited by:

*Chuanlang Zhan, Institute of Chemistry (CAS), China*

#### Reviewed by:

*Kunpeng Guo, Taiyuan University of Technology, China Chunhui Duan, South China University of Technology, China*

#### \*Correspondence:

*Guo Chen chenguo@shu.edu.cn Hisahiro Sasabe h-sasabe@yz.yamagata-u.ac.jp*

#### Specialty section:

*This article was submitted to Organic Chemistry, a section of the journal Frontiers in Chemistry*

Received: *17 June 2018* Accepted: *22 August 2018* Published: *11 September 2018*

#### Citation:

*Chen G, Ling Z, Wei B, Zhang J, Hong Z, Sasabe H and Kido J (2018) Comparison of the Solution and Vacuum-Processed Squaraine:Fullerene Small-Molecule Bulk Heterojunction Solar Cells. Front. Chem. 6:412. doi: 10.3389/fchem.2018.00412* Squaraine dyes have shown promising properties for high performance organic solar cells owing to their advantages of intense absorption and high absorption coefficients in the visible and near-infrared (NIR) regions. In this work, to directly compare the photovoltaic performance of solution- and vacuum-processed small-molecule bulk heterojunction (SMBHJ) solar cells, we employed a squaraine small molecular dye, 2,4-bis[4-(N,N-diisobutylamino)-2,6-dihydroxyphenyl] squaraine (DIBSQ), as an electron donor combined with fullerene acceptors to fabricate SMBHJ cells either from solution or vacuum deposition process. The solution-processed SMBHJ cell possesses a power conversion efficiency (PCE) of ∼4.3%, while the vacuum-processed cell provides a PCE of ∼6.3%. Comparison of the device performance shows that the vacuumprocessed SMBHJ cells possess higher short-circuit current density, fill factor and thus higher PCE than the solution-processed devices, which should be assigned to more efficient charge transport and charge extraction in the vacuum-processed SMBHJ cells. However, solution-processed SMBHJ cells demonstrate more pronounced temperaturedependent device performance and higher device stability. This study indicates the great potential of DIBSQ in photovoltaic application via both of solution and vacuum processing techniques.

Keywords: organic solar cells, bulk heterojunction, squaraine dye, solution-process, vacuum-process

### INTRODUCTION

Organic solar cells (OSCs) have attracted much attention as a green solar energy technology for cost-effective, renewable energy sources because of their advantages of low-cost, light-weight, large-area manufacturing, and mechanical flexibility (Yu et al., 1995; Günes et al., 2007; Li et al., 2012; He et al., 2015; Duan et al., 2017). An efficient OSC must absorb over a broad spectral range from visible to near-infrared (NIR) wavelengths (350∼950 nm) and convert the incident light effectively into charges (Yu et al., 1995). Bulk heterojunction (BHJ) cell with broad spectral absorbing photoactive layer has great potential to realize high power conversion efficiency (PCE) because the donor/acceptor system in this kind of device can efficiently overcome the strong exciton binding energy and thus achieve efficient charge generation (Lin et al., 2012; Chen et al., 2017b; Tang et al., 2017; Duan et al., 2018; Qi et al., 2018). Generally, there are two processing techniques to fabricate BHJ cells, i.e., solution processing and vacuum evaporation. These two deposition routes require the raw materials with much different properties, in particular solubility and thermal stability, respectively (Kronenberg et al., 2010; Hu et al., 2017). In general, the conjugated polymers with high solubility are deposited by solution process while small molecules (SMs) with high-thermal stability are deposited by vacuum evaporation. Solution processing is a relatively low-cost and fast method for depositing thin films. However, the stacking of multilayer layers becomes a great challenge because the interface erosion issue between the different solution-processed stacked layers (Hu et al., 2017). In comparison, vacuum deposition has several exceptive advantages such as accurate control of the thin film thickness and evaporation rate of the materials, and easy fabrication of multilayer devices by successive deposition of the materials. However, the vacuum conditions lead to high production costs (Kronenberg et al., 2010).

Recently, much effort has been dedicated to develop NIR absorbing polymer or small molecular donors combined with the fullerene acceptors to construct broad spectral absorbing BHJ layers (Li, 2012; Roncali et al., 2014; Chen et al., 2016b; Huang et al., 2016; Jiang et al., 2016; Du et al., 2018). Squaraine (SQ) dye is a kind of NIR absorbing small molecular donor material due to its advantages of simple synthetic route, high photochemical and thermal stability and high absorption coefficient in the NIR region (Ajayaghosh, 2005; Sasabe et al., 2014; Chen et al., 2015, 2017b). More recently, some NIR absorbing SQ donor materials with symmetric or asymmetrical molecular structures have been developed for high-performance BHJ cells by several groups: Silvestri et al. firstly introduced the hydrazine endcapped symmetric SQ donor into the BHJ system combined with [6,6]-phenyl-C61-butyric acid methyl ester (PCBM) acceptor and obtained a PCE of over 1% (Silvestri et al., 2008), and 2% of PCE was then obtained from the alkenyl-functionalized symmetric SQ and [6,6]-phenyl-C71-butyric acid methyl ester (PC71BM) based BHJ cell (Bagnis et al., 2010). Würthner's group reported several symmetric SQs based BHJ cells and achieved a PCE of 1.8% with unusually high short circuit current (Jsc) of 12.6 mA/cm<sup>2</sup> (Mayerhöffer et al., 2009). Wei et al. reported PCEs of over 5% from symmetric SQ and PC71BM based BHJ cells (Wei et al., 2011). Chen et al. further promoted the PCE of symmetric SQ based BHJ up to over 6% by using co-evaporation technology (Chen et al., 2012a). Huang's group has synthesized a series of NIR absorbing asymmetric SQ donors combined with PC71BM acceptor to realize PCEs of over 6% (Wu et al., 2018).

Among all the above mentioned SQ dyes, the symmetric SQ dye 2,4-bis[4-(N,N-diisobutylamino)-2,6-dihydroxyphenyl] squaraine (DIBSQ) has shown promising properties for highly efficient BHJ cells (Chen et al., 2012a): DIBSQ has a wide absorption spectra in the NIR region with high absorption coefficient of over 3 × 10<sup>5</sup> cm−<sup>1</sup> , and its highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) levels are located at −5.3 and −3.5 eV, respectively (Chen et al., 2016a), which provides a higher photocurrent and higher Voc for the BHJ cells. The most unique property of DIBSQ is that the DIBSQ thin film can be deposited by using both of vacuum evaporation and solution processing technologies because the DIBSQ has high thermal

stability and moderate solubility in organic solvents. To the best of our knowledge, it is rare that a small molecular thin film can be fabricated by using both of solution process and vacuum evaporation techniques (Chen et al., 2013). This unique property of DIBSQ indicates the photoactive layers of the DIBSQ:Fullerene based SMBHJ cells can be fabricated by spin-coating DIBSQ:PC71BM mixed solution or by vacuum co-evaporation of DIBSQ and C70, which offers the possibility to directly compare the device performance of two processing techniques.

In this work, to directly compare the two processing techniques for BHJs, i.e., solution process and vacuum evaporation process, we fabricated a DIBSQ:PC71BM SMBHJ cell by using spin-coating processing and a DIBSQ:C<sup>70</sup> SMBHJ cell by using vacuum co-evaporation processing. Then the device performance and film properties of the SMBHJ active layers were characterized and systematically compared.

## EXPERIMENTAL SECTION

#### Materials

DIBSQ was synthesized according to the procedure reported by Tian et al. (2002), The purity of DIBSQ was proved to be over 99.9% by NMR and elemental analysis. MoO3, bathocuproine (BCP) and PC71BM were commercially available and used as received. Commercially available C<sup>70</sup> was sublimated 3 times before use. Patterned indium-tine-oxide (ITO) glass substrates were successively cleaned by using detergent, deionized water, acetone and isopropanol in ultrasonic bath, respectively, finally the cleaned ITO substrates were kept in an oven at 80◦C for 12 h to be completely dried.

## Film Characterization

UV-vis absorption spectra were carried out using a UV-vis-NIR spectrophotometer (SHIMADZU, UV-3150). Atomic force microscopy (AFM) images were performed in air on a scanning probe microscope (Nanonavi SPA-400SPM, Japan) using a tapping mode. X-ray diffraction (XRD) patterns were measured using a high-resolution XRD diffractometer (SmartLab, Rigaku Co.). The films for these measurements were prepared by using the same fabrication conditions for the devices to enable accurate comparisons.

### Device Fabrication and Characterization

The DIBSQ:PC71BM BHJ cells with the structure of ITO/MoO<sup>3</sup> (5 nm)/DIBSQ:PC71BM (60 nm, 1:5)/BCP (10 nm)/Al (100 nm) (as shown in **Figure 1**) were fabricated as follows: the cleaned ITO substrates were exposed to UV ozone for 30 min and immediately transferred into a high-vacuum chamber for deposition of 5 nm-thick MoO<sup>3</sup> at a base pressure of 1 × 10−<sup>5</sup> Pa. The substrates were then transferred into a nitrogen-filled glove box, 60 nm-thick photoactive layers were fabricated by spincoating DIBSQ:PC71BM solution (20 mg mL−<sup>1</sup> in chloroform with donor:acceptor weight ratio of 1:5) on the MoO<sup>3</sup> coated ITO surface at a rate of 4,000 revolutions per minute (rpm), then the DIBSQ:PC71BM blend films were thermally annealed at 70◦C for 10 min. Finally, the substrates were transferred back to the high-vacuum chamber where 10 nm-thick BCP and 100 nm-thick Al cathode were successively deposited. The DIBSQ:C<sup>70</sup> BHJ cells with the structure of ITO/MoO<sup>3</sup> (5 nm)/DIBSQ:C<sup>70</sup> (60 nm, 1:5)/BCP (10 nm)/Al (100 nm) were fabricated in a high vacuum chamber with a pressure of 5 × 10−<sup>6</sup> Pa, where 5 nm MoO<sup>3</sup> was deposited on the pre-cleaned ITO surface, then the 60 nm-thick DIBSQ:C<sup>70</sup> photoactive layer were prepared by co-depositing DIBSQ and C<sup>70</sup> with a blending ratio of 1:5, finally the devices were completed by evaporating 10 nm-thick BCP and 100 nm-thick Al cathode. The active area of solar cells is 0.04 cm<sup>2</sup> . Hole-only devices with the structure of ITO/MoO<sup>3</sup> (5 nm)/DIBSQ:PC71BM (60 nm, 1:5) or DIBSQ:C<sup>70</sup> (60 nm, 1:5)/MoO<sup>3</sup> (5 nm)/Al (100 nm) and electron-only devices with the structure of ITO/Cs2CO<sup>3</sup> (1 nm)/DIBSQ:PC71BM (60 nm, 1:5) or DIBSQ:C<sup>70</sup> (60 nm, 1:5)/BCP (10 nm)/Al (100 nm) were fabricated to characterize hole and electron mobility in photoactive blend films by using space-charge-limited current (SCLC) method (Shrotriya et al., 2006), respectively. The current density-voltage (J-V) and external quantum efficiency (EQE) of BHJ cells were measured on a CEP-2000 integrated system (Bunkou Keiki Co.) under standard measurement conditions. The device performance data were averaged from 16 individually fabricated BHJ cells.

#### RESULTS AND DISCUSSION

**Figure 1** represents the architectures of the DIBSQ:Fullerene SMBHJ cells (**Figure 1a**), the corresponding energy-level diagram (**Figure 1b**) and the molecular structures of the organic materials used in the device (**Figure 1c**). In the SMBHJ device, MoO<sup>3</sup> is adopted as hole-transporting layer. As demonstrated in our previous research (Chen et al., 2012b), the MoO<sup>3</sup> layer effectively deepened the work function of ITO to −5.7 eV, and thus delivered a larger open circuit voltage (Voc) for the DIBSQ:Fullerene SMBHJ cells compared with those based on the typical ITO/PEDOT:PSS anode. The DIBSQ:Fullerene blend film is employed as the photoactive layer, while the bathocuproine (BCP) is used as electron-transporting layer. For comparison, the same ITO/MoO<sup>3</sup> (5 nm) cathode and BCP (10 nm)/Al (100 nm) anode were employed in all devices, as well as the film thickness and the blend ratio of DIBSQ and Fullerene for all the devices were also kept same as 60 nm and 1:5 ratio, respectively. The only difference is the fabrication processing and the corresponding acceptor materials used for the photoactive layers: the solution-processed DIBSQ:PC71BM SMBHJ cells were prepared by spin-coating the DIBSQ:PC71BM blend solution, while the vacuum-processed DIBSQ:C<sup>70</sup> SMBHJ cells were fabricated by co-evaporating the DIBSQ and C<sup>70</sup> in the vacuum condition.

The UV-Vis absorption spectra of the neat films of DIBSQ, C70, PC71BM and the blend films of DIBSQ:PC71BM and DIBSQ:C<sup>70</sup> were characterized. As shown in **Figure 2a**, the DIBSQ film displays an absorption band in the region of 500– 800 nm with an absorption peak at 700 nm; the C<sup>70</sup> and PC71BM film show wide absorption bands between 300 and 740 nm, thus the blend films of both DIBSQ:PC71BM and DIBSQ:C<sup>70</sup> display strong and broad absorption covering the region from visible to NIR, which leads to larger light-harvesting, and thus potentially obtains higher Jsc of the DIBSQ:Fullerene SMBHJ cells (Chen et al., 2014).

**Figure 3a** exhibits the J-V curves of the SMBHJ cells under 100 mW/cm<sup>2</sup> AM 1.5 illumination. As shown in **Figure 3a**

TABLE 1 | Comparison of device performance of solution-processed DIBSQ:PC71BM and vacuum-processed DIBSQ:C70 SMBHJ cells.


*<sup>a</sup>Carrier mobility data were originated from a space-charge-limited current (SCLC) model. The device performance data were averaged from 16 individually fabricated BHJ cells.*

and **Table 1**, the solution-processed DIBSQ:PC71BM SMBHJ cell possesses a PCE of 4.26% with a Jsc of 10.60 mA/cm<sup>2</sup> , a Voc of 0.94 V and a fill factor (FF) of 0.43. While the vacuum-processed DIBSQ:C<sup>70</sup> SMBHJ cell provides a PCE of 6.32% with a Jsc of 13.69 mA/cm<sup>2</sup> , Voc of 0.87 V and FF of 0.53. Comparison of the device performance shows that the vacuum-processed BHJ cells possess higher Jsc and FF and thus higher PCE than the solutionprocessed devices, which should be assigned to more efficient charge transport and charge extraction in the vacuum-processed SMBHJ cells owing to much higher and more balanced hole and electron mobilities in their active layers (**Table 1**) (Chen et al., 2017b), as discussed in a subsequent section. The lower Voc in the vacuum-processed DIBSQ:C<sup>70</sup> SMBHJ cells can be explained by the deeper LUMO level of C<sup>70</sup> acceptor compared with that of PC71BM acceptor, as shown in **Figure 1b**. Normally, the Voc of BHJs depends on the energy level difference between the HOMO of the donor (HOMO Donor) and LUMO of the acceptor (LUMO Acceptor) (Chen et al., 2017b), the deeper LUMO level of C<sup>70</sup> leads to smaller |HOMO Donor| – |LUMO Acceptor| value and thus smaller Voc. As shown in **Table 1**, comparing with the solutionprocessed DIBSQ:PC71BM SMBHJ cell, the series resistance (Rs) of the DBSQ:C<sup>70</sup> SMBHJ cell obviously decreases from 16.2 to 6.8 cm<sup>2</sup> , whereas the shunt resistance (Rsh) increase from 2.9 × 10<sup>3</sup> to 4.0 × 10<sup>3</sup> cm<sup>2</sup> in the same time. The decreased R<sup>s</sup>

and increased Rsh also contribute to higher FF of the vacuumprocessed DIBSQ:C<sup>70</sup> SMBHJ device (Chen et al., 2017b).

**Figure 3b** displays the J-V curves of the SMBHJ cells under dark condition. The solution-processed DIBSQ:PC71BM SMBHJ cell demonstrates a turn-on voltage of approximately 0.8–1.0 V while it is approximately 0.7–0.9 V for the vacuum-processed DIBSQ:C<sup>70</sup> device, which indicates that the superior limit of the attainable Voc in SMBHJ cell, i.e., the built-in potential (Vbi) across the SMBHJ cell, obviously decreases by using the vacuum-processed DIBSQ:C<sup>70</sup> active layer (He et al., 2011; Chen et al., 2017b). This observation means that solution-processed DIBSQ:PC71BM SMBHJ device possesses larger Vbi than the vacuum-processed DIBSQ:C<sup>70</sup> SMBHJ device, which should be ascribed to poorer charge-transporting properties in the solutionprocessed SMBHJ cells. As shown in **Figure 3c**, the EQE spectra of both devices consist of three major peaks at 370, 500, and 700 nm approximately. The first two peaks located at 370 and 500 nm correspond to the absorption of PC71BM or C70, and the third peak at 700 nm is from the absorption of DIBSQ, which is consistent with the absorption spectra of the blend DIBSQ:PC71BM and DIBSQ:C<sup>70</sup> films (**Figure 2b**). Even though the photo response in the longer wavelength region are almost same for the solution- and vacuum-processed SMBHJ cells, the photo response in the near ultraviolet and visible region for the vacuum-processed DIBSQ:C<sup>70</sup> SMBHJ cell is significantly enhanced up to 76 and 72%, respectively, which explains that why the vacuum-processed SMBHJ cell possesses higher Jsc than that of solution-processed SMBHJ cell. The normalized EQE spectra (**Figure S1**) shows that the photo response in the longer wavelength region of the solution-processed DIBSQ:PC71BM SMBHJ cell is much stronger than that of the DIBSQ:C<sup>70</sup> SMBHJ cell, which is consistent with the observation in the normalized absorption spectra of DIBSQ:Fullerene films (**Figure 2b**). Moreover, the obviously higher and longer photo response from 750 to 850 nm should be ascribed to the absorption of charge transfer states between DIBSQ and C<sup>70</sup> (Wang et al., 2013), which also contributes to the higher Jsc in the DIBSQ:C<sup>70</sup> SMBHJ cell.

To further study the effect of fabrication processing on the SMBHJ device performance, we also characterized the photocurrent density (Jph) vs. effective voltage (Veff) of the DIBSQ:PC71BM and DIBSQ:C<sup>70</sup> SMBHJ cells, respectively, as demonstrated in **Figure 3d**. Here, Jph and Veff are defined by

$$J\_{\rm ph} = \rm \rm L - J\rm \tag{1}$$

$$V\_{\text{eff}} = V\_0 - V\_{\text{a}} \tag{2}$$

where J<sup>L</sup> and J<sup>D</sup> represent the current density measured under AM 1.5G illumination and in the dark at an applied bias voltage Vb, respectively. V<sup>0</sup> is the built-in potential, which can by identified as the voltage at Jph = 0 (Wu et al., 2016). **Figure 3d** displays that the Jph increases linearly with Veff at low Veff range (<0.1 V), and then gradually saturates at a high Veff. In general, it is supposed that all photogenerated excitons are dissociated into free carriers and extracted by electrodes at high Veff (Chen et al., 2017b). At lower Veff region, higher Jph can be found for the solution-processed SMBHJ device than that of the vacuum-processed device, which reflects the charge collection in the vacuum-processed SMBHJ cells is more efficient than that in the solution-processed SMBHJ cells. The enhanced charge collection efficiency leads to improved device performance in the vacuum-processed devices, as discussed in the previous sections.

Carrier mobilities in BSQ:Fullerene blend films deposited from different fabrication processing were also characterized and compared, as shown in **Figure 4**. Hole-only diodes were fabricated using the structure of ITO/MoO<sup>3</sup> (5 nm)/DIBSQ:PC71BM (60 nm, 1:5) or DIBSQ:C<sup>70</sup> (60 nm, 1:5)/MoO<sup>3</sup> (5 nm)/Al (100 nm), while the electron-only diodes were prepared using the structure of ITO/Cs2CO<sup>3</sup> (1 nm)/DIBSQ:PC71BM (60 nm, 1:5) or DIBSQ:C<sup>70</sup> (60 nm, 1:5)/BCP (10 nm)/Al (100 nm) for electrons. Where the DIBSQ:PC71BM and DIBSQ:C<sup>70</sup> blend films were prepared by using the same solution- and vacuum-processing, respectively, as those in the SMBHJ devices. The J-V curves of the hole- and electron-only diodes were characterized as shown in **Figure 4**, and then the carrier mobilities were calculated by fitting the J-V data using the SCLC method according to the following equation (3) (Wang et al., 2015):

$$J = \frac{9}{8} \mu \varepsilon\_0 \varepsilon\_r \frac{V^2}{d^3} \tag{3}$$

where J represents the current density, µ represents effective carrier mobility, ε<sup>0</sup> and ε<sup>r</sup> represent the absolute dielectric constant and relative dielectric constant, respectively, d represents the thickness of the DBSQ:PC71BM film and V is the applied voltage (Chen et al., 2017b). The SCLC-estimated hole and electron mobility data were summarized in **Table 1**. The hole and electron mobility in vacuum-processed DIBSQ:C<sup>70</sup> films are about 4.7 and 1.8 times greater than that in the solution-processed DBISQ:PC71BM film, respectively, which shall be ascribed to the closer intermolecular distance in the denser vacuum-processed DIBSQ:C<sup>70</sup> film. Meanwhile the hole and electron mobilities in vacuum-processed DIBSQ:C<sup>70</sup> films are much more balanced than those in solution-processed DBISQ:PC71BM films. More balanced hole and electron mobilities leads to more efficient charge extraction and thus much higher device performance in the DIBSQ:C<sup>70</sup> SMBHJ devices (Chen et al., 2017b).

The light-intensity dependence of device performance for the DIBSQ:Fullerene SMBHJ cell was also characterized and compared, as shown in **Figure 5**. For both of the solution- and vacuum SMBHJ cells, the Jsc is linearly proportional to the light intensity. The slope value (0.91) nears the ideal value (1.0) for OSC device. The Voc of the device shows a sub-linear trend with saturation when the light intensity reaches 100 mW/cm<sup>2</sup> . The slope is roughly close to kT/q, demonstrating that bimolecular recombination dominates, where k is Boltzmann's constant, T is the temperature, and q is elementary charge (Koster et al., 2005; Chen et al., 2017a). The FF decreases with increasing the light intensity, which indicates that the recombination loss of the device is sensitive to both carrier density and electrical field. The PCEs of the solution-processed DIBSQ:PC71BM cell and DIBSQ:C<sup>70</sup> SMBHJ cell improve up to 6.21 and 8.06% at 3.5 mW cm−<sup>2</sup> , respectively, respective to suppressed nongerminated recombination (Chen et al., 2012a), which is a strong indication of the great potential of the DIBSQ:Fullerene SMBHJ cell for the commercial application in the lower light intensity ambient.

To deeply understand the device performance of DIBSQ:Fullerene BHJ cells, we used AFM and XRD characterizations to study the film morphology of the BHJ active layer fabricated by using solution- or vacuum processing, as depicted in **Figure 6**, **Figure S2**. From the AFM images, a small root-mean-square (RMS) roughness of 1.36 and 0.61 nm for

FIGURE 7 | (A) *J*-*V* characteristics of the solution-processed DIBSQ:PC71BM SMBHJ cells at 25 and 80◦C; and (B) *J*-*V* characteristics of the vacuum-processed DIBSQ:C<sup>70</sup> SMBHJ cells at 25 and 80◦C.

the solution-processed DIBSQ:PC71BM and vacuum-processed DIBSQ:C<sup>70</sup> films were determined, respectively. And no peaks can be observed in the XRD patterns. These observations indicate that both of the solution- and vacuum-processed photoactive layers are amorphous. Though the amorphous feature of the photoactive layers is not conducive to charge transport, the small roughness of the photoactive layers are critical to avoid leakage current (Chen et al., 2012b), as demonstrated in **Figure 3b**.

Considering DIBSQ:PC71BM SMBHJ cell has pronounced temperature-dependent performance, which has been systemically studied by our group (Chen et al., 2016a). In this work, we also test the device performance of vacuumprocessed DIBSQ:C<sup>70</sup> SMBHJ cell at higher temperature, and then compare the device performance with that of the solutionprocessed DIBSQ:PC71BM SMBHJ cell at the same testing temperature, as shown in **Figure 7**. For the solution-processed SMBHJ cell, the PCE will significantly increase up to 5.07% with increased Jsc and FF at a testing temperature of 80◦C. While the Voc decreases in the same time owing to the carrier recombination of the SMBHJ device (Chen et al., 2016a). As a result, the PCE is near 20% enhancement at 80◦C comparing with the device performance at 25◦C. However, we can not observe the similar temperature-dependent device performance from the vacuum-processed DIBSQ:C<sup>70</sup> SMBHJ system. The DIBSQ:C<sup>70</sup> SMBHJ cell at 80◦C shows a slightly increased Jsc and FF while slightly decreased Voc and thus a similar PCE compared with the device performance at 25◦C, which can be explained by the almost unchanged carrier mobilities in the DIBSQ:C<sup>70</sup> film compared with those at 25◦C (**Table S1**). This observation is different from that in the DIBSQ:PC71BM SMBHJ cell, in which the carrier mobility will significant increase and thus the device performance will increase in the same time with increasing the testing temperature.

Besides PCE, stability is another important parameter to evaluate the performance of an OSC cell (Ecker et al., 2011). **Figure 8** shows the degradation of normalized PCE of the solution- and vacuum-processed BHJ cells in 1 year, and all the measurements were carried out with all devices kept in air with glass encapsulation. After keeping in air one year, the PCE of the solution-processed DIBSQ:PC71BM SMBHJ cell still remains 93.4% of the initial value, while the PCE of the vacuum-processed DIBSQ:C<sup>70</sup> SMBHJ cell remains 77.3% of the initial number. The higher device stability of the solutionprocessed DIBSQ:PC71BM SMBHJ cell demonstrates that the stable DIBSQ and PC71BM materials employed in the device can efficiently resist the chemical and photochemical degradation

#### REFERENCES


inside the SMBHJ cells, which is a very important property for the commercial application because many OSC materials can realize higher PCE while short lifetime.

#### CONCLUSIONS

In summary, to directly compare the device performance of solution- and vacuum-processed SMBHJ solar cells, we employed a SQ dye, which can be deposited by using both of the solution and vacuum processing, as electron donor combined with fullerene as acceptor to construct solutionand vacuum-processed DIBSQ:Fullerene SMBHJ cells. Then the device performance were characterized and compared. The results demonstrates that the vacuum-processed cell provides a ∼47% higher PCE than that of the solution-processed SMBHJ cell due to more efficient charge transport and charge extraction in the vacuum-processed SMBHJ cells. However, solution-processed SMBHJ cells demonstrate more pronounced temperature-dependent device performance and higher device stability. The light intensity-dependent device performance for both of the solution- and vacuum-processed DIBSQ:Fullerene SMBHJ cell indicates their promising application in the lower light intensity ambient. This study indicates the great potential of DIBSQ in photovoltaic application via both of solution and vacuum processing techniques.

### AUTHOR CONTRIBUTIONS

GC, HS, and JK designed experiments, GC and ZL carried out experiments, BW and ZH analyzed experimental results GC, JZ, and HS wrote the manuscript.

#### ACKNOWLEDGMENTS

This work is financially supported by the Shanghai Pujiang Program (16PJ1403300), the National Natural Scientific Foundation of China (61604093), the Natural Science Foundation of Shanghai (16ZR1411000), and the Science and Technology Commission of Shanghai Municipality Program (17DZ2281700).

#### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem. 2018.00412/full#supplementary-material


polymer solar cells. Synthetic Met. 192, 10–14. doi: 10.1016/j.synthmet.2014. 02.018


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Chen, Ling, Wei, Zhang, Hong, Sasabe and Kido. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

## Urea-Doped ZnO Films as the Electron Transport Layer for High Efficiency Inverted Polymer Solar Cells

Zongtao Wang<sup>1</sup> , Zhongqiang Wang<sup>1</sup> \*, Ruqin Zhang<sup>1</sup> , Kunpeng Guo<sup>1</sup> \*, Yuezhen Wu<sup>1</sup> , Hua Wang<sup>1</sup> , Yuying Hao<sup>1</sup> and Guo Chen<sup>2</sup>

*<sup>1</sup> Key Laboratory of Interface Science and Engineering in Advanced Materials, Ministry of Education, Research Center of Advanced Materials Science and Technology, Taiyuan University of Technology, Taiyuan, China, <sup>2</sup> Key Laboratory of Advanced Display and System Applications, Ministry of Education, Shanghai University, Shanghai, China*

#### Edited by:

*Donghong Yu, Aalborg University, Denmark*

#### Reviewed by:

*Xichang Bao, Qingdao Institute of Bioenergy and Bioprocess Technology (CAS), China Weiwei Li, Institute of Chemistry (CAS), China*

#### \*Correspondence:

*Zhongqiang Wang wangzhongqiang@tyut.edu.cn Kunpeng Guo guokunpeng@tyut.edu.cn*

#### Specialty section:

*This article was submitted to Organic Chemistry, a section of the journal Frontiers in Chemistry*

Received: *30 May 2018* Accepted: *20 August 2018* Published: *07 September 2018*

#### Citation:

*Wang Z, Wang Z, Zhang R, Guo K, Wu Y, Wang H, Hao Y and Chen G (2018) Urea-Doped ZnO Films as the Electron Transport Layer for High Efficiency Inverted Polymer Solar Cells. Front. Chem. 6:398. doi: 10.3389/fchem.2018.00398* In this paper, urea-doped ZnO (U-ZnO) is investigated as a modified electron transport layer (ETL) in inverted polymer solar cells (PSCs). Using a blend of Poly{4,8-bis[(2-ethylhexyl)oxy] benzo [1,2-b:4,5-b'] dithiophene-2,6-diyl-alt-3-fluoro-2-[(2-ethylhexyl)carbonyl] thieno [3,4-b] thiophene-4,6-diyl}(PTB7), and [6,6]-phenyl-C71-butyric acid methyl ester (PC71BM) as light absorber, a champion power conversion efficiency (PCE) of 9.15% for U-ZnO ETL based PSCs was obtained, which is 15% higher than that of the pure ZnO ETL based PSCs (7.76%). It was demonstrated that urea helps to passivate defects in ZnO ETL, resulting in enhanced exciton dissociation, suppressed charge recombination and efficient charge extraction efficiency. This work suggests that the utilization of the U-ZnO ETL offer promising potential for achieving highly efficient PSCs.

Keywords: polymer solar cells, ETL, ZnO, urea, PCE

### INTRODUCTION

In past decades, the need of green and sustainable energy has become more and more pressing. Polymer solar cells (PSCs) as one of the most potential renewable energy technologies have attracted wide attention due to its merits, such as low cost, light weight, and the capability of fabricating flexible large-area modules (Zhang et al., 2016; Sun et al., 2017; Tran et al., 2017; Hou et al., 2018; Li et al., 2018). To improve the performance of PSCs, enormous efforts have been made by the worldwide researchers. Benefited from the progresses of new optoelectronic materials and interfacial engineering, the power conversion efficiency (PCE) have reached over 14% in single-junction PSCs (Gao et al., 2016; Liu et al., 2016; Wu et al., 2016; Li et al., 2017; Peng et al., 2017). Meantime, it was found the interfacial engineering plays key role in determining the efficiency and the stability of PSCs (Liu, Z. et al., 2017). For example, it has been demonstrated that the application of anode and electrode interlayer leads to 20% improvement compared to the reference device without electrode interlayer (Wu et al., 2016). The interfacial contact at the interface between active layer and electrode usually polished by inorganic and organic materials (such as graphene, carbon quantum dots, PEIE, PEI), which benefits the charge transport and extraction in PSCs, resulting in enhanced device performance (Bi et al., 2018).

In conventional structure devices, low work-function metal and air sensitive materials are widely used as electron extraction layer (Liu, D. et al., 2017; Huai et al., 2018), which limits the stability of devices. Hence, the inverted device structure is developed in PSCs, which switches the position of anode and cathode, preventing the interfacial deterioration and physical degradation of organic layers (Wang et al., 2012; Cheng et al., 2018). Therefore, the device stability can be significantly improved in inverted structure devices. In addition, the inverted PSCs showed excellent performance due to the efficient charge extraction and suppressed recombination loss compared to the conventional structure devices (Wang et al., 2017).

In inverted PSCs, the metal-oxide materials have been utilized as buffer layers to extract and transport charge carriers, such as TiOx, CsOx, SnO2, ZnO (Tozlu et al., 2017; Tran et al., 2017; Wang et al., 2017; Jung et al., 2018). Due to the excellent electron extraction property, high electron mobility and easy processing, environment-friendly ZnO is widely used in the inverted PSCs (Vohra et al., 2015).. However, Sol-Gel processed ZnO films exhibit certain defects on the surface and inside, which limit the application as electron transport layer (ETL) in the highly efficient PSCs (Gu et al., 2014). These defects are apt to cause electron trap loss in PSCs, resulting in the severe charge carrier recombination loss. Additionally, poor contact between inorganic and organic materials also cause bad electrical contact, leading to recombination loss (Han et al., 2016).

To facilitate the electron extraction in PSCs, many attempts have been made to polish the interface between ZnO ETL and active layer (Zhang et al., 2018). The double-layer structure of ZnO/organic was developed to modify ZnO ETL and prevent direct contact of ZnO ETL with active layer (Zhang et al., 2018). Compared to pristine ZnO ETL, double-layer structure ZnO/organic ETL showed impressive enhancement in PCE. Following this principle, various materials, such as ionic liquids, polyelectrolytes, self-assembled monolayers, have been introduced to form double-layer structure ETL in PSCs (Dai et al., 2011; Zhang et al., 2018). However, new interfaces are introduced into the devices after the application of double-layer structure ETL, resulting in new interfacial contact. Thus, it is highly desirable but challenged to develop efficient and simple ETL in PSCs.

In order to address the above mentioned problems, we envisaging that modifying ZnO with suitable material would help to eliminate or passivate defects in ZnO films. To realize this, we noticed urea would be a good dopant candidate. This is because the possibility of formation of coordinate bonds between the lone pair of electrons on the two nitrogen atom of the amino groups and the Lewis acid sites of the ZnO surface would enable defects reducing.

In this work, urea-doped ZnO (U-ZnO) was investigated as an electron transport layer (ETL) in inverted PSCs with a blend of Poly{4,8-bis[(2-ethylhexyl)oxy]benzo[1,2-b:4,5-b']dithiophene-2,6-diyl-alt-3-fluoro-2-[(2-ethylhexyl)carbonyl]thieno[3,4 b]thiophene-4,6-diyl} (PTB7), and [6,6]-phenyl-C71-butyric acid methyl ester (PC71BM) as active layer. With the device structure of ITO/ U-ZnO (40 nm)/PTB7:PC71BM(95 nm)/MoO3(5 nm)/Al (80 nm), an optimal PCE value of 9.15% was achieved, which is 15% higher than that of the pure ZnO ETL based reference device (7.76%). The performance enhancement of the U-ZnO ETL based PSCs may be attributed to the promotion of exciton dissociation, the suppression of charge recombination and the improvement of charge extraction capability. This investigation proves that the U-ZnO ETL offers promising potential for achieving high efficiency in PSCs.

## EXPERIMENTAL

### General Information

PTB7, PC71BM, and MoO<sup>3</sup> were purchased from 1-Material Company of United States, Luminescence Technology Corporation of Taiwan and Rieke Company of China, respectively. Anhydrous zinc acetate [Zn(CH3COO)2] and 2-methoxyethanol (CH3OCH2CH2OH) were obtained from Energy Chemical. Ethanolamine (NH2CH2CH2OH) was purchased from Sigma-Aldrich Company. All materials and solvents were used as purchased without further purification.

## Device Fabrication

To obtain the ZnO precursor solution, 0.836 g of Zn(CH3COO)<sup>2</sup> and 0.28 g of NH2CH2CH2OH were dissolved in 10 mL of CH3OCH2CH2OH. Different ratios of urea were added into CH3OCH2CH2OH to prepare the U-ZnO precursor. All the precursor solutions were stirred overnight in air. A blend of PTB7:PC71BM (1:1.5 wt %, 25 mg mL−<sup>1</sup> ) was dissolved in the mixed solvents of CB and DIO (97:3 vol%) and was stirred overnight at 50◦C. Pre-patterned ITO substrates (15 Ω/square) were cleaned in detergent, acetone, and isopropanol for 30 min in sequence. Then the substrates were dried overnight in an oven with a temperature of 50◦C. The ultraviolet-ozone treated ITO substrates were used to spin coat ZnO precursor or U-ZnO precursor at 4000 rpm in air. Then the ZnO-coated or ZnOcoated substrates were annealed at 200◦C for 1 h in air. After that, the substrates were put into a N2-filled glovebox for active layer spin-coating. Then the samples were quickly transfer into vacuum chamber for electrode deposition. Finally, 5 nm MoO<sup>3</sup> and 80 nm Al were deposited in a vacuum chamber with base press of 1 × 10−<sup>4</sup> Pa.

## Characterization

The current density-voltage (J-V) characteristics, PCE and external efficiency quantum (EQE) were recorded by a Newport solar simulator system. The J-V characteristics were recorded under AM 1.5G illumination with light intensity of 100 mW cm−<sup>2</sup> . A calibrated silicon diode was set as the reference, which exhibited a response from 300 to 800 nm. The photoluminescence (PL) samples were measured by A Hitachi F-7000 spectrofluorophotometer.

## RESULTS AND DISCUSSION

The device architecture of inverted PSCs was shown in **Figure 1A**, with a structure of ITO/U-ZnO(40 nm)/PTB7:PC71BM(95 nm)/MoO3(5 nm)/Al. According to our previous study, the thicknesses of ETL and active layer were set to the optimal values of 40 and 95 nm in the device fabrication, respectively(Zhang et al., 2018). The molecular structures of PTB7 and PC71BM were illustrated in **Figure 1B**.

As a ETL in PSCs, high transmittance is the priority, which is highly related to the photovoltaic performance. As shown in Supporting Information **Figure S1**, the U-ZnO ETLs show high transmittance, which is slightly lower than pure ZnO ETL. Moreover, Supporting Information **Figure S2** compares the light absorption in PTB7:PC71BM active layers. The light absorption spectra overlap each other, indicating good transmittance of U-ZnO ETL. To optimize the photovoltaic performance of PSCs, the urea doped concentration was tuned from 0 to 5 mg mL−<sup>1</sup> in ZnO precursor solution. **Figure 2A** shows J-V characteristics and **Figure 2B** shows EQE photoresponse spectra of pure ZnO and U-ZnO ETLs based inverted PSCs. A summary of the corresponding photovoltaic parameters is listed in **Table 1**. The reference device exhibited an open circuit voltage(Voc) of 0.73 V, a short circuit current density(Jsc)of 15.38 mA/cm<sup>2</sup> , a fill factor (FF) of 68.75%, and a PCE of 7.76%. In comparison with the reference device, the U-ZnO ETL based devices showed an obvious improvement in Jsc and FF, resulting in enhanced PCE. As a result of variations of Jsc, FF and Voc in inverted PSCs, a champion efficiency of 9.15% was obtained with a Voc of 0.74 V, a Jsc of 17.31 mA/cm<sup>2</sup> , and a FF of 71.43% for the device using 3 mg mL−<sup>1</sup> U-ZnO as the ETL. As listed in **Table 1**, the values of Jsc, FF and PCE were increased by increasing the concentration of urea from 0 to 3 mg mL−<sup>1</sup> and then decreased at high concentration > 3 mg mL−<sup>1</sup> . The increased Jsc was supported by the EQE profiles of the devices. The Supporting Information **Figure S3** plots the current density (16.86 mA/cm2) of champion cell calculated from EQE,

showing lower discrepancy with the Jsc (17.31 mA/cm<sup>2</sup> ) extracted from J-V curve. The enhanced photovoltaic performance in PTB7:PC71BM based devices means the superior property of U-ZnO ETL in inverted PSCs.

To investigate the interface contact between U-ZnO ETL and active layer, the contact angle of water droplets on pure ZnO and U-ZnO ETLs were measured in this study (Zisman, 2008; Peng et al., 2017; Han et al., 2018). As shown in **Figure 3**, the contact angles of water on 0, 3, 5 mg mL−<sup>1</sup> U-ZnO ETLs were 26, 31.5 and 29.5◦ , respectively. In comparison with pure ZnO ETL, the U-ZnO ETL show much higher hydrophobicity, resulting in better interface contact between ETLs and photoactive layers. Thus, the better interface contact is beneficial to the improved performance of U-ZnO ETL based inverted PSCs.

It is known that the morphology of ETL is critical in determining the performance of PSCs(Tran et al., 2017). Hence, the surfacemorphology of the ETLs without and with the urea-doping are studied by atomic force microscopy (AFM), and the top images are displayed in **Figure 4**. The Root-Mean-Square (RMS) roughness of ZnO ETL without ureadoping is 1.19 nm. The RMS roughness is decreased to 0.74 and 1.14 nm in U-ZnO with 3 mg mL−<sup>1</sup> and 5 mg mL−<sup>1</sup> , respectively. The smooth surface after urea-doping modifies the contact between ETLs and active layers, and infects the

TABLE 1 | Photovoltaic parameters of inverted PSCs based on PTB7:PC71BM using U-ZnO as the ETL under AM 1.5G irradiation (100 m W cm <sup>−</sup><sup>2</sup> ). The values of Jcal are calculated from the related EQE spectra.


charge transport and collection in PSCs, leading to photocurrent improvement. On the other hand, the 2-D and 3-D images of PTB7:PC71BM active layers on U-ZnO ETLs are shown in Supporting Information **Figure S4**. AFM images of active layers reveal little morphology variation. The RMS surface roughness is slightly changed.

The space-charge-limited current (SCLC) method can be utilized to study the effect of urea on charge transport behavior inside the PSCs (Gupta et al., 2018). Hence, the electron mobility in ZnO and U-ZnO based devices were extracted by SCLC model. The electron-only device structures were ITO/ZnO or U-ZnO(40 nm)/PTB7:PC71BM(95 nm)/BCP(8 nm)/Al. As shown in Supporting Information **Figure S5**, the electron-only device J-V characteristics were fitted by SCLC model. The electron mobility in PSCs based on 0, 3, 5 mg mL−<sup>1</sup> urea-doping U-ZnO ETLs are 1.87 × 10−<sup>4</sup> , 1.14 × 10−<sup>3</sup> , and 3.84 × 10−<sup>4</sup> cm<sup>2</sup> V −1 S −1 , respectively. The superior electron mobility after ureadoping is beneficial to the electrical properties in PSCs, leading to efficient charge transport and extraction.

To understand the contribution of U-ZnO ETL in PSCs, the dependence of J-V characteristics with incident light intensity were compared in U-ZnO ETL based devices (Xiao et al., 2018). The plots in **Figure 5A** were fitted with power law, then the values of α were obtained (Schilinsky et al., 2002). As shown in **Figure 5A**, each α value is close to unity, indicating efficient charge transport and collection in these inverted PSCs at short circuit condition (Huang et al., 2018). Slightly improved α was observed after introduction of urea in precursor solution, which means more efficient charge transport and collection in inverted PSCs using U-ZnO as the ETL.

Furthermore, photo generated charge carriers recombine again inside of cells under the open circuit condition, which directly reflects the recombination loss of cells. Therefore,the dependence of Voc with incident light intensity was studied in this work. The Voc shows linear dependence with semi logarithmic incident light intensity with a slope of KT e (Yang et al., 2018), where K is the Boltzmann constant, T is the temperature in Kelvin, and e is the elementary charge. **Figure 5B** shows the linear dependence of Voc as a function with incident light intensity and the extracted slope. It can be seen from the extracted slope that the values of slope are decreased after adding urea into ZnO precursor. Compared to the slope of 1.23 KT e in PSCs using pure ZnO as the ETL, a smaller slope of 1.11 KT e is obtained when the concentration of urea in ZnO precursor solution is 3 mg mL−<sup>1</sup> , indicating the suppressed interfacial trap defects of the U-ZnO ETL. Consequently, the introduction of urea in ZnO precursor solution could passivate the defects of Sol-Gel processed ZnO ETL, resulting in improved Jsc and FF in U-ZnO ETL based PSCs.

To further study the enhancement of Jsc, the maximum exciton generation rate (Gmax) of PSCs was calculated in this study (Zhang et al., 2017). **Figure 6A** shows the dependence of photocurrent density (Jph) with the effective voltage (Veff ). Apparently, two different regions were observed in Jph-Veff characteristics. Under low effective voltage, the Jph shows linear dependence with Veff . Then, it gradually approaches saturated value (Jsat) under high effective voltage. The values of Gmax are showed in **Figure 6C**. An obvious improvement in Gmax for U-ZnO ETL based PSCs is clearly seen in **Figure 6C**, which means efficient charge carrier transport and extraction. The enhanced Gmax highly contributes to the enhanced Jsc in U-ZnO ETL based PSCs.

In organic PSCs, only a portion of excitons dissociate into free charge carriers due to the unique optoelectronic conversion behavior. The exciton dissociation probability P (E, T) is related to electric field (E) and temperature (T). Hence, the value of P (E, T) under zero bias is deduced from **Figure 6B**. The value of P (E, T) increased from 81.6% in the control device to 88.2% in the U-ZnO ETL based device, implying that U-ZnO ETL can promote excitons dissociation. The improved excitons dissociation probability also contributes to the enhancement of Jsc.

Due to the priority of exciton quenching in PSCs, photoluminescence (PL) spectra was used to analyze PL effect of pure ZnO and U-ZnO ETLs (Anger et al., 2006). **Figure 7A** shows the PL spectra of samples with a structure of ITO/ETL (40 nm)/PTB7 (50 nm)/MoO<sup>3</sup> (5 nm)/Al (10 nm). In comparison with pure ZnO ETL, significant enhancement of PL

intensity was observed in the U-ZnO ETL based samples, which indicated suppressed exciton quenching at the interface of ETL and active layer. The suppressed exciton quenching is beneficial to the improvement of Jsc.

Electric impedance spectroscopy (EIS) is usually utilized to analyze the electrical behavior in PSCs (Zhang et al., 2018). Here, EIS spectrum is used to reveal the charge recombination inside of devices. The Nyquist plot of the impedance spectroscopy under dark condition was plotted in **Figure 7B**. Shorter diameter was observed after introduction of urea, meaning a lower transport resistance in U-ZnO ETL based devices. The decreased transport resistance reflects better contact between U-ZnO ETL and active layer, resulting in an efficient charge transport probability. The EIS spectra analyses also prove the efficient charge transport and extraction in U-ZnO ETL based PSCs.

#### CONCLUSIONS

In summary, U-ZnO as the ETL was applied in PSCs. The advantages of U-ZnO ETL were analyzed in PTB7:PC71BM based inverted PSCs. The introduction of urea helps to passivate the defects of Sol-Gel processed ZnO ETL, polish the interface contact, and promote the exciton dissociation. In comparison with the devices using pure ZnO as the ETL, an impressive improvement was observed in PTB7:PC71BM based PSCs with the U-ZnO ETL. A champion efficiency of 9.15% was obtained with ∼15% enhancement compared to the efficiency of 7.76% in pure ZnO ETL based PSCs. Our results suggest that U-ZnO ETL have great potential in organic PSCs.

### REFERENCES


#### AUTHOR CONTRIBUTIONS

Device fabrication and photovoltaic performance studies were carried out by ZoW, RZ, and YW. ZhW, KG, HW, YH, and GC contributed to discussions. ZhW led the project, and prepared the manuscript. All authors contributed to the manuscript.

## FUNDING

This study was supported by National Natural Science Foundation of China (NSFC) (Grant No. 61704118), Shanxi Provincial Natural Science Foundation of China (Grant No. 201601D021050). This study was also supported by the Qualified Personal Foundation of Taiyuan University of Technology (800101-02030017).

#### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem. 2018.00398/full#supplementary-material


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Wang, Wang, Zhang, Guo, Wu, Wang, Hao and Chen. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

## Electron Acceptors With a Truxene Core and Perylene Diimide Branches for Organic Solar Cells: The Effect of Ring-Fusion

Kaiwen Lin† , Shiliang Wang† , Zhenfeng Wang, Qingwu Yin, Xi Liu, Jianchao Jia, Xiao'e Jia, Peng Luo, Xiaofang Jiang, Chunhui Duan\*, Fei Huang\* and Yong Cao

*State Key Laboratory of Luminescent Materials and Devices, Institute of Polymer Optoelectronic Materials and Devices, South China University of Technology, Guangzhou, China*

#### Edited by:

*Chuanlang Zhan, Institute of Chemistry (CAS), China*

#### Reviewed by:

*Ling Teng Ye, Harbin Institute of Technology, China Gregory C. Welch, University of Calgary, Canada*

#### \*Correspondence:

*Chunhui Duan duanchunhui@scut.edu.cn Fei Huang msfhuang@scut.edu.cn*

*†These authors have contributed equally to this work*

#### Specialty section:

*This article was submitted to Organic Chemistry, a section of the journal Frontiers in Chemistry*

Received: *26 April 2018* Accepted: *12 July 2018* Published: *04 September 2018*

#### Citation:

*Lin K, Wang S, Wang Z, Yin Q, Liu X, Jia J, Jia X, Luo P, Jiang X, Duan C, Huang F and Cao Y (2018) Electron Acceptors With a Truxene Core and Perylene Diimide Branches for Organic Solar Cells: The Effect of Ring-Fusion. Front. Chem. 6:328. doi: 10.3389/fchem.2018.00328* In this work, a star-shaped planar acceptor named FTr-3PDI was synthesized via ring-fusion between truxene core and three bay-linked perylene diimide (PDI) branches. Compared to the unfused non-planar acceptor Tr-3PDI, FTr-3PDI exhibits better structural rigidity and planarity, as well as more effective conjugation between truxene core and PDI branches. As a result, FTr-3PDI shows up-shifted energy levels, enhanced light absorption coefficient, increased electron mobility, and more favorable phase separation morphology in bulk-heterojunction (BHJ) blend films as compared to Tr-3PDI. Consequently, FTr-3PDI afforded higher power conversion efficiency (PCE) in BHJ solar cells when blended with a polymer donor PTB7-Th. This work demonstrates that ring-fusion is a promising molecular design strategy to combine the merits of truxene and PDI for non-fullerene acceptors used in organic solar cells.

Keywords: organic solar cells, star-shaped electron acceptors, truxene, perylene diimide, ring-fusion

## INTRODUCTION

Recently, non-fullerene electron acceptors have received considerable attention in the community of organic solar cells (OSCs) due to their energy level tunability, intense optical absorption properties, and potential for low-cost and large-scale fabrication (Cheng et al., 2018; Hou et al., 2018; Yan et al., 2018). Among them, perylene diimide (PDI) derivatives are widely investigated in bulk-heterojunction (BHJ) OSCs because of their intense light absorption and high electron mobility (Zhan et al., 2007, 2011; Lin et al., 2014; Sun et al., 2015; Hendsbee et al., 2016; Liu J. et al., 2016; Liu Z. T. et al., 2016; Meng et al., 2016a). Despite these favorable properties, PDI monomer shows low device performance due to the formation of large aggregated nanostructure and undesired large crystalline domains caused by the large coplanar structure of PDI block, which hamper the exciton diffusion and separation process (Sharenko et al., 2013; Liu S. Y. et al., 2015). To overcome these drawbacks, an effective strategy is to develop non-coplanar PDI-based molecules via forming twisted intramolecular structures (Zhong et al., 2014, 2016; Lin et al., 2016; Zhang et al., 2016; Duan et al., 2017a; Liu X. et al., 2017; Liu et al., 2018). For example, a lot of star-shaped electron acceptors with PDI branches were reported recently based on this design guideline (Lin et al., 2014, 2016; Liu Y. H. et al., 2015; Lee et al., 2016; Duan et al., 2017a; Zhang A. D. et al., 2017). Although these star-shaped PDI electron acceptors can avoid forming large crystalline domains, their highly twisted architectures decrease the intermolecular contact and orbital overlapping between PDI π-planes, thus hampering electron hopping between molecules. Therefore, the key point to develop high-performance PDI electron acceptors is to obtain a balance between highly twisted non-planar structures for forming proper phase separation in blend films and strong intermolecular interaction for supporting sufficient charge transport ability.

Recently, several studies showed that oxidative ring-fusion between the PDI branches and the central aromatic core of PDI-based molecules is an effective strategy to achieve such an exquisite balance (Hartnett et al., 2016; Meng et al., 2016b, 2017; Zhong et al., 2016; Wang B. et al., 2017; Zhang J. Q. et al., 2017). The resulting fused PDI molecules exhibit better structural rigidity and planarity, as well as more effective conjugation between the aromatic core and PDI branches. Meanwhile, the fused PDI molecules show stronger intermolecular π-π stacking and higher electron mobility. Moreover, these fused PDI acceptors can lead to desirable film morphology with proper domain size and high domain purity in BHJ blends when blended with donor polymers (Meng et al., 2016b, 2017; Wang B. et al., 2017; Zhang J. Q. et al., 2017). Therefore, the fused PDI acceptors display significantly improved photovoltaic performance compared to their carbon-carbon single bond connected counterparts (**Scheme 1**) (Li et al., 2016; Meng et al., 2016b, 2017; Liu X. F. et al., 2017; Wang B. et al., 2017; Zhang J. Q. et al., 2017). Actually, the best-performing OSCs based on PDI acceptors was achieved by a star-shaped fused PDI molecule named FTTB-PDI4, which afforded a power conversion efficiency (PCE) of 10.58% (Zhang J. Q. et al., 2017).

Among various central cores for star-shaped electron acceptors, truxene has been proved to be promising for constructing high-performance optoelectronic materials (Nielsen et al., 2013, 2014; Lin et al., 2018; Wu et al., 2018). The rigid coplanar structure and unique C3h symmetry contribute to well-delocalized electronic structure in extended dimensionality for the resulting star-shaped conjugated molecules, which in turn result in strong light absorption and effective charge transport. Recently, Peng's group reported a state-of-the-art truxenebased electron acceptor for application in OSCs, which yielded impressive PCE exceeding 10% (Wu et al., 2018). These results suggested the promising prospect of truxene for constructing high-performance electron acceptors.

Inspired by these achievements, herein, we report the design and synthesis of two star-shaped acceptors named Tr-3PDI and FTr-3PDI, (**Scheme 2A**) where the truxene core and PDI branches are linked by carbon-carbon single bonds or via ring-fusion, respectively. We further evaluated their potential as electron acceptors in OSCs with poly[4,8-bis(5-(2-ethylhexyl)thiophen-2-yl)benzo[1,2-b;4,5 b]dithiophene-2,6-diyl-alt-(4-(2-ethylhexyl)-3-fluorothieno[3,4 b]thiophene)-2-carboxylate-2,6-diyl] (PTB7-Th) as the donor. The solar cells based on Tr-3PDI and FTr-3PDI exhibited a PCE of 2.2 and 3.8%, respectively. The better device performance of the fused acceptor FTr-3PDI is attributable to more favorable energy level alignment with the polymer donor PTB7-Th, more intense light absorption, stronger intermolecular packing, higher electron mobility, and more proper morphology in blend film. This work suggests the potential of ring-fusion strategy for constructing high performance PDI electron acceptors based on truxene core.

## EXPERIMENTAL SECTION

## Materials and Synthesis

the B3LYP/6-31G(d) level.

All reagents were obtained from commercial sources and used without further purification, unless otherwise specified. **Scheme 2A** shows the synthetic routes of Tr-3PDI and FTr-3PDI. The detailed synthesis procedures are described as following.

#### Tr-3PDI

A mixture of 2,2′ ,2′′-(5,5,10,10,15,15-hexahexyl-10,15-dihydro-5H-diindeno[1,2-a:1′ ,2′ -c]fluorene-2,7,12-triyl)tris(4,4,5,5-

tetramethyl-1,3,2-dioxaborolane (truxene boronic acid pinacol ester, 0.613 g, 0.5 mmol) and 5-Bromo-2,9-bis(1 pentylhexyl)anthra[2,1,9-def:6,5,10-d′ e ′ f ′ ]diisoquinoline-1,3,8, 10(2H,9H)-tetrone (monobromo-PDI, 1.746 g, 2.25 mmol) in anhydrous dimethylformamide (40 mL) was degassed for 30 min before Pd(PPh3)<sup>4</sup> (58 mg, 0.05 mmol) and K2CO<sup>3</sup> aqueous solution (2 M, 10 mL) was added. The solution was heated at 95◦C for 48 h. Water and dichloromethane were added, and the organic layer was dried over MgSO4. After removing the solvent, the crude product was chromatographically purified on silica gel column (eluted with ethyl acetate:petrolem ether = 1:20) to afford Tr-3PDI as a brownish-red solid (0.95 g, 65%). <sup>1</sup>H NMR (500 MHz, CDCl3) δ: 8.74 (m, 18H), 8.60 (m, 6H), 7.62 (m, 6H), 5.25 (m, 6H), 3.04 (m, 6H), 2.17 (m, 30H), 1.29 (m, 112H), 0.86 (m, 72H). <sup>13</sup>C NMR (125MHz, CDCl3) δ: 165.00, 163.91, 155.86, 155.72, 146.31, 142.03, 141.28, 140.62, 138.11, 135.10, 134.57, 132.77, 131.59, 129.98, 129.46, 128.77, 128.34, 127.78, 126.99, 126.80, 123.65, 122.84, 122.26, 56.45, 56.40, 54.95, 54.67, 37.20, 37.07, 32.52, 32.34, 31.91, 31.84, 31.83, 31.81, 31.59, 29.85, 29.80, 29.62, 29.39, 26.78, 26.69, 24.37, 24.13, 22.73, 22.66, 22.64, 22.60, 22.50, 22.43, 14.19,14.16, 14.14. MS (MALDI-TOF) calculated for C201H246N6O12, 2938.21; found, 2937.88.

#### FTr-3PDI

Tr-3PDI (293.7 mg, 0.1 mmol) was dissolved in 20 mL chlorobenzene before adding a catalytic amount of iodine (about 2 mg). The resultant mixture was stirred for 1 h under lab environment. After the reactivation process, kept the closed stand-up bottle exposing to irradiation of 500 W mercury lamp for 5 h at room temperature. The color of the solvent turned to brownish-yellow from brownish-red. After the reaction, the solvent was concentrated and the residue was purified by silica gel column chromatography (hexane:dichloromethane = 1:1) to afford a brownish-yellow solid (263.8 mg, 90%). <sup>1</sup>H NMR (500 MHz, CDCl3) δ: 10.65 (m, 9H), 9.67 (s, 3H), 9.37 (m, 6H), 9.17 (m, 6H), 5.62 (m, 6H), 3.83 (m, 6H), 3.17 (m, 6H), 2.63 (m, 12H), 2.10 (m, 12H), 1.47 (m, 90H), 1.38 (m, 15H), 0.91 (m, 80H), 0.36 (m, 18H). <sup>13</sup>C NMR (125MHz, CDCl3) δ: 165.06, 164.52, 155.41, 149.23, 141.71, 139.02, 134.27, 133.92, 129.76, 129.33, 128.73, 127.83, 127.71, 125.41, 125.27, 124.98, 123.50, 123.36, 119.83, 117.66, 57.60, 55.26, 38.26, 32.81, 32.03, 31.56, 29.57, 27.01, 24.91, 22.85, 22.81, 22.32, 14.30, 14.25, 13.84, 13.82. MS (MALDI-TOF) calculated for C201H240N6O12, 2932.16; found, 2931.94.

#### Instruments and Characterization

<sup>1</sup>H and <sup>13</sup>C NMR spectra were tested on a Bruker AV-500 with tetramethylsilane (TMS) as an internal reference. MALDI-TOF-MS was performed by using a Bruker Agilent1290/maXis impact. UV-vis spectra were measured on a HP 8453 spectrophotometer. Thermogravimetric (TGA) analysis was measured on a NETZSCH TG 209 at a heating rate of 10◦C min−<sup>1</sup> with a nitrogen flow rate of 20 mL min−<sup>1</sup> . Cyclic voltammetry data were measured on a CHI600D electrochemical workstation with Bu4NPF<sup>6</sup> (0.1 M) in acetonitrile as the electrolyte, a carbon electrode and a saturated calomel electrode as the working and reference electrodes, respectively. The thin films were coated on a glassy carbon working electrode. The scan rate was 100 mV s −1 . The geometry was optimized by Density Functional Theory (DFT) calculations performed at the B3LYP/6-31G(d) level to optimize the ground state geometries of the acceptor molecules using the Gaussian 09. The transient photocurrent of devices was measured by applying 500 nm laser pulses with a pulse width of 120 fs to the devices, which produced a transient voltage signal on a 50 resistor and recorded by an oscilloscope (Tektronix EDS 3052C). The laser pulses were generated from optical parametric amplifier (TOPAS-Prime) pumped by a mode-locked Ti:sapphire oscillator seeded regenerative amplifier with a pulse energy of 1.3 mJ at 800 nm and a repetition rate of 1 KHz (Spectra Physics Spitfire Ace). The atom force microscopy (AFM) images were obtained from a NanoMan VS microscopy under tapping mode. The transmission electron microscopy (TEM) images were characterized with a JEM-2100F instrument.

## Fabrication and Characterization of Solar Cells

The devices of indium tin oxide (ITO)/poly(3,4 ethylenedioxythiophene): poly(styrenesulfonate) (PEDOT:PSS)/ PTB7-Th:acceptor/poly[(9,9-bis(3′ -((N,N-dimethyl)-N-

ethylammonium)-propyl)-2,7-fluorene)-alt-2,7-(9,9-

dioctylfluorene)]dibromide (PFN-Br)/Al were fabricated through the following procedures. The ITO-coated glass substrate was cleaned in an ultrasonic bath with deionized water, acetone, and isopropanol, each process was approximately 15 min, and then dried under a stream of dry nitrogen. PEDOT:PSS (Heraeus Clevios PVPA 4083) was spin-coated on top of the above ITO and annealed in air at 150◦C for 10 min. Subsequently, the blend solutions of PTB7-Th and truxene-PDI acceptors were prepared by simultaneously dissolving both materials with the optimized weight ratio in ortho-dichlorobenzene and spin-coated on the ITO/PEDOT:PSS electrode (at 1,600 rpm for 60 s) to form an active layer with thickness of about 100 nm. Then PFN-Br and Al layer were thermally deposited onto the active layer through a shadow mask at a vacuum of 5 × 10−<sup>5</sup> Pa. During the test, an aperture with an area of 3.14 mm<sup>2</sup> was used. The current density–voltage (J–V) curves were measured on a computer-controlled Keithley 2400 source meter under 1 sun, the AM 1.5 G spectra came from a class solar simulator (Enlitech, Taiwan), and the light intensity was 100 mW cm−<sup>2</sup> as calibrated by a China General Certification Center-certified reference monocrystal silicon cell (Enlitech). Before the J–V measurement, a physical mask with an aperture with precise area of 0.04 cm<sup>2</sup> was used to define the device area. The external quantum efficiency (EQE) spectra were measured on a commercial QE measurement system (QE-R3011, Enlitech).

## Fabrication and Characterization of Single-Carrier Devices

The charge carrier mobilities of PTB7-Th:truxene-PDI acceptor blend films were determined from single-carrier devices with space-charge-limited current (SCLC) model. The device structures of the electron only and hole only devices are ITO/ZnO/PTB7-Th:acceptor/Ca/Al and ITO/PEDOT:PSS/PTB7-Th: acceptor/MoO3/Ag respectively. The mobilities were determined by fitting the dark J–V current to the model of a single carrier SCLC using the equation: J = 9ε0εrµV 2 /8d 3 , where J is the current density, d is the thickness of the blend films, ε<sup>0</sup> is the permittivity of free space, ε<sup>r</sup> is the relative dielectric constant of the transport medium, and µ is the charge carrier mobility. V = Vapp-Vbi, where Vapp is the applied voltage and Vbi is the built-in voltage. The carrier mobility can be calculated from the slope of the J 1/2 -V curves.

## RESULTS AND DISCUSSION

#### Synthesis and Characterization

The synthetic routes to Tr-3PDI and FTr-3PDI are shown in **Scheme 2A**. Tr-3PDI was synthesized via Suzuki cross-coupling reaction between corresponding truxene boronic acid pinacol ester (Lin et al., 2018) and monobromo-PDI (Gao et al., 2017) using Pd(PPh3)<sup>4</sup> as the catalyst. FTr-3PDI was obtained with an excellent yield (90%) from Tr-3PDI by dissolving in chlorobenzene containing a catalytic amount of iodine and exposed to irradiation. Tr-3PDI and FTr-3PDI are characterized by <sup>1</sup>H NMR, <sup>13</sup>C NMR, and mass spectra (Figures S1–S6). The optimized geometries of Tr-3PDI and FTr-3PDI are simulated using density functional theoretical (DFT) calculations at the B3LYP/6-31G(d) level (**Scheme 2B**). Clearly, Tr-3PDI exhibits higher twisted structure with a large dihedral angle over 50◦ owing to the steric hindrance effect. After the oxidative ringfusion, each PDI moiety is tethered to truxene through benzene rings, resulting in an overall planarity structure because of the high rigidity and coplanarity of truxene core. Both acceptors are soluble in common organic solvents such as dichloromethane, chloroform, chlorobenzene, and ortho-dichlorobenzene at room temperature. The reason is that there are six hexyl chains on the truxene core, providing outstanding solubility for the resulting compounds.

The thermal properties of Tr-3PDI and FTr-3PDI were analyzed by thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC). As shown in Figure S7, both truxene-PDI acceptors have decomposition temperature with 5% weight loss above 400◦C. Moreover, there is no clear phase transition in DSC curves, which is indicative of the amorphous nature of Tr-3PDI and FTr-3PDI.

The UV-vis absorption spectra of the two acceptors in solutions and as thin films are shown in **Figure 1A**, Figure S8, and the relevant data are summarized in **Table 1**. Both Tr-3PDI and FTr-3PDI show two absorption bands with one in the short wavelength region of 300–400 nm and one in the longer wavelength region of 400–600 nm. The intense absorption in the short wavelength region is attributable to the large coplanar

TABLE 1 | Optical and electrochemical properties of Tr-3PDI and FTr-3PDI.


*<sup>a</sup>Calculated from Eopt <sup>g</sup>* = *1240/*λ *film onset eV; <sup>b</sup>Calculated from EHOMO* <sup>=</sup> *–e(Eonse ox - EFc*/*Fc*++*4.8) eV; <sup>c</sup>Calculated from ELUMO* = *–e(Eonset red -EFc*/*Fc*++*4.8) eV.*

core of truxene. FTr-3PDI shows little difference in normalized absorption spectra from the solution state to the film state, while the solid state Tr-3PDI has extended and redshifted absorption compared to the solution state. Notably, although the two compounds have very similar absorption maxima in both solution and solid state, FTr-3PDI shows considerably blueshifted absorption onset as compared to Tr-3PDI, which could be related to the reduced conformational disorder via ring-fusion and then weakens the intramolecular charge transfer between truxene and PDI moieties. In addition, FTr-3PDI exhibits higher absorption coefficient than Tr-3PDI (Figure S8). The optical band gaps (Eg) are calculated to be 1.96 eV for Tr-3PDI, and 2.23 eV for FTr-3PDI (**Table 1**).

The energy levels of the acceptors were determined by cyclic voltammetry (CV) experiments. The half-wave potential of Fc/Fc<sup>+</sup> was measured to be 0.36 V, and the energy levels of the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) were estimated from the onset oxidation (E onset ox ) and reduction (E onset red ) potentials by equations: EHOMO = –e(E onset ox -EFc/Fc++4.8) and ELUMO = –e(E onset red -EFc/Fc++4.8), respectively (Li et al., 1999). The CV curves are shown in **Figure 1B**, and the relevant data are listed in **Table 1**. The HOMO/LUMO levels are −6.09/−3.64 eV for Tr-3PDI and −6.11/−3.44 eV for FTr-3PDI, respectively. The slightly up-shifted LUMO level of FTr-3PDI will help to offer a higher open-circuit voltage (Voc), and the down-shifted HOMO level is favorable for hole transfer from excited acceptor phase to donor phase in BHJ OSCs (Duan et al., 2016a, 2017b, 2018; Jia et al., 2017).

## Photovoltaic Properties

The photovoltaic properties of Tr-3PDI and FTr-3PDI were evaluated in OSCs under AM1.5G illumination at 100 mW cm−<sup>2</sup> with a device structure of ITO/PEDOT:PSS/PTB7- Th:acceptor/PFN-Br/Ag (**Figure 2A**). PTB7-Th was used as the donor because of its strong optical absorption at long-wavelength region (Figure S9), which can achieve complementary absorption with our truxene-based acceptors (Zhang et al., 2015; McAfee et al., 2017; Welsh et al., 2018). The schematic energy diagram of individual components is displayed in **Figure 2B**, suggesting proper energy level alignment of each layer in the device. The devices were fully optimized in terms of host solvent, donor/acceptor weight ratios, active layer thickness, solvent additives, thermal annealing, and solvent annealing. The current density–voltage (J–V) curves of the optimized devices are shown in **Figure 2C**, and the photovoltaic parameters are summarized in **Table 1**. The device parameters under various conditions are collected in Tables S1–S8. The optimized device of Tr-3PDI afforded a PCE of 2.2% along with a Voc of 0.92 V, a short-circuit current density (Jsc) of 6.5 mA cm−<sup>2</sup> , and a fill factor (FF) of 0.37. The ring-fused acceptor FTr-3PDI produced a higher PCE of 3.8% along with a Voc of 1.02 V, a Jsc of 8.1 mA cm−<sup>2</sup> , and an FF of 0.46 (**Table 2**). The higher Voc of FTr-3PDI is consistent with the up-shifted LUMO level. The difference in Jsc of the solar cells can be explained by their external quantum efficiency (EQE) spectra (**Figure 2D**) (Duan et al., 2016b; Wu et al., 2016). The PTB7-Th:FTr-3PDI blend film show higher EQE than PTB7-Th:Tr-3PDI almost in the whole spectral range of 300–800 nm, which is because of the enhanced light absorption of the former and more efficient charge generation. The PTB7-Th:FTr-3PDI device also shows higher FF than the PTB7-Th:Tr-3PDI device, suggesting improved charge transport, reduced charge recombination, and more optimal phase separated morphology (Duan et al., 2011; Xie et al., 2012).

## Charge Transport and Recombination

The charge transport were investigated in single-carrier devices with a device structure of ITO/ZnO/ active layer /Ca/Al

TABLE 2 | Photovoltaic parameters of OSCs based on PTB7-Th and truxene-PDI acceptors under AM1.5G illumination at 100 mW cm−<sup>2</sup> .

based on PTB7-Th and truxene-PDI acceptors.


for electron only devices and ITO/PEDOT:PSS/ Active layer /MoO3/Ag for hole only devices, respectively. The electron and hole mobilities were acquired by fitting the J–V with space-charge-limited current (SCLC) model. The J–V curves of the devices for pure acceptors and blend films are shown in Figures S10, S11. As shown in **Table 3**, the FTr-3PDI pure film exhibits higher electron mobility (µe) of 2.4 × 10−<sup>6</sup> cm<sup>2</sup> V −1 s −1 than Tr-3PDI film (3.2 × 10−<sup>7</sup> cm<sup>2</sup> V −1 s −1 ), which support that the ring-fusion strategy is successful. As for blend films, the hole mobilities (µh) were estimated to be 1.2 × 10−<sup>3</sup> cm<sup>2</sup> V −1 s −1 for PTB7-Th:Tr-3PDI and 8.2 × 10−<sup>3</sup> cm<sup>2</sup> V −1 s −1 for PTB7-Th:FTr-3PDI, which are comparable with the value that obtained from PTB7-Th:fullerene devices (Huang et al., 2016). In contrast, theµ<sup>e</sup> of the blend films of PTB7- Th:truxene-PDI acceptors were measured to be 5.8 × 10−<sup>6</sup> cm<sup>2</sup> V −1 s −1 for PTB7-Th:Tr-3PDI and 1.3 × 10−<sup>5</sup> cm<sup>2</sup> V −1 s −1 for PTB7-Th:FTr-3PDI, which are more than two orders of magnitude lower than that of PTB7-Th:fullerene film (Lin et al., 2015). The low electron mobility and highly imbalanced µe/µ<sup>h</sup> seriously obstruct the charge transport and resulted in more bimolecular recombination, which in turn led to low FF and Jsc. For the solar cells with very imbalanced µe/µh, the device performance will be determined by the slower charge carrier, which is electron in these cases. The higher electron mobility in PTB7-Th:FTr-3PDI will thus result in better device performance.

To study the charge-recombination of these devices, we investigated the photocurrent (Jsc) as a function of light intensity (Pin, from 1 to 100 mW cm−<sup>2</sup> ), with the relevant characteristics plotted in **Figure 3A**. Generally, Jsc and Pin follow the relationship of Jsc∝P S in. If all free carriers are swept out and collected at the electrodes prior to recombination, the slope (S) should be equal to 1, while S<1 indicates some extent of bimolecular recombination (Kyaw et al., 2013). The S-values of the two devices are 0.89 for PTB7-Th:FTr-3PDI and 0.88 for PTB7-Th:Tr-3PDI, respectively, which indicates the existence of some extent of bimolecular recombination. The charge recombination mechanism of the truxene-PDIbased OSCs are also investigated by estimating the slope (α) of Voc vs. lnP (P is light intensity). In principle, the slope α approaching kBT/q implies that the device has only bimolecular recombination, and the slope α approaching 2kBT/q


TABLE 3 | The relevant parameters related to charge transport and recombination of PTB7-Th: truxene-PDI acceptor devices.

suggests that the monomolecular recombination or trap-assisted recombination dominates in OSCs (where T, kB, and q are the Kelvin temperature, Boltzmann constant, and elementary charge, respectively) (Koster et al., 2005; Lu et al., 2015). The α values for PTB7-Th:Tr-3PDI and PTB7-Th:FTr-3PDI are 2.1 and 1.9, respectively, which indicate considerable monomolecular recombination or trap-based recombination in these devices (**Figure 3B** and **Table 3**).

Transient photocurrent (TPC) and transient photovoltage (TPV) measurements were used to study the charge recombination dynamics and charge-extraction process in OSCs. From TPC analysis (**Figure 3C**), the charge-extraction time of the PTB7-Th:FTr-3PDI based device (0.15 µs) is slightly shorter than the PTB7-Th:Tr-3PDI based device (0.16 µs), suggesting increased charge extraction rate (Jin et al., 2016). From TPV analysis (**Figure 3D**), the charge carrier lifetime increased from 9.72 µs for the PTB7-Th:Tr-3PDI device to 15.31 µs for the PTB7-Th:FTr-3PDI device (**Table 3**), indicating reduced recombination loss for the PTB7-Th:FTr-3PDI device (Shuttle et al., 2008). The increased charge extraction rate and longer carrier lifetime thus explained the improved FF value and the higher PCE of the PTB7-Th:FTr-3PDI device.

#### Morphology Characterization

The morphology of the active layers was studied by atom force microscopy (AFM) and transmission electron microscopy (TEM). The AFM images and TEM images of the blend films are shown in **Figure 4**. The blend film of PTB7-Th:Tr-3PDI (**Figures 4A,C**) is homogeneous with a root-mean-square (RMS) surface roughness of 0.72 nm. The uniform film reveals intimately mixed blends without noteworthy phase

separation (Duan et al., 2017c; Wang J. Y. et al., 2017; Wen et al., 2018). With such a morphology, charge transport is impeded. The film of PTB7-Th:FTr-3PDI exhibits obvious phase separation with granulate features (**Figures 4B,D**), resulting in a relative coarse surface with a RMS surface roughness of 3.88 nm. The less phase-separated morphology of PTB7- Th:Tr-3PDI film could be a reason of the enhanced charge recombination and imbalanced hole/electron transport, which is in accordance with the analysis demonstrated above based on TPC, TPV, and charge carrier mobility measurements.

## CONCLUSION

In summary, two electron acceptors with a truxene core and three PDIs branches linked by carbon-carbon single bonds (Tr-3PDI) or adjoin benzene ring (FTr-3PDI) are designed and developed. The FTr-3PDI shows up-shifted energy levels, enhanced absorption, improved charge mobility, and more favorable morphology as compared to Tr-3PDI. These merits further lead to higher Voc, Jsc, and FF in resulting OSCs, respectively. The OSCs of PTB7-Th:FTr-3PDI blend shows a PCE of 3.8%, which is almost two times higher than that of PTB7-Th:Tr-3PDI blend. This work demonstrates a successful construction of star-shaped non-fullerene electron acceptor materials based on a truxene core and multiple PDI branches via ring-fusion to improve the performance of OSCs.

## AUTHOR CONTRIBUTIONS

KL, FH and YC: Designed experiments; KL, SW, ZW, QY, XL, and XJ: Carried out experiments; JJ, XfJ, and PL: Analyzed experimental results; KL and CD: Wrote the manuscript.

## FUNDING

This work was supported by the Ministry of Science and Technology (Nos. 2014CB643501, 2017YFA0206600). The research was also financially supported by the Recruitment Program of Global Youth Experts of China, the Natural Science Foundation of China (Nos. 21520102006 and 91633301).

## SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem. 2018.00328/full#supplementary-material

## REFERENCES


geometry enabling efficient organic solar cells. Adv. Mater. 28, 69–76. doi: 10.1002/adma.201504010


for polymer solar cells. Polymer 53, 5675–5683. doi: 10.1016/j.polymer.2012. 10.008


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Lin, Wang, Wang, Yin, Liu, Jia, Jia, Luo, Jiang, Duan, Huang and Cao. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

## Fluorination Induced Donor to Acceptor Transformation in A1–D–A2–D–A1-Type Photovoltaic Small Molecules

Ruimin Zhou1,2,3,4†, Benzheng Xia1†, Huan Li <sup>1</sup> , Zhen Wang<sup>1</sup> , Yang Yang<sup>1</sup> , Jianqi Zhang<sup>1</sup> , Bo W. Laursen<sup>4</sup> , Kun Lu<sup>1</sup> \* and Zhixiang Wei <sup>1</sup> \*

*<sup>1</sup> CAS Key Laboratory of Nanosystem and Hierarchical Fabrication, CAS Center for Excellence in Nanoscience, National Center for Nanoscience and Technology, Beijing, China, <sup>2</sup> Sino-Danish College, University of Chinese Academy of Sciences, Beijing, China, <sup>3</sup> Sino-Danish Center for Education and Research, Beijing, China, <sup>4</sup> Nano-Science Center & Department of Chemistry, University of Copenhagen, Copenhagen, Denmark*

#### Edited by:

*Chuanlang Zhan, Institute of Chemistry (CAS), China*

#### Reviewed by:

*Yingping Zou, Central South University, China Hui Huang, University of Chinese Academy of Sciences (UCAS), China*

#### \*Correspondence:

*Kun Lu lvk@nanoctr.cn Zhixiang Wei weizx@nanoctr.cn*

*†These authors share joint first authorship*

#### Specialty section:

*This article was submitted to Organic Chemistry, a section of the journal Frontiers in Chemistry*

Received: *12 June 2018* Accepted: *09 August 2018* Published: *30 August 2018*

#### Citation:

*Zhou R, Xia B, Li H, Wang Z, Yang Y, Zhang J, Laursen BW, Lu K and Wei Z (2018) Fluorination Induced Donor to Acceptor Transformation in A1–D–A2–D–A1-Type Photovoltaic Small Molecules. Front. Chem. 6:384. doi: 10.3389/fchem.2018.00384* With the development of diversity of non-fullerene acceptors, it is found that there is no clear boundary between electron donors and electron acceptors. Modulation of the electron donating and withdrawing properties of organic semiconductors is necessary for organic photovoltaics research. In this work, we designed and synthesized three A1–D–A2–D–A1-type (A represents acceptor unit and D represents donor unit) small molecules, named as M-0F, M-1F, and M-2F, respectively containing zero, one, and two fluorine atoms in the terminal acceptor segments (A1), respectively. Fluorination substitution was found to be able to downshift the HOMO and LUMO energy level, red-shift the absorption, and enhance the electron mobility. The M-0F exhibited the highest efficiency of 5.99% as a donor in fullerene-containing system and the lowest efficiency of 0.58% as an acceptor in fullerene-free system. While the M-2F performed the lowest efficiency of 0.97% as the donor and the highest efficiency of 2.65% as the acceptor. The electron-donating and electron-withdrawing property of M-1F are in-between that of M-0F and M-2F. Among the three molecules, the electron mobility is increased while the hole mobility is decreased with increasing fluorination. This work provides a typical example of tuning of the electron donating and withdrawing property without changes to the backbone of the conjugated molecules, which is important for further designing high performance solution processable small molecules.

Keywords: organic photovoltaics, small molecules, fluorination substitution, donors, acceptors

## INTRODUCTION

The bulk-heterojunction (BHJ) organic solar cells (OSCs) show a promising prospect for low-cost and renewable energy technology because of their unique advantages of light weight, easy-fabrication, and the capability to be fabricated into large area flexible devices (Forrest, 2004; Brabec et al., 2010; Kumar and Chand, 2012). In a typical BHJ organic photovoltaic (OPV) device, the heterojunction usually consists of a p-type electron donor and an n-type electron acceptor, which is the photoactive part for converting solar light to electricity. The p-type electron donor can be polymers or organic small molecules. The n-type electron acceptors include fullerene derivatives, such as [6, 6]-phenyl-C61/C71-butyric acid methyl ester (PC61BM/PC71BM) and non-fullerene electron acceptors. For a highly-efficient BHJ OPV, its active layer should possess the following features: (a) a broad absorption spectrum with a high extinction coefficient to utilize more solar photons; (b) a suitable molecular energy level alignment between the involved molecular orbitals on donor and acceptor to offer a sufficient driving force for efficient charge separation; (c) a bicontinuous network with nanoscale phase separation to facilitate exciton diffusion and charge separation; and (d) high charge mobility to facilitate charge transport (Scharber et al., 2006; Chen and Cao, 2009; Cheng et al., 2009; Beaujuge and Frechet, 2011; Henson et al., 2012; Li, 2012; Xu and Yu, 2014; Etxebarria et al., 2015; Zhang H. et al., 2015). Among these key features, the difference in LUMO energy of donor and acceptor is usually larger than 0.3 eV in fullerene system (Scharber and Sariciftci, 2013). However, with the development and application of non-fullerene electron acceptors, it is found that there is no clear boundary between electron donor and electron acceptor materials, and the difference of LUMO or HOMO energies in non-fullerene system could be very small. For instance, the polymer acceptor P-BNBPfBT has a LUMO of −3.6 eV, and PTB7-Th donor has a LUMO level of −3.42 eV. The device based on PTB7-Th/P-BNBP-fBT showed a power conversation efficiency (PCE) as high as 6.26%, even though the difference in LUMO levels is only 0.18 eV (Long et al., 2016). However, polymer materials normally have a large polydispersity in molecular weight increasing the complexity of the materials. Thus, when it comes to fine tuning of the molecular levels, the use of solution processable small molecules is important to understand the relationship between molecular structures and their electron donating/withdrawing properties (Zhu et al., 2016; Wu et al., 2017).

At present, the small molecules could be designed as donor or acceptors based on different molecular architectures. For small molecule donors the conjugated skeleton is most often relatively planar. For example, oligothiophene-based small molecule donor DRCN7T exhibited an impressive optimized PCE of 9.30% using PC71BM acceptor (Zhang Q. et al., 2015). Our group synthesized

the A1–D–A2–D–A1 structure electron-donating small molecule with a PCE of more than 9% with enhanced molecular planarity and crystallinity (Yuan et al., 2016). On the other hand, currently non-fullerene acceptors are mainly concentrated on fused ring acceptors (Cheng et al., 2018; Hou et al., 2018). The acceptors based on perylene diimide (PDI) or naphthalene diimide (NDI) have shown relatively good performance in PSCs (Facchetti, 2013; Zhang et al., 2013; Hartnett et al., 2014; Li et al., 2014; Lin et al., 2014; Jung et al., 2015; Liu et al., 2015; Sun et al., 2015; Yang et al., 2016; Lei et al., 2018). Most of the highperformance PDI and NDI-based fullerene-free acceptors have twisted backbones to decrease the planarity, self-aggregation, and crystalline domains of simple rylene diimides. Besides the rylene-based fullerene-free electron acceptors, electron-donating extended fused rings, e.g., indacenodithiophene (IDT) and indacenodithieno[3,2-b]thiophene (IDTT), were widely used in small molecule acceptors because their LUMO levels can be readily tuned by flanking with different electron-withdrawing groups and the steric effect of tetrahexylphenyl substituents on the coplanar backbone can reduce the intermolecular interactions while weaken the stacking of donor units and promote the stacking of acceptor units (Lin et al., 2015; Wu et al., 2015; Lin and Zhan, 2016; Huang et al., 2017; Yu et al., 2017). Another example was present by Hou et al., the steric hindrance caused by side chains could convert two isomers to donors and acceptors separately (Liu et al., 2018). In this case, molecules with weak π-π stacking structure is more likely to work as acceptor materials. Although recent progress showed that small molecules based on the same backbone can be changed from donor to

TABLE 1 | Optical and electrochemical data of compounds M1-0F, M1-1F, and M1-2F.


accepters by changing the side chain hindrance, it is still unclear whether one can slightly tune the energy levels of planer small molecules and charge mobility to realize the transformation from donor to acceptor.

Herein, we designed and synthesized three novel A1–D– A2–D–A1-type small molecules with dialkoxyphenyldithiophene (PDT) as D unit, difluorinated benzothiadiazole (2FBT) as A2 unit and strong electron-withdraw group 1, 1-dicyanomethylene-3-indanone (IC) substituted with 0–2 fluorine substituents as end-capped A1, namely M-0F, M-1F, and M-2F. All three molecules exhibit good planarity. The molecular structures are shown in **Scheme 1**. Fluorine is the most electronegative element with relatively small van der Waals radius. It can serve as an electron-withdrawing group without introducing undesirable steric hindrance. The introduction of fluorine onto the conjugated backbone of polymers or small molecules for BHJ solar cells could lower the LUMO and HOMO level (Price et al., 2011; Zhou et al., 2011; Zhang et al., 2017; Zhao et al., 2017). Fluorination IC also promotes intermolecular interactions

through forming non-covalent F–S and F–H bonds, which can be favorable for charge transport (Sakamoto et al., 2001; Lei et al., 2014; Kim et al., 2015). The target molecules show red-shifted absorption spectrum and their LUMO and HOMO energies decreased with increasing fluorination. As donor, M-0F exhibited the highest PCE of 5.99%, M-1F performed a PCE of 2.60%, and M-2F showed the lowest PCE of 0.97% with PC71BM as acceptor. On the other hand, as acceptor, M-0F showed the lowest PCE of 0.58%, M-1F performed medium PCE of 1.85%, and M-2F achieved the highest PCE of 2.65% with polymer PBDB-T as donor. The energy level difference affects the performance of the molecule to a certain extent, M-0F exhibited the same LUMO level as PBDB-T, and cannot provide enough driving force for charge transfer and separation, but for M-2F, its lower LUMO is relatively better-match with PBDB-T although the LUMO difference is small. On the other hand, the hole and electron mobility also show that transition from hole transport to electron transport can be obtained via fluorination.

## RESULTS AND DISCUSSION

#### Molecular Synthesis and Properties

All three target molecules M-0F, M-1F, M-2F were synthesized according to the route shown in **Scheme 1**. To ensure the solubility of these three materials, the long branched side chain octyldodecyloxy was selected. Compound PDT2FBT-CHO was obtained through a Vilsmeier reaction from O-1. The three target molecules were synthesized by a Knoevenagel reaction. The synthetic routes, purification methods, and the nuclear magnetic

TABLE 2 | Device performance of M-0F, M-1F, and M-2F with PC71BM as acceptors.


resonance spectroscopy data are provided in the Supporting Information.

**Figures 1A,B** show the normalized UV–Vis absorption spectra of M-0F, M-1F, and M-2F measured in chloroform solution and in thin solid films. Detailed parameters are listed in **Table 1**. In solution, the three molecules show similar absorption spectra where the introduction of fluorine results in a red-shift of the absorption spectrum with the max absorption peak shifting from 623 to 630 and 638 nm. In thin films, all three molecules show more red-shifted and broader absorption spectra. The absorption peaks of these three compounds red-shift gradually from 697 to 705 nm and then to 712 nm in thin solid films. The strong shoulder peaks show that the three molecules have good π-π stacking in the film (Badgujar et al., 2016). The optical band gaps of M-0F, M-1F, and M-2F are calculated to be 1.64, 1.59, and 1.60 eV from the absorption edge, respectively.

The cyclic voltammetry (CV) was used to evaluate the electrochemical properties of the three small molecules and the results are shown in **Figures 1C,D** with reversible reduction waves and quasi-reversible oxidation waves. The HOMO and LUMO energy levels are calculated from the onset oxidation and reduction potentials, assuming the absolute energy level of FeCp+/<sup>0</sup> 2 to be 4.8 eV below vacuum. The equation of HOMO energy levels is EHOMO = –e (Eox + 4.8–E1/2, (Fc/Fc+)) and the equation of the LUMO energy levels is ELUMO = –e (ERED + 4.8–E1/2, (Fc/Fc+)) (Li et al., 1999). Due to the strong electron withdraw ability of fluorine atom, the HOMO and LUMO of both M-1F, M-2F downshift. The HOMO levels of M-0F, M-1F, and M-2F are estimated to be −5.45, −5.50, and −5.54 eV and LUMO levels are −3.52, −3.54 to −3.55 eV (**Figure 1D**).

#### Photovoltaic Device Characterization

To investigate the photovoltaic behaviors of the three molecules as the electron donor and acceptor in OPV devices, we selected two materials, PC71BM as an acceptor (Wienk et al., 2003; Thompson and Fréchet, 2008) and PBDB-T as a donor polymer (Zhao et al., 2016) to blend with M-0F, M-1F, M-2F to fabricated conventional device structure: ITO/PEDOT:PSS/Donor:Acceptor/Ca/Al. We comparatively studied the photovoltaic performance of M-0F, M-1F, and M-2F as donor materials. The devices were made of blend of small molecule donor and PC71BM with chloroform as solvent by spin-coating. The photovoltaic properties of the devices were characterized under illumination of simulated solar light, AM1.5G (100 mWcm−<sup>2</sup> ). The optimized current density vs. voltage (J-V) curves and their corresponding external quantum efficiency (EQE) curves are displayed in **Figure 2**. The corresponding photovoltaic performance data is summarized in **Table 2**. The photovoltaic performance data of other condition is summarized in **Tables S1**, **S2**. The device based on M-0F: PC71BM exhibited a high VOC of 1.01 V, a JSC of 9.54 mAcm−<sup>2</sup> and a high FF of 62.41%, resulting in the highest optimal PCE of 5.99%. For device M-1F: PC71BM, PCE was 2.60% with VOC of 0.91 V, JSC of 5.14 mA/cm<sup>2</sup> and FF 55.71%. However, for device M-2F: PC71BM, the PCE was only 0.97% with a VOC of 0.93 V, a JSC of 1.80 mAcm−<sup>2</sup> and a FF of 57.65%. It can be concluded that as an electron donor in blends with PC71BM, M-0F displays a superior photovoltaic performance compared

M-2F/PC71BM blend films.

to M-1F while M-1F has better photovoltaic performance than M-2F. The results show that the device performance based the three molecules as donors decreases with increased number of fluorine atoms attached to the terminal acceptor units (A1).

We also studied photovoltaic performance of M-0F, M-1F, and M-2F as electron acceptor material with PBDB-T as donor material. The blend of small molecule acceptor and PBDB-T was made with chlorobenzene as solvent by spin-coating. It was found that the PCE of a M-0F:PBDB-T device was only 0.58% with low JSC and FF; the PCE of a device based on M-1F:PBDB-T was 1.85%; while the device based on M-2F:PBDB-T displayed a PCE as high as 2.65%, with a VOC of 0.95, a JSC of 6.25 mAcm−<sup>2</sup> , and a FF of 44.54%. It can be concluded that, as electron acceptors after blending with PBDB-T, M-2F displays a superior photovoltaic performance compared to M-0F and M-1F. This means, that as acceptors and donors the series of small molecules display opposite trend in device performances. With increasing fluorination, the device performance as donor decreases while improves as acceptor.

#### Photoluminescence Quenching Effects

In order to investigate the donor property of M-0F, M-1F, and M-2F, the fluorescence spectra were measured in the range of 680–900 nm with exciting wavelength at 650 nm. As shown in **Figure S1A**, the photoluminescence (PL) of M-0F, M-1F can be partly quenched by PC71BM, whereas the PL of M-2F almost can't be quenched by PC71BM, indicating that the exciton dissociation in the blend of M-0F/PC71BM, M-1F/PC71BM should be more efficient than that in the blend of M-2F/PC71BM. For studying acceptor property of these three small molecules, the exciting wavelength at 740 nm was chosen as shown in **Figure S1B**. From the film of M-0F/ PBDB-T, M-1F/ PBDB-T to M-2F/ PBDB-T, the PL quenching effect is gradually stronger, indicating that the exciton dissociation is more efficient gradually. These results are in consistent with device performance.

#### Film Morphology and Microstructure

Using atomic force microscopy (AFM) and transmission electronic microscopy (TEM), the phase separation morphology of the blends for the two systems were investigated. As shown in **Figure 3**, the AFM images are consistent with the TEM images. For the PC71BM system, it can be seen that M-0F, M-1F, and M-2F have good compatibility with PC71BM and good phase separation. The surface roughness is 8.317, 4.236, and 4.617 nm for blend films M-0F:PC71BM, M-1F:PC71BM, and

corresponding 2D GIWAXS patterns.

M-2F:PC71BM, respectively. The M-0F:PC71BM blend has the larger size of aggregation phase region, and the proper domain size for efficient exciton diffusion and dissociation contributing to high JSC and FF. For M-1F:PC71BM and M-2F:PC71BM blend film, the aggregation size is so small resulting in low JSC and FF.

For the non-fullerene system shown in **Figure 4**, the surface roughness of PBDB-T:M-0F, PBDB-T:M-1F, and PBDB-T:M-2F



blend film are 11.034, 6.672, and 4.356 nm. The surface of PBDB-T:M-2F is smoother. From TEM image, we can see that PBDB-T:M-2F formed a uniform nanoscale phase separation with nanofiber and continuous interpenetrating network structure, and the phase separation size is suitable, which facilitates the exciton diffusion and charge transfer and improve JSC and FF. However, For the PBDB-T:M-0F and PBDB-T:M-1F blends large domain size can be seen, this is not conducive to the charge transfer, which cause a negative impact on the JSC and FF.

In order to further study the molecular stacking and crystallization properties of the active layer, grazing incidence wide-angle X-ray scattering (GIWAXS) on the neat films and blend films was employed. The results of pristine M-0F, M-1F, M-2F, and PBDB-T films are shown in **Figure S2**. In-plane direction M-0F, M-1F, and M-2F all show π-π stacking peaks at 1.72 Å−<sup>1</sup> (d-spacing: 3.65 Å), 1.77Å−<sup>1</sup> (d-spacing: 3.55 Å), 1.78 Å −1 (d-spacing: 3.53 Å), which indicates the higher crystallinity of M-1F, and M-2F compared with M-0F in the neat films. **Figure 5** shows the two-dimensional GIWAXS patterns and the one-dimensional GIWAXS cuts along in-plane and out-of-plane directions of M-0F:PC71BM, M-1F:PC71BM, and M-2F:PC71BM blended films. As can be seen from **Figure 5**, M-0F, M-1F, and M-2F all show very obvious diffraction peaks of (100), (200), (300), and (010), indicating the good crystallinity M-0F, M-1F, and M-2F with the orderly aggregation state structure. Besides, these corresponding peaks have almost the same location, which indicate the same d-spacing. The positions of (100) peak is ∼0.298 Å−<sup>1</sup> correspongding to a d-spacing of ∼21.08 Å. The positions of (010) q is ∼1.869 Å−<sup>1</sup> , so the d-spacing is ∼3.36 Å. The π-π stacking of M-1F is stronger while the hole mobility is not the highest compared with M-0F, M-2F, indicating that the performance of the donor and acceptor has been changed.

**Figure 6** shows the GIWAXS of PBDB-T:M-0F, PBDB-T:M-1F, and PBDB-T:M-2F blend films. Similar with PC71BM system, M-0F, M-1F, and M-2F all show very obvious diffraction peaks of (100) (200), (300), and (010), these corresponding peaks have almost the same location. The positions of (100) q is∼ 0.30 Å−<sup>1</sup> , so the d-spacing is ∼20.94 Å. The positions of (010) q is ∼1.83 Å −1 , so the d-spacing is ∼3.43 Å. The GIWAXS results show the good crystallization of M-0F, M-1F, and M-2F.

#### Hole and Electron Mobility

In order to study the origin of the observed differences in photovoltaic properties of the three new materials, the hole and electron mobility of the three pure materials and blend films at the best performance condition were tested using space-limited charge (SCLC) method (Blom et al., 2005). The hole and electron mobility curves measured by the SCLC method are shown in Supplementary Information **Figure S3**. The data of hole and electron mobility under different conditions is summarized in **Table 3**. The hole and electron mobility for pristine M-0F are 2.49 × 10−<sup>4</sup> and 1.01 × 10−<sup>6</sup> cm<sup>2</sup> V −1 S −1 , respectively. For pristine M-1F, the hole and electron mobility are 1.78 × 10−<sup>5</sup> and 3.03 × 10−<sup>6</sup> cm<sup>2</sup> V −1 S −1 . For pristine M-2F, the hole and electron mobility are 6.42 × 10−<sup>6</sup> and 1.26 × 10−<sup>5</sup> cm<sup>2</sup> V −1 S −1 . This indicates the transition from hole transport to electron transport. Among M-1F and M-2F, the electron mobility of M-2F is higher than its hole mobility, so M-2F has better n-type semiconductor property than M-1F. At the corresponding best performance condition, from M-0F:PC71BM, M-1F:PC71BM to M-2F:PC71BM, both the hole and electron mobility decrease. From PBDB-T:M-0F, PBDB-T:M-1F to PBDB-T:M-2F, both the hole and electron mobility increase, which is consistent with photovoltaic performance results.

On the basis of above observations, we can conclude that M-0F has good donor property with better phase separation and higher hole mobility in fullerene system. while M-2F has the best acceptor properties with nanofiber interpenetrating network morphology and higher electron mobility in nonfullerene system.

#### CONCLUSION

In conclusion, we designed and synthesized three novel A1- D-A2-D-A1 small molecules (M-0F, M-1F, and M-2F) with the strong electron-withdraw group 1,1-dicyanomethylene-3 indanone (IC) substituted with varying number of fluorine atoms. With increasing number of fluorine atoms, the absorption spectra are red-shifted, and both the LUMO and HOMO energies are decreased. M-0F exhibited excellent electrondonating properties while M-2F showed excellent electronaccepting properties. Our work further proves that the definition of donor and acceptor is without clear boundaries, which offer wide potential for molecular design. The regulation of energy levels and carrier mobility is one of the effective ways to achieve this transformation. Also, the boundary between donor materials and acceptor materials is worth exploring for understanding the deep mechanism of exciton dissociation and charge transfer in BHJ active layers.

#### AUTHOR CONTRIBUTIONS

RZ designed the project, finished the synthesis, characterization, wrote the first draft of the manuscript; BX designed the project, finished the device, and characterization of photovoltaic performance. BL, KL, and ZW guided the project and helped to revise the manuscript. The other authors gave the contribution in synthesis or data analysis.

## ACKNOWLEDGMENTS

We acknowledge the Ministry of Science and Technology of China (Nos. 2016YFA0200700 and 2016YFF0203803), the National Natural Science Foundation of China (Grant

#### REFERENCES


Nos. 21474022, 21125420, 21603044, 51673049, and 21603044), the Beijing Nova Program, the Youth Innovation Promotion Association CAS, and the Chinese Academy of Sciences.

#### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem. 2018.00384/full#supplementary-material


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Zhou, Xia, Li, Wang, Yang, Zhang, Laursen, Lu and Wei. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

## Constructing Desired Vertical Component Distribution Within a PBDB-T:ITIC-M Photoactive Layer via Fine-Tuning the Surface Free Energy of a Titanium Chelate Cathode Buffer Layer

Yiming Bai 1,2, Bo Yang<sup>1</sup> , Xiaohan Chen<sup>1</sup> , Fuzhi Wang1,2, Tasawar Hayat 3,4, Ahmed Alsaedi <sup>4</sup> and Zhan'ao Tan1,2 \*

#### Edited by:

*Chuanlang Zhan, Institute of Chemistry (CAS), China*

#### Reviewed by:

*Xiaoyan Du, i-MEET Institute Materials for Electronics and Energy Technology, Germany Ziyi Ge, Ningbo Institute of Industrial Technology (CAS), China Zhixiang Wei, National Center for Nanoscience and Technology (CAS), China*

> \*Correspondence: *Zhan'ao Tan tanzhanao@ncepu.edu.cn*

#### Specialty section:

*This article was submitted to Organic Chemistry, a section of the journal Frontiers in Chemistry*

Received: *26 April 2018* Accepted: *25 June 2018* Published: *20 August 2018*

#### Citation:

*Bai Y, Yang B, Chen X, Wang F, Hayat T, Alsaedi A and Tan Z (2018) Constructing Desired Vertical Component Distribution Within a PBDB-T:ITIC-M Photoactive Layer via Fine-Tuning the Surface Free Energy of a Titanium Chelate Cathode Buffer Layer. Front. Chem. 6:292. doi: 10.3389/fchem.2018.00292*

*<sup>1</sup> State Key Laboratory of Alternate Electrical Power System with Renewable Energy Sources, North China Electric Power University, Beijing, China, <sup>2</sup> Beijing Key Laboratory of Energy Safety and Clean Utilization, North China Electric Power University, Beijing, China, <sup>3</sup> Department of Mathematics, Quiad-I-Azam University, Islamabad, Pakistan, <sup>4</sup> NAAM Research Group, Faculty of Science, King Abdulaziz University, Jeddah, Saudi Arabia*

Rationally controlling the vertical component distribution within a photoactive layer is crucial for efficient polymer solar cells (PSCs). Herein, fine-tuning the surface free energy (SFE) of the titanium(IV) oxide bis(2,4-pentanedionate) (TOPD) cathode buffer layer is proposed to achieve a desired perpendicular component distribution for the PBDB-T:ITIC-M photoactive layer. The Owens-Wendt method is adopted to precisely calculate the SFE of TOPD film jointly based on the water contact angle and the diiodomethane contact angle. We find that the SFE of TOPD film increases as the annealing temperature rises, and the subtle SFE change causes the profound vertical component distribution within the bulk region of PBDB-T:ITIC-M. The results of secondary-ion mass spectroscopy visibly demonstrate that the TOPD film with an SFE of 48.71 mJ/cm<sup>2</sup> , which is very close to that of the ITIC film (43.98 mJ/cm<sup>2</sup> ), tends to form desired vertical component distribution. Consequently, compared with conventional bulk heterojunction devices, the power conversion efficiency increases from 9.00 to 10.20% benefiting from the short circuit current density increase from 14.76 to 16.88 mA/cm<sup>2</sup> . Our findings confirm that the SFE adjustment is an effective way of constructing the desired vertical component distribution and therefore achieving high-efficiency PSCs.

Keywords: polymer solar cells, vertical component distribution, surface free energy, cathode buffer layer, annealing temperature

## INTRODUCTION

Stimulated by the need for a clean renewable energy source, there has been considerable interest in exploring polymer solar cells (PSCs) due to their unique properties of low cost, light weight, and flexibility (Krebs et al., 2010; Li G. et al., 2012; Zhao et al., 2017). The state-of-the-art device configuration is the sandwich bulk heterojunction (BHJ), blending the conjugated polymer donor compactly with a fullerene or fullerene-free acceptor (Ouyang et al., 2015; Jiang et al., 2017; Peng et al., 2018). Very recently, the best power conversion efficiency (PCE) of single-junction PSCs based on the fullerene acceptor has exceeded 11.7% (Zhao et al., 2016), and that of fullerene-free PSCs has reached up to 14%. This can be attributed to the development of new electron donors and the matching acceptors, device structures and novel interfacial layers (Xiao et al., 2017).

The BHJ photoactive layer, as the main functional layer for light absorption, exciton generation, dissociation, and transportation, is commonly fabricated by spin-coating the mixed solution of electron donors and acceptors (Heriot and Jones, 2005; Lu et al., 2013; Xie et al., 2014; Bin et al., 2016). The mixed components will form a vertical phase separated photoactive layer during the film-drying process, which ensures a large interfacial area for efficient exciton dissociation and facilitates the charge transportation and selective collection via the formation of bi-continuous interpenetrating networks (Qiu et al., 2008; Xu et al., 2010; Meier et al., 2011). Hence, an indepth understanding about the vertical component distribution within the photoactive layer is absolutely imperative for realizing efficient PSCs. However, its mechanism is complex and still vague, which involves thermodynamics, dynamics, free-surface, and interface effects during the blend formation process (Kim et al., 2017).

A number of studies on fullerene PSCs have illustrated that the perpendicular component distribution of photoactive blends is greatly influenced by the processing conditions, such as solvent soaking, solvent flush treatment, and solvent additives (Li C. Z. et al., 2012; Heo et al., 2014; Van Franeker et al., 2015). Li et al. introduced a mixed solvent-soaking approach to obtain an interpenetrating network composed of highly crystalline regioregular poly(3-hexylthiophene) (P3HT) and [6,6]-phenyl-C61-butyric acid methyl ester (PC61BM) nano-aggregates (Hui et al., 2011). The 2-chlorophenol flush treatment is also a simple and feasible way to acquire an optimal vertical composition profile of photoactive blends, leading to an increased PCE from 6.18 to 10.15% for inverted PSCs based on poly[4,8-bis(5-(2-ethylhexyl)thiophen-2-yl)benzo[1,2 b:4,5-b0]dithiophenealt-3-fluorothieno[3,4-b]thiophene-2-

carboxylate] (PTB7-Th):[6,6]-phenyl-C71-butyric acid methyl ester (PC71BM) blends (Wang et al., 2016). The solvent additive 1,8-diiodooctane (DIO) with high boiling point, using the solubility of the different components to change and affect the donor and acceptor phase distributions, can selectively dissolve PCBM and facilitate long-range diffusion of PCBM to form a bicontinuous pathway for electron and hole transportation at the latter stage of the film-drying process (Xiao et al., 2014). However, few investigations have been conducted to elucidate the influence of processing parameters on the perpendicular component distribution of fullerene-free photoactive blends though it well exists, and this is an area warranting urgent exploration (Yan et al., 2017).

Evidence also shows that the free energy of electron donors and acceptors, and the substrate surface onto which they are deposited has a direct impact on the perpendicular component distribution (Jones et al., 1989; Jasieniak et al., 2016). Kim et al. found that P3HTs with hydroxyl-, ethyl-, perfluoro-, and bromoend groups have different surface free energies (SFEs), and the vertical stratification in their blends with PCBM can be tuned as the surface energy difference between electron donors and acceptors (Kim et al., 2010). It is well-known that the system is the most stable if and only if the systematic energy is the minimum. Jones put forward that any surface energy difference between the pure components allows the photoactive blends to minimize its total free energy by increasing the surface concentration of the low-energy component (Jones et al., 1989). In contrast, Bjöström and Tillack believed that the variation of the substrate SFE affects the perpendicular component distribution within the bulk region rather than the air surface region (Björström et al., 2005; Tillack et al., 2011). Germack found that for the similar P3HT:PCBM blends (the surface energies for polymers P3HT and PCBM are about 23 and 45 mJ/m<sup>2</sup> , respectively), their vertical component distributions are influenced by the substrates (Germack et al., 2009). Namely, if the P3HT:PCBM blends are deposited on poly(3,4 ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS) with a surface energy of 45 mJ/m<sup>2</sup> , the PCBM is enriched near the substrate surface, while the P3HT is enriched near the substrate interface as well as the free surface if the P3HT:PCBM blends are deposited on a poly(thienothiophene):Nafion substrate with a surface energy of 23 mJ/m<sup>2</sup> (Germack et al., 2010). The aforementioned reports are helpful for constructing ideal vertical morphology and further achieving efficient PSCs. However, most of the studies primarily focus on fullerene PSCs, and few works are available on fullerene-free PSCs, especially on finding modulation approaches to realize the desired vertical phase separation.

Hence, the present work is aimed at exploring the influence of processing parameters on the SFE of titanium(IV) oxide bis (2,4-pentanedionate) (TOPD) cathode buffer layer and further elucidating the impact of SFE on the vertical component distribution within the (poly[(2,6-(4,8-bis(5-(2-ethylhexyl)thiophen -2-yl)benzo[1,2-b:4,5-b′ ]dithiophene)-co-(1,3-bis(5-thiophene-2-yl)-5,7-bis(2-ethylhexyl)benzo[1,2-c:4,5-c]dithiophene-4,8 dione)] PBDB-T:ITIC-M (3,9-bis((Z)-1-(6-(dicyanomethylene)- 2-methyl-5,6-dihydro-6H-cyclopenta[b]thiophen-6-one-5-yl) ethylene)-5,5,11,11-tetrakis(4-hexylphenyl)dithieno[2,3-d:2′ ,3′ d ′ ]-sindaceno[1,2-b:5,6-b′ ]dithiophene) photoactive layer. Herein, the SFE of TOPD films annealed at different temperatures was quantified jointly from the water contact angle and the diiodomethane contact angle according to the Owens-Wendt (OW) method (Owens and Wendt, 1969). We found that the SFE of TOPD changes as the annealing temperature increases, and the PBDB-T:ITIC-M photoactive layer with the desired vertical component distribution is obtained via fine controlling the SFE of the TOPD layer, leading to a PCE high up to 10.20% for inverted PSCs. The results of time-offlight secondary-ion mass spectroscopy (TOF-SIMS) visually present the optimized vertical concentration distribution, and the space-charge-limited current (SCLC) method elucidates that the rational vertical component distribution guarantees fully exciton dissociation and facilitates charge transportation.

## EXPERIMENTAL SECTION

materials involved in the i-PSCs.

## Materials and Instrumentations

Patterned indium tin oxide (ITO) glass with a sheet resistance of 10 /sq was purchased from CSG HOLDING Co., Ltd. (Shenzhen, China). Both MoO<sup>3</sup> (purity > 99.0%) and DIO (purity > 98.0%) were purchased from Sigma Aldrich (St. Louis, MO). The TOPD was purchased from Alfa Aesar (Shanghai, China). PBDB-T and ITIC-M were purchased from Solarmer Materials Inc. (Beijing, China), and their molecular structures are displayed in **Figure 1A**. All these commercially available materials were used as received without further purification.

Ultraviolet–visible (UV-Vis) absorption spectra for TOPD films before and after annealing were measured by a Hitachi U-3010 UV-Vis spectrophotometer. X-ray diffraction (XRD) patterns were recorded on a Siemens D5005 diffractometer using CuKα radiation at 40 kV and 20 mA. The contact angle images were investigated using a profilometer of Dektak XT (Bruker) under ambient conditions. The perpendicular component distribution was analyzed using a TOF-SIMS from ION-TOF GmbH. The Ambios Technology XP-2 surface profilometer was employed to evaluate the film thicknesses involved in the device.

## Device Design, Fabrication, and Characterization

A well-designed device architecture coupled with a desired perpendicular component distribution in photoactive blends is anticipated to prepare efficient PSCs. **Figure 1B** plots the inverted device structure, where PBDB-T:ITIC-M blend is sandwiched between the TOPD-coated ITO cathode and the high work function (WF) MoO<sup>3</sup> anode. **Figure 1C** demonstrates the energy levels of the materials involved in the devices. The PBDB-T:ITIC-M blend absorbs incident photons and produces excitons. The excitons diffuse toward and dissociate at the PBDB-T:ITIC-M interfaces to yield free electrons and holes. The free electrons can breezily transport from the active layer to the cathode through the TOPD layer due to the similar lowest unoccupied

contact angle images of (D) pure PBDB-T, (E) pure ITIC-M, and (F) PBDB-T:ITIC-M on the TOPD modified substrate.

molecular orbital (LUMO) of TOPD (−3.86 eV) and ITIC-M (−3.98 eV). Meanwhile, the high up to 1.25 eV hole barrier between the highest occupied molecular orbital (HOMO) level of TOPD (−6.81 eV) and ITIC-M (−5.56 eV) can effectively block the transportation of the hole from PBDB-T to ITO, reducing carrier recombination losses at the interface (Bai et al., 2018). Furthermore, MoO<sup>3</sup> with WF of 5.30 eV facilitates hole transfer from the active layer to the Al anode (Bai et al., 2017).

All ITO substrates were successively ultrasonically cleaned twice by detergent, deionized water, acetone, and isopropanol. For the regular control device with the architecture of ITO/PEDOT:PSS/PBDB-T:ITIC-M/Ca/Al, the pre-cleaned and dried ITO substrates were treated under UV-ozone (UVO) exposure for 15 min to improve its surface smoothness and WF (Bai et al., 2018). Then a 30 nm PEDOT:PSS layer was deposited by spin-coating its aqueous solution at 2,000 rpm for 45 s, and baked at 150◦C for 10 min in air. On the other hand, for the inverted PSCs with the structure of ITO/TOPD/PBDB-T:ITIC-M/Ca/Al, the clean ITO substrates without UVO treatment were transferred into the nitrogen-filled glovebox for the following process. After that, the isopropanol solution with optimized

concentrations of TOPD (1,000 rpm/min, 12 mg/mL) were spincoated on ITO, and finally annealed at different temperatures (80–120◦C) for 5 min, and the 14-nm TOPD films with excellent robustness were obtained.

The photosensitive layer was prepared by spin-coating the PBDB-T:ITIC-M chlorobenzene solution (1:1 weight ratio, polymer concentration of 10 mg/mL) with 5% volume ratio of DIO additive on the ITO/PEDOT:PSS and ITO/TOPD substrates at optimized 1,900 rpm for 60 s. Subsequently, the samples were annealed at 100◦C for 10 min to obtain the PBDB-T:ITIC-M layer with the thickness of ∼100 nm (Bai et al., 2017). Finally, the anode of Ca(10 nm)/Al(100 nm) for the control devices or MoO3(24 nm)/Al(100 nm) for the inverted PSCs was thermally deposited on the active layer under a base pressure of 5 × 10−<sup>5</sup> Pa (Luo et al., 2018).

Device characterization of current density–voltage (J-V) performance was conducted in a nitrogen-filled glovebox using a Keithley 2400 Source Measure Unit under simulated AM1.5G solar irradiation with the light intensity of 100 mW/cm<sup>2</sup> (from SAN-EI LTD, AAA grade). The incident photon to electron conversion efficiency (IPCE) was measured using the QE-R system (Enli Tech., Kaohsiung, Taiwan) in air at room temperature. The intensity of each wavelength both in J-V and IPCE was calibrated with the standard single crystalline silicon photovoltaic device purchased from the national renewable energy laboratory. Electron mobility was

FIGURE 5 | (A) S

− 2 and CN

and the TOPD layer is annealed at 90◦C.

measured employing the SCLC method for devices with the structure of ITO/Al/TOPD/Al, ITO/Al/PBDB-T:ITIC-M/Al, and ITO/TOPD/PBDB-T:ITIC-M/Al. The results are plotted as ln(JL<sup>3</sup> /V<sup>2</sup> ) vs. (V/L) 0.5. Electron mobility was calculated from the intercept of the corresponding lines on the axis of ln(JL<sup>3</sup> /V<sup>2</sup> ) (Malliaras et al., 1998).

#### RESULTS AND DISCUSSION

## OW Method for Determination the SFE of Solid Films

The measurement of the contact angle of sessile drops deposited on different film surfaces is one of the powerful approaches to evaluate the film SFEs and further modulate the vertical concentration distribution of the electron donor and acceptor (Clark et al., 2013). The OW method is the most common approach for polymeric materials so far, in which water and diiodomethane are used. According to OW principal assumptions, the SFE includes dispersion and polar two components. The former represents the dispersion interaction occurring on an interface and the latter is a sum of polar, hydrogen, inductive, and acid–base interactions. The SFE is evaluated with the OW method using the following set of equations (Zenkiewicz, ˙ 2007):

2

$$(\gamma\_{\rm Sd}\gamma\_{\rm Wd})^{\frac{1}{2}} + (\gamma\_{\rm Sp}\gamma\_{\rm Wp})^{\frac{1}{2}} = 0.5\gamma\_{\rm W}(1 + \cos\Theta\_W) \tag{1}$$

and CN<sup>−</sup> in the PBDTBDD:ITIC-M blend films,

<sup>−</sup> intensity as a function of sputter time. (B) Three-dimensional concentration profiles of S<sup>−</sup>

FIGURE 6 | (A) *J-V* curves, (B) PCE with error bars, and (C) IPCE spectra for conventional devices and inverted devices without and with TOPD. The *J-V* curves are measured under the illumination of AM 1.5G at 100 mW/cm<sup>2</sup> .

TABLE 1 | Photovoltaic parameters (averaged over 12 individual devices) of the conventional devices and the i-PSCs with and without TOPD under 100 mW/cm<sup>2</sup> .


*<sup>a</sup>Series resistance (Rs) for PSCs in the dark is obtained at 1 V.*

$$(\chi\_{\rm Sd}\chi\_{\rm Dd})^{\frac{1}{2}} + (\chi\_{\rm Sp}\chi\_{\rm Dp})^{\frac{1}{2}} = 0.5\chi\_{\rm D}(1 + \cos\Theta\_{\rm D})\tag{2}$$

$$\text{\textit{\textit{\textit{\chi}s}}} = \text{\textit{\textit{\chi}}}\_{\text{Sd}} + \text{\textit{\textit{\chi}s}}\_{\text{\textit{\textquotesingle}p}} \tag{3}$$

where W, D, and S represent the polar liquid of water, the dispersion liquid diiodomethane, and the solid film; γ<sup>s</sup> , γsd, γsp are the SFE, the SFE dispersion component, and the SFE polar component of the solid film; γWd, γDd, γWp, γDp are the dispersion component and the polar component of water and diiodomethane, and these values are extracted from Owens and Wendt (1969); 2<sup>W</sup> and 2<sup>D</sup> are the contact angles of water and diiodomethane, respectively.

**Figures 2A–F** demonstrate the average water contact angle (WCA) and the diiodomethane contact angle (DCA) of the pure ITIC-M, pure PBDB-T, and PBDB-T:ITIC-M blend films on TOPD-coated ITO. For pure PBDB-T, pure ITIC-M, and PBDB-T:ITIC-M blend film, the WCAs are 101.20◦ (**Figure 2A**), 89.16◦ (**Figure 2B**), and 100.85◦ (**Figure 2C**), respectively; the DCAs are 48.51◦ (**Figure 2D**), 30.67◦ (**Figure 2E**), and 45.32◦ (**Figure 2F**), respectively. Correspondingly, the SFEs for pure PBDB-T and pure ITIC-M are 35.67 and 43.98 mJ/cm<sup>2</sup> calculated by the OW method, respectively. Apparently, the SFE of PBDB-T is lower than that of ITIC-M, and we can foresee that a much higher proportion of PBDB-T will accumulate at the top surface of PBDB-T:ITIC-M blend film to lower the systematic SFE. As expected, the SFE of the PBDB-T:ITIC-M blend film is 37.56 mJ/m<sup>2</sup> calculated from the OW method, which visibly indicates that the blend film minimizes its total free energy by increasing the surface concentration of the low-energy component PBDB-T (Jasieniak et al., 2016).

To construct a desired vertical component distribution within PBDB-T:ITIC-M blend film, TOPD was annealed at different temperatures to change its SFE and further modulate the vertical concentration distribution of ITIC-M. **Figure 3** presents the WCA and DCA of TOPD films before and after annealing at different temperatures. The WCA and DCA for TOPD film without annealing are only 26.01◦ (**Figure 3A**) and 7.41◦ (**Figure 3G**), which indicates the high SFE of 70.53 mJ/cm<sup>2</sup> . The WCA for TOPD film annealed at 80, 90 100, 110, and 120 ◦C are 73.03◦ (**Figure 3B**), 69.89◦ (**Figure 3C**), 66.62◦ (**Figure 3D**), 65.53◦ (**Figure 3E**), and 60.52◦ (**Figure 3F**); and their corresponding DCA are 25.00◦ (**Figure 3H**), 24.59◦ (**Figure 3I**), 24.01◦ (**Figure 3J**), 23.29◦ (**Figure 3K**), and 21.89◦ (**Figure 3L**), respectively. The SFEs are 47.74, 48.71, 49.9, 50.46, 52.46 mJ/cm<sup>2</sup> for TOPD film annealed at 80, 90 100, 110, and 120◦C according to the OW method, as shown in **Figure 3**. Obviously, the SFE increases gradually with the rise in the annealing temperature, and these subtle changes certainly adjust the vertical concentration distribution of ITIC-M.

#### Component Distribution of PBDB-T:ITIC-M Blends at the Air Surface

To clarify the influence of substrate SFEs on the component distribution at the air surface, the WCA and DCA of PBDB-T:ITIC-M on TOPD film without or with annealing are illustrated in **Figures 4A-L**, respectively. As can be seen, both

the WCA and DCA do not strongly depend on substrate SFEs, and their corresponding values remain almost the same whether TOPD is annealed or not. The WCAs change slightly around 100◦ and the DCAs alter around 50◦ , which suggests that there is no distinct difference about the component distribution of the electron donor and acceptor at the air surface. Hence, the changes of substrate SFEs hardly affect the component distribution at the top surface region (Björström et al., 2005; Tillack et al., 2011).

## Vertical Component Distribution Within the PBDB-T:ITIC-M Photoactive Layer

The photoactive layers deposited on three typical TOPD cathode buffer layers (before annealing, annealed at 90 and 120◦C) were selected to test the TOF-SIMS and further elucidate how SFEs affect the vertical component distribution. **Figure 5** plots the intensity change in the signals of S<sup>−</sup> 2 and CN<sup>−</sup> with sputter time since they are characteristic species for PBDB-T and ITIC-M, respectively. As can be seen, the PBDB-T concentration located at the air surface is distinctly higher than that found in the bulk whether TOPD is annealed or not. However, the PBDB-T concentration distribution almost decreases linearly with sputter time for TOPD without annealing; the PBDB-T concentration distribution reaches a maximum value at 6 s for TOPD annealed at 90 and 120◦C. When sputter time is beyond 55 s, in increasing order, the PBDB-T concentration distribution value for TOPD annealed at 120◦C, next to TOPD baked at 90◦C, is lower than that of TOPD before annealing. Obviously, the higher PBDB-T concentration distribution at the substrate interface has a detrimental effect on the collection of carriers for PSCs (Chen et al., 2009). Therefore, it is reasonable to postulate that the PSCs with TOPD baked at 90◦C afford high performance.

The distributions of ITIC-M greatly rely on heat treatment. For TOPD film without annealing, the ITIC-M distribution in the blends increases rapidly during the first 30 s and remains constant after 30 s deviating away from the ideal distribution of the electron acceptor. For TOPD film baked at 90 and 120◦C, their ITIC-M distributions reach a maximum value at 24 and 30 s, respectively, and the signal intensity of the former is stronger than that of the latter. There is no obvious change for signals of S − 2 and CN<sup>−</sup> after the sputter time beyond 150 s. From the results obtained up till now, three aspects have to be addressed. The first is that the air surface is always enriched with a lower-surfaceenergy polymer component and the influence of substrate SFEs is negligible, which is in accordance with the contact angle measurement results (Tanaka et al., 2010). The second is that the substrate SFEs have a more direct and remarkable effect on the vertical component distribution within the active layer, both for the electron donor and acceptor. Lastly, the TOPD film baked at 90◦C with the SFE of 48.71 mJ/cm<sup>2</sup> , which is very close to that of the ITIC film with an SFE of 43.98 mJ/cm<sup>2</sup> , tends to form a desired vertical component distribution facilitating carrier transportaion.

The mechanism of the vertical component distribution is complicated, including thermodynamics, kinetics, surface free energy, and selective dissolubility. In this work, we focus on the SFE of the TOPD layer, which roots from the kinetics of the molecular rearrangement in the blend films (Karagiannidis et al., 2011). As we know, the driving force for the lower SFE constituent accumulation to the high energy surface (air) is the lowering of the overall free energy of the system (Xu et al., 2010), leading to a larger concentration of PBDB-T at the surface. Theoretically, a complete demixing is expected to occur and the formation of a bilayer is thermodynamically possible if the thermal annealing above the polymer's glass transition temperature and the macromolecules obtain the appropriate mobility to rearrange (Klein et al., 2000). However, in the case of the PBDB-T:ITIC-M system, some ITIC-M molecules will probably diffuse into the PBDB-T layer driven by the different surface free energies of TOPD layers to achieve a more thermodynamically favorable component distribution, namely a phase separation will eventually come to a limit (Treat et al., 2011). Hence, the air–film interface is enriched with polymer, while the substrate–film interface is enriched with ITIC-M regardless of the annealing temperature. Films deposited on annealed TOPD show more ITIC-M close to the top surface (higher CN<sup>−</sup> to S<sup>−</sup> 2 signal at around 30 s) induced by the different

SFEs of the TOPD layer. Nevertheless, the underlying mechanism of the vertical component distribution is still under way for all researchers, which is also the area warranting our further study.

#### Photovoltaic Performance and Electron Mobility

To further demonstrate the interplay between the vertical component distribution and device performance, J-V results, PCE with error bars and IPCE spectra of the control device and i-PSCs with TOPD cathode buffer layers are shown in **Figure 6**. The key parameters of PCE, short-circuit current density (Jsc), open-circuit voltage (Voc), and fill factor (FF) under the illumination of AM1.5G, 100 mW/cm<sup>2</sup> (averaged over 12 individual devices), are compared in **Table 1**. The control device with PEDOT:PSS shows a PCE of 9.00%, with a Voc of 0.913 V, a Jsc of 14.76 mA/cm<sup>2</sup> , and an FF of 66.78%. The inverted device with an unannealed TOPD layer provides the lowest performance; the PCE, Jsc, Voc, and FF are 5.40%, 13.21 mA/cm<sup>2</sup> , 0.861 V, and 47.46%, respectivley. After annealing, the four parameters are all enhanced significantly, and the device with TOPD baked at 90◦C affords the highest device metrics, with a PCE of 10.20%, a Jsc of 16.88 mA/cm<sup>2</sup> , a Voc of 0.916 V, and an FF of 65.33%. The improvement in Jsc and FF benefits from the vertical concentration distribution of PBDB-T and ITIC-M in the active layer, which fortify the charge separation and transportation (Ma et al., 2014). The five annealing temperatures affect PCEs in descending order as 90, 80, 100, 110, 120◦C. The results of IPCE in **Figure 6B** are in agreement with the aforementioned results, which strongly confirm that the vertical concentration distribution of PBDB-T and ITIC-M within the blend film is affected by the substrate SFE greatly.

To illustrate how charge transportation and collection can be affected in the photoactive blends having different vertical component distributions, the J-V curves of singleelectron devices (Ahmed and Nakazato, 1996) with the structure of ITO/Al/TOPD/Al, ITO/Al/PBDB-T:ITIC-M/Al, and ITO/TOPD/PBDB-T:ITIC-M/Al are displayed in **Figure 7A**, in which the TOPD film annealed at 90◦C. The electron mobilities of TOPD, PBDB-T:ITIC-M, and TOPD/PBDB-T:ITIC-M are 8.56 × 10−<sup>3</sup> , 1.04 × 10−<sup>3</sup> , and 2.85 × 10−<sup>3</sup> cm<sup>2</sup> V −1 s −1 , respectively. The change of the electron mobility as a function of the annealing temperature in **Figure 7B** shows that TOPD with heat treatment manifests higher charge mobility than that of TOPD film without annealing, and a maximum is passed through. Without annealing, the electron mobilities of TOPD and TOPD/PBDB-T:ITIC-M are 9.51 × 10−<sup>5</sup> and 3.16 × 10−<sup>5</sup> cm<sup>2</sup> V −1 s −1 , respectively. On the other hand, the electron mobilities of TOPD baked at 80, 90, 100, and 110◦C are 5.74 × 10−<sup>3</sup> , 8.56 × 10−<sup>3</sup> , 3.48 × 10−<sup>3</sup> , and 4.71 × 10−<sup>4</sup> cm<sup>2</sup> V −1 s −1 , and the corresponding electron mobilities of TOPD/PBDB-T:ITIC-M are 1.91 × 10−<sup>3</sup> , 2.85 × 10−<sup>3</sup> , 7.77 × 10−<sup>4</sup> , and 1.56 × 10−<sup>4</sup> cm<sup>2</sup> V −1 s −1 , respectively. Therefore, the increased electron mobility for TOPD after heat treatment is in favor of charge transportation and collection. Our results confirm that 90◦C is an appropriate annealing temperature for TOPD from the standpoint of charge transportation and collection, which can increase Jsc and FF of devices.

## Other Structural and Optical Properties of TOPD Films

Other structural and optical information of TOPD film besides the SFE is explored to elaborate the device performance enhancements. **Figure 8A** shows the XRD patterns for TOPD films before and after annealing. Apparently, there does not appear to be any characteristic peak of TOPD except that of ITO, and all TOPD films are in the amorphous state whether they are annealed or not. But the peak of the glass substrate at 2θ = 22◦ illustrates obvious changes before and after annealing. Namely, the XRD curve of TOPD/ITO before annealing overlaps well with the bare ITO, and the peak intensity is very strong. On the other hand, there is no obvious peak at 2θ = 22◦ for TOPD annealed 80 or 90◦C, and the peak intensity gradually increases as the annealing temperature increases from 100 to 130◦C. This suggests that 80 and 90◦C are the suitable temperatures for forming uniform and compact film on the surface of ITO. The different peak intensities for TOPD annealed at higher temperature are caused by different aggregation behaviors of TOPD films. The AFM images of TOPD annealed at different temperatures are shown in **Figure 8B**. Evidently, the root-mean-square (rms) roughness rises with the increase in the annealing temperature of TOPD, and they are 2.86, 3.43, and 4.01 nm for TOPD annealed at 90, 110, and 130◦C, respectively. The suggests that fine and weak aggregation behaviors occur at lower annealing temperature, and the smooth and uniform film formed at lower temperature is favorable for effective charge transport and collection. This is in accordance with the result of the XRD results.

**Figure 8C** plots the absorption spectra of TOPD films before and after annealing at different temperatures. It is evident that the absorption of TOPD films decreases distinctly after annealing, which is caused by the organic residual solvents because there is no chemical component change before and after thermal annealing (Bai et al., 2018). The TOPD film baked at 90◦C demonstrates the weakest absorption, which ensures more light harvesting in the photoactive layer. On the whole, the absorption first decreases slightly as the annealing temperature rises from 80 to 90◦C, and then increases as the annealing temperature rises from 90 to 120◦C. Therefore, TOPD film baked at 90◦C shows superior structural and optical properties in addition to its appropriate SFE.

### REFERENCES


## CONCLUSIONS

Fine-tuning the SFE of the TOPD cathode buffer layer has been explored in this work, with the aim of unraveling the underlying mechanism and rational controlling the vertical distribution of the electron donor and acceptor. Our studies confirm that the SFE of TOPD increases gradually with the rise in the annealing temperature, and these subtle changes certainly cause the profound vertical component distribution within the bulk region of the PBDB-T:ITIC-M. The results of TOF-SIMS visibly demonstrate that TOPD film baked at 90◦C with the SFE of 48.71 mJ/cm<sup>2</sup> , which is very close to that of the ITIC film with the SFE of 43.98 mJ/cm<sup>2</sup> , tends to form a desired vertical component distribution facilitating charge transportation. Consequently, compared with conventional BHJ devices without tuning the donor and acceptor concentration, the PCE increases from 9.00 to 10.20% benefiting from the short circuit current density increase from 14.76 to 16.88 mA/cm<sup>2</sup> . The results obtained in this work allow the conclusion that modulation of the SFE of the substrate is a feasible way to control the vertical component distribution of the electron donor and acceptor. This approach holds great potential for practical application of high-efficiency PSCs.

## AUTHOR CONTRIBUTIONS

All authors listed have made a substantial, direct, and intellectual contribution to the work, and approved it for publication.

### FUNDING

Financial support by the National Natural Science Foundation of China [grant numbers 61006050, 51573042], the Natural Science Foundation of Beijing [grant number 2151004], and the Fundamental Research Funds for the Central Universities [grant numbers 2016MS50, JB2015RCJ02, 2016YQ06] in China is gratefully acknowledged.

## ACKNOWLEDGMENTS

Many thanks to Dr. Xingwang Zhang and Zhigang Yin for generous advice and help.

substituted 2d-conjugated polymer as donor. Nat. Commun. 7:13651. doi: 10.1038/ncomms13651


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Bai, Yang, Chen, Wang, Hayat, Alsaedi and Tan. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

## BN Embedded Polycyclic π-Conjugated Systems: Synthesis, Optoelectronic Properties, and Photovoltaic Applications

#### Jianhua Huang\* and Yuqing Li

*College of Materials Science and Engineering, Huaqiao University, Xiamen, China*

#### Edited by:

*Chuanlang Zhan, Institute of Chemistry (CAS), China*

#### Reviewed by:

*Liu Jun, Changchun Institute of Applied Chemistry (CAS), China Guo Chen, Shanghai University, China*

> \*Correspondence: *Jianhua Huang huangjianhua@hqu.edu.cn*

#### Specialty section:

*This article was submitted to Organic Chemistry, a section of the journal Frontiers in Chemistry*

Received: *18 June 2018* Accepted: *19 July 2018* Published: *07 August 2018*

#### Citation:

*Huang J and Li Y (2018) BN Embedded Polycyclic* π*-Conjugated Systems: Synthesis, Optoelectronic Properties, and Photovoltaic Applications. Front. Chem. 6:341. doi: 10.3389/fchem.2018.00341* In the periodic table of elements, boron (B, atomic number, 5) and nitrogen (N, atomic number, 7) are neighboring to the carbon (C, atomic number, 6). Thus, the total electronic number of two carbons (12) is equal to the electronic sum of one boron (5) and one nitrogen (7). Accordingly, replacing two carbons with one boron and one nitrogen in a π-conjugated structure gives an isoelectronic system, i.e., the BN perturbed π-conjugated system, comparing to their all-carbon analogs. The BN embedded π-conjugated systems have unique properties, e.g., optical absorption, emission, energy levels, bandgaps, and packing order in contrast to their all-carbon analogs and have been intensively studied in terms of novel synthesis, photophysical characterizations, and electronic applications in recent years. In this review, we try to summarize the synthesis methods, optoelectronic properties, and progress in organic photovoltaic (OPV) applications of the representative BN embedded polycyclic π-conjugated systems. Firstly, the narrative will be commenced with a general introduction to the BN units, i.e., B←N coordination bond, B-N covalent bond, and N-B←N group. Then, the representative synthesis strategies toward π-conjugated systems containing B←N coordination bond, B-N covalent bond, and N-B←N group will be summarized. Afterwards, the frontier orbital energy levels, optical absorption, packing order in solid state, charge transportation ability, and photovoltaic performances of typical BN embedded π-conjugated systems will be discussed. Finally, a prospect will be proposed on the OPV materials of BN doped π-conjugated systems, especially their potential applications to the small molecules organic solar cells.

Keywords: BN-embedded unit, isoelectronic structure, π-conjugated material, organic solar cell, device performance

## INTRODUCTION

The past years have witnessed a fruitful advance of organic conjugated materials and great enthusiasm was fueled to develop novel π-molecules and judiciously apply them to organic electronic devices, e.g., organic field effect transistors (OFETs) (Gsänger et al., 2016; Li M. et al., 2018), organic light emitting diodes (OLEDs) (Grimsdale et al., 2009), organic solar cells (OSCs) (Lu et al., 2015; Zhan and Yao, 2016), organic thermoelectric devices (OTEDs) (Shi et al., 2015; Huang et al., 2016; Lim et al., 2018), and organic photodetectors (OTDs) (Wang et al., 2016a; Benavides et al., 2018; Murto et al., 2018). Especially, the bulk-heterojunction (BHJ)-type OSCs adopting organic semiconductors as photo-sensitive layers have been considered a promising candidate for the next generation of green energy due to solution processability, low cost, light weight, flexibility features of organic materials. The photosensitive layers of OSCs are blends of an electron-donor (p-type) and an electron-acceptor (n-type) with nano-phase separated morphology. Although the fullerene derivatives, e.g., phenyl-C61-butyric acid methyl ester (PC61BM) and phenyl-C71-butyric acid methyl ester (PC71BM) have been the dominant acceptor materials in a long time (Sariciftci et al., 1992), both the electrondonor and electron-acceptor materials have been extended to π-conjugated linear molecules, star molecules, oligomers, and polymers in recent years. To achieve satisfactory power conversion efficiency (PCE), the photo-sensitive materials should be featured by the following points (**Figure 1A**), (1) strong light-harvesting ability resulting from wide absorption band and strong absorption coefficients; (2) appropriate energy level alignment between the p-type and n-type materials to ensure efficient built-in field and driving force for exciton dissociation; (3) proper aggregation and crystallization ability for both of ptype and n-type materials to form well-defined blend film with desirable micro-morphology, e.g., domain sizes and molecular stacking order; (4) fairly well-charge carrier mobility, i.e., electron and hole mobility to facilitate the charge transportation and collection. These features are closely related to the material properties and device preparation technics. Thanks to the continuing devotion on material design and device optimization, the PCE of single junction photovoltaic devices based on organic semiconductors have been promoted from the initial 1% in 1986 (Tang, 1986) to 10–13% recently (Gupta et al., 2013; Chen et al., 2015; Zhang Z. G. et al., 2017; Li W. et al., 2018), illuminating the bright future of OSCs for low-cost and portable energy provision. However, in contrast to the inorganic and hybrid photovoltaics, for example, the silicon solar cells and perovskite solar cells, whose efficiencies are commonly on the magnitude of ca. 20% (Sun, 2015; Meng et al., 2016), the OSCs have a large offset to promotion. In fact, theoretical models based on Shockley– Queisser detailed balance approach predicted a reachable PCE of 20–24% for OSCs (Janssen and Nelson, 2013), Moreover, in the current stage, excellent photovoltaic materials capable of accomplishing efficiencies higher than 10% are limited. As such, large amount of fundamental explorations on developing novel photovoltaic materials are required to thrust the overall progress of OSCs.

The most popular strategy to construct the organic photovoltaic materials involves the covalently bonding of various conjugated units with electron-rich (D) or electrondeficient (A) nature to obtain D-A type linear molecules, star molecules, oligomers, and polymers. These D and A π-electronic units are basic building blockings that critically determine the optoelectronic properties and photovoltaic performances of the photovoltaic materials. Consequently, the design and structural tailoring of π-electronic units are essential for the construction of photovoltaic materials. To now, outstanding D units such as oligothiophene (OT), fluorene (Fl), cyclopentadithiophene (CPT), benzodithiophene, (BDT), and indacenodithiophene (IDT) and A units including perylene diimide (PDI), naphthalene diimide (NDI), diketopyrrolopyrrole (DPP), isoindigo (IID), thieno[3,4-c]pyrrole-4,6-dione (TPD), benzothiadiazole (BT), benzotriazole (BTz), benzo[1,2-c:4,5-c′ ]dithiophene-4,8-dione (BDD), rhodanine (Rh), cyano indone (IC), and N,N′ -diethyl thiobarbituric acid (TBA) were revealed in literature (**Figure 1B**) (Lu et al., 2015; Zhan and Yao, 2016). Additionally, classic dye molecules such as phthalocyanine (Pc), porphyrin (Pr), and squaraine (SQ) are also frequently reported for construction of OPV materials (Chen et al., 2012, 2014; Chen G. et al., 2015). Developing novel D or A π-electronic units has being an energetic realm. A typical strategy of introducing heteroatoms including O, N, P, S, Se, Si, Ge, and B, etc. into the polycyclic aromatic hydrocarbons (PAH) backbone is widely used to tailor the properties of π-electronic units (Stepien et al., 2017).

When one gives a glance to the periodic table of elements, it's a cinch to perceive the neighbor elements of carbon (C, atomic number, 6), i.e., boron (B, atomic number, 5) and nitrogen (N, atomic number, 7). Thus, the total electronic number of two carbons (12) is equal to the electronic sum of one boron (5) and one nitrogen (7). Accordingly, replacing two carbons with one boron and one nitrogen in a π-conjugated structure gives an isoelectronic system, i.e., the BN embedded π-conjugated system, in contrast to its all-carbon analogs. The bonds between the B and N can be formed as coordinated bond (B←N) and covalent bond (B-N), corresponding to the isoelectronic units of C–C and C=C (**Figure 2A**), respectively. Replacing CC unit with BN unit in the conjugated skeleton is favorable for property adjustment. On the one hand, the BN would alter the electronic nature of the conjugated backbone due to the different electron-negativity of heteroatoms with that of carbon atom. On the other hand, the BN also enhances the dipolarity of hydrocarbon skeletons and thus boost the inter-molecular interactions. Additionally,

replacing carbons with BN usually maintains the good coplanarity and rigidity of the backbones. All these features are desirable to design novel π-electronic units for photovoltaic materials construction. Taking the isoelectronic compounds of ethane and ammonia-borane (NH3←BH3, AB) for an example, ethane is gaseous at ambient temperature with a dipole moment of zero and weak inter-molecular interactions (Pritchard and Kern, 1969), whereas AB is a solid state at room temperature with a strong dipole moment of 5.2 D and inter-molecular BH. . . NH interactions (Leroy et al., 1993). For B←N embedded aromatic systems, the first report was in 1963 by Morrison et al. (Letsinger and MacLean, 1963). Recently, a series of conjugated materials containing B←N bonds have been revealed for OSCs application (Dou et al., 2015, 2017). For the B-N covalent bond embedded aromatic structures, the research history has been almost one century since the first synthesis of borazine in 1926 (Stock and Pohland, 1926). In 1950s and 1960s, Dewar and coworkers conducted pioneering work on synthesis of BN doped PAH (Dewar et al., 1958; Dewar and Dietz, 1959; Chissick et al., 1960; Davies et al., 1967). Since then, little progress in this field has been made due to limited characterization means at that time. Recently, the B-N embedded polycyclic aromatic systems are experiencing a renaissance with fast development of synthesis protocols and widely application to H<sup>2</sup> storage, OLEDs, and OFETs (Jaska et al., 2006, 2007; Bosdet et al., 2007a; Liu and Marder, 2008; Campbell et al., 2012; Hashimoto et al., 2014; Wang et al., 2015a,b, 2016b; Beniwal et al., 2017; Ishibashi et al., 2017). Additionally, N–B←N group, the comprehensive form of B←N coordination and B-N covalent bonds, also widely appears in the conjugated units, such as BODIPY (Loudet and Burgess, 2007). All of these BN perturbed structures have unique properties and are intensively interested in terms of synthesis routes, optoelectronic properties, and electronic device performances. However, studies on the photovoltaic applications of BN embedded π-electronic units are still in infancy. In this review, we are going to summarize the synthesis routes toward π-electronic units containing B←N coordination bond, B–N covalent bond, and N–B←N group (**Figures 2B–D**), and discuss their optoelectronic properties, as well as their applications in photovoltaic devices.

## SYNTHESIS ROUTES

## Synthesis of π-Electronic Units Containing B←N Coordination Bond

#### Alkyl Lithium (e.g., n-BuLi)/ Aryl Boron (e.g., BPh3) System

In 2002, Erker et al. reported an intramolecular nucleophilic aromatic substitution reaction, using CH3Li and B(C6F5)<sup>3</sup> to prepare the tricyclic fused structures containing B←N coordination (Dominik et al., 2002). As shown in **Figure 3**, the starting N-Methylimidazole **1** was coordinated with strong Lewis acid of B(C6F5)<sup>3</sup> to form adduct **2**. Deprotonation at the C-2 position of imidazole heterocycle was accomplished by treatment with CH3Li, affording intermediate **3**, which experienced a rapid intramolecular nucleophilic aromatic substitution reaction with

one of the adjacent C6F<sup>5</sup> groups to generate fused π-electron unit **4,** containing B←N coordination bond. One year later, they replaced the starting reactant with 1-methylbenzimidazole **5**. Using the same routes, larger conjugated π-electron unit **6** was obtained (Vagedes et al., 2003). Similarly, in 2010, B←N perturbed structures with further extended conjugation (**7** and **8**) were synthesized using the same strategy (Job et al., 2010). In 2006, Yamaguchi et al. reported the synthesis of π-electron systems **10** containing B←N coordination from **9** by n-BuLi and Mes2BF (Mes<sup>2</sup> = 2, 4, 6-Me3C6H2), whose mechanism involves the coordination between thiazole N and aryl B and consequent electrophilic attack of electron-deficient boron to the β-site of adjacent thiophene (Wakamiya et al., 2006). Recently, Liu et al. utilized this route to produce a stable electron-deficient unit **11**. By co-polymerizing with D or A units, they constructed a series of novel photovoltaic polymers with outstanding performances (Dou et al., 2015).

#### BX3/Hindered Base System

This synthetic method can be traced back to 1963, when Maclean et al. passed BCl<sup>3</sup> into melts of 2-phenylbenzirnidazole (**12**) at 300◦C and subsequently hydrolyzed in moisture air producing **13** (Letsinger and MacLean, 1963), as shown in **Figure 4**. The C-H borylation was considered to be reversible and the byproduct HCl should be sequestered to improve the reaction yield. Accordingly, in 2010, Murakami et al. improved the method by adding a hindered base, Et2N(i-Pr) to absorb the protic by-product (Ishida et al., 2010). They used **14** as starting reactant, by adding 3 eq BBr<sup>3</sup> and 1 eq Et2N(i-Pr) at 0◦C, achieving **15**. **15** was not stable in moisture due to the electrondrawing property of Br, endowing B atom strong electrophilic. Further functionalization at B atom can be realized by adding organometallic reagents to substitute the Br atoms, affording a series of stable ladder-type π-units containing B←N bonds (**16a**-**16f**). The reaction mechanism was proposed as follow: the Lewis acid-base coordination between **14** and BBr<sup>3</sup> provided **17**; then, another BBr<sup>3</sup> captured a Br<sup>−</sup> from **17**, leading to trivalent cationic boron species **18**; finally, the cationic boron attacked the neighboring aromatic unit, generating the circular **16**. After that, several conjugated units containing B←N coordination with tunable emission and aggregating-induced emission properties were synthesized by this method (Wong et al., 2012, 2016; Zhao et al., 2013). Recently, based on this strategy, fused πelectronic units with good co-planarity, red-shifted absorption, and depressed energy levels of **21**, **23**, and **25** have been synthesized from **20**, **22**, and **24**, respectively (Yusuf et al., 2016; Zhu et al., 2016; Li Y. et al., 2018). These π-electronic units containing B←N coordination are potentially useful to construct organic semiconductors for electronic device applications. As this strategy involves the electrophilic attack on the aromatic units, the electron-rich nature of the aromatic cycles is critical to the C-H borylation. Ingleson and Turner et al. employed **26** as precursor to conduct the C-H borylation (Crossley et al., 2015). It's found that the C-H borylation occurred on the thiophene rather than the fluorene, presumably due to the more electronicrich nature of thiophene than fluorene, facilitating electrophilic attack on the thiophene. Not only the small molecules, but also the polymers can undergo this reaction. Ingleson and Turner et al. also applied this method to modify the copolymer **28**, yielding near-infrared emitting polymer **29** (Crossley et al., 2017).

Although the BX3/hindered base reaction condition has been demonstrated to be widely applicable to the C-H borylation, it's invalid in some cases. For example, Ingleson and Turner et al. found that precursor **30** can not afford **31** upon adding BX3, e.g., BCl<sup>3</sup> or BBr<sup>3</sup> and the hindered base, e.g., EtN(i-Pr)<sup>2</sup> or 2,4,6-trit-butylpyridine (TBP) (Crossley et al., 2015). Otherwise, the C-H borylation occurred by adding excess BCl<sup>3</sup> (ca. 4 eq), 2 eq of TBP and 4 eq of AlCl3, yielding intermediate **32**, which was readily transformed to **31** by adding Bu4NCl. The addition of 4 eq of AlCl<sup>3</sup> was regarded to be essential to ensure the fully conversion to intermediate **32**. The function of AlCl<sup>3</sup> herein is similar to its effect in the classic Friedel–Crafts reaction. It's worth to note although the structure units of **30** is the same to the repeating units of polymer **29**, the C-H borylation conditions were different for this two precursor, indicating the different reaction law in small molecules and polymers for C-H borylation. Recently, 2D conjugated units containing B←N coordination (**34** and **35**) were reported, also based on this method (Liu K. et al., 2017).

## Synthesis of π-Electronic Units Containing B-N Covalent Bonds

#### Electrophilic Cyclization Between Boron and Aromatic Units

This method involves the Friedel-Crafts cyclization, in which BX<sup>3</sup> and Lewis acid are usually required to complete the cyclization (**Figure 5**). In 1958, Dewar et al. initially conducted the synthesis work from **36** by adding BCl<sup>3</sup> and AlCl<sup>3</sup> to obtain 9,10-azaboraphenanthrenes (**38**) (Dewar et al., 1958). Further modification on the B atom led to a series of BN-substituted

phenanthrene derivatives (**39**). Consequently, a family of B-N embedded PAHs was synthesized via the similar strategies (Dewar and Dietz, 1959; Chissick et al., 1960; Dewar and Poesche, 1963, 1964). In 2013, Pei and coworkers reported BN-substituted tetrathienonaphthalene derivatives (**41**) (Wang X. Y. et al., 2013), starting from **40** by adding BBr<sup>3</sup> and Et3N. BBr<sup>3</sup> attacked the imine and consequently electrophilic attacked the β-site of thiophene to finish the cyclization. Latterly, they revealed extended π-conjugated structure **43** with similar cyclization methods (Wang et al., 2014). It's worth to note that the conjugation of **42** is more extended than **40**, leading to weaker electron-rich of imine groups in **42**. Accordingly, the n-BuLi was required to facilitate the attack of BBr<sup>3</sup> to imine in **43**. Similarly, starting from **44**, Nakamura et al. utilized n-BuLi and BBr<sup>3</sup> to prepare the intermediate **45** (Hatakeyama et al., 2011). Due to weaker electron-rich properties of phenyl than thiophene, the Lewis acid such as AlCl<sup>3</sup> and hindered base TBP were required to complete the electrophilic cyclization to produce **46**. Liu et al. reported the synthesis of B-N embedded tetracene **48** and **49**, starting from **47** (Ishibashi et al., 2014, 2017). Pei et al. also reported synthesis of heterocoronene (**51**) by adding PhBCl<sup>2</sup> and Et3N to **50** and heated to 180◦C in o-DCB (Wang et al., 2015c). Based on similar methods, laddertype conjugated units substituted by B-N covalent bonds were also synthesized (Wang X. et al., 2013; Zhou et al., 2016). Wang et al. revealed an electrophilic cyclization between B and methyl located on phenyl, obtaining unsaturated **53**, which was subjected to photoelimination, leading to **54** (Lu et al., 2013; Ko et al., 2014; Yang et al., 2015, 2016, 2017).

#### Chelation of Aromatic N and B Precursor

This method involves the chelation of B precursor and N Lewis base to eliminate a by-product (**Figure 6**). In 2003, Piers et al. reported the chelation of **55** with pyridazine and benzo[c]cinnoline, eliminating Me3SiCl to obtain **56** and **57**, respectively (Emslie et al., 2003). Latterly in 2006, this method was applied to synthesize **58** (Jaska et al., 2006). In 2007, they furtherrevealed synthesis of **61** by chelation of **59** with 2-ethynylpyridine, leading to intermediate **60**, which experienced smooth cyclization without any catalyst (Bosdet et al., 2007a). In another aspect, the chelation of **59** with 2, 5-diethynyl-pyridine afforded **62**, which required catalytic amount of PdCl<sup>2</sup> to complete the second cyclization of the ethynyl group to obtain the pyrene analog with internalized B-N substitution **63** (Bosdet et al., 2007b). The chelation and ethynyl-cyclization strategies were spread widely to prepare a series of PAHs embedded with B-N bonds (Jaska et al., 2007; Bosdet et al., 2010; Benedikt et al., 2013).

## Synthesis of π-Electronic Units Containing N–B←N Groups

A family of conjugated molecules containing N–B←N groups has been intensively explored as fluorescence dyes with high

absorption coefficients and fluorescence quantum efficiency, which were widely utilized in OSCs, OLEDs, sensing, and imaging, etc (Li et al., 2013; Fu et al., 2015; Lin et al., 2015; Dou et al., 2017). The synthesis routes toward N–B←N groups usually require a precursor equipped with an amino group and an aromatic nitrogen at suitable position for chelation of the boron. BF3•OEt2/Et3N is most widely used reaction condition (**Figure 7**). Taking the typical dye boron dipyrromethene (BODIPY) as an example, the precursor **64**, usually synthesized from pyrrole derivatives and aldehydes, is readily to obtain the BODIPY skeleton **65** by adding BF3•OEt2/Et3N (Loudet and Burgess, 2007). This reaction condition is widely applicable to the precursors with the features of containing amino group and aromatic nitrogen atoms at appropriate positions, e.g., **66**, **67**, **68**, and **69** (Araneda et al., 2011; Nawn et al., 2013; Hao et al., 2014; Qiu et al., 2016).

### OPTOELECTRONIC PROPERTIES AND OPV APPLICATIONS

## π-Electronic Units Containing B←N Coordination Bonds

In a long time, the inter-molecular coordination between B and N has been well-demonstrated to adjust the optoelectronic properties of conjugated molecules. It's well-known that the B atom is an electron-deficient center (Lewis acid) due to the existence of an unoccupied orbital while the N atom is an electron-rich center (Lewis base) owing to the existence of un-bonded pair of electron. As such, typical Lewis acid-base coordination between B and N atoms occurs when molecules containing N and B atoms are mixed together (Maria and Gal, 1985; Piers, 2005). It's has been revealed that the optoelectronic properties of conjugated molecules containing N atoms can be readily amendable when mixed with boride Lewis acid, e.g., BF3, BCl3, BBr3, and B(C6F5)<sup>3</sup> (BCF). In 2009, Bazan and coworkers reported the bandgap control of benzothiadiazole-based oligomers via Lewis acid of B(C6F5)<sup>3</sup> (Welch et al., 2009). As shown in **Figures 8A,B**, upon stoichiometric coordination with BCF, the absorption band of **74** red-shifted and the optical bandgap (E ◦ g pt) decreased from 2.15 to 1.60 eV. In 2011, they further implemented the method to a series of oligomers and polymers. The HOMO and LUMO, estimated from ultraviolet photoelectron spectroscopy (UPS) were found to synergetic lowering due to the introduction of electron-deficient center B to the conjugated backbone (**Figure 8C**) (Welch and Bazan, 2011). It's interpreted that the Lewis acid BCF pulled the electron density away from the conjugated backbone, altering the electron topology and leading to decreased HOMO/LUMO and optical bandgap. In 2017, the strategy was employed to the dye molecules of 7-azaisoindigo, by using BF<sup>3</sup> to amend the energy levels and optical absorption (Randell et al., 2017). Recently, we synthesized a series of pyridine end-capped diketopyrrolopyrrole (DPP) dye molecules and systematically explored the optical bandgap alteration upon coordinating with BCF (Huang et al., 2018). The effects of stoichiometry and equilibrium of the Lewis

acid-base interactions on the optical bandgaps were studied (**Figure 8D**).

Lewis acid-base complexation of conjugated polymers containing aromatic N atoms with BCF can also adjust the performances of OFET and OLED devices. Heeney et al. synthesized two indenopyrazine-based copolymers, with which the OFET devices were prepared. They explored the effects of doping BCF to the copolymers on the device performances (Han et al., 2016). It's found that by doping the polymers with BCF in a certain amount, e.g., 0.075 equiv, the hole mobility can be increased up to 11-fold along with the reduced threshold voltages. Otherwise, increased the amount of BCF to a critical amount, the OFET performances would be adversely affected. It's deduced by the authors that moderate amount of BCF leads to effective traps filling and positive effects on the device operation while the excess amount of BCF gives rise to defect formation and structural disorder, which negatively affects the device performances. Bazan and Nguyen et al. studied the color turning of OLED by BCF doping (Zalar et al., 2012). They selected a fluorescent copolymer of fluorene and pyridine. By doping the polymer with BCF, the OLED emission color red-shifted obviously.

Although the inter-molecular B←N interactions are effective to adjust the optoelectronic properties of conjugated molecules, this method are not applicable for OPV devices because the inter-molecular B←N complexation is unstable and the boride molecule dopants may lead to defect formation and hinder the molecular order packing. Consequently, incorporating B←N bonds into the molecular skeletons are more feasible for OPV applications. Recent development of synthesis protocols promoted the birth of several π-units containing B←N bonds (**Figure 9**). As the frontier orbital energy levels, i.e., HOMOs and LUMOs and optical absorption are critical parameters for the PCEs of OPV devices, we summarized these parameters of some B←N embedded π-units, as shown in **Table 1**. For **20** and **22** (**Figure 4**), after introducing the BBr<sup>2</sup> groups into the backbones, the LUMOs were decreased significantly by 0.96 and 0.53 eV, corresponding to **21** and **23**, respectively (Zhu et al., 2016). Moreover, remarkably red-shifted absorption band to near-infrared region occurred after introducing the BBr<sup>2</sup> groups

into backbones (**Figure 10A**). Similarly, LUMOs of **79**, **81**, and **83** also lowered obviously in comparison to their precursors of **78**, **80**, and **82** (Crossley et al., 2015). It's worth to note that the LUMO depressed remarkably with the slightly changed HOMOs, leading to decreased bandgaps. An interesting comparison from precursor **84** (**86**), to inter-molecular B←N complex **85** (**87**), and to cyclization product **7** (**8**) further demonstrates the outstanding ability of B←N unit to depress the LUMO energy levels of πunits (Job et al., 2010). Not only compared to their precursors before cyclization, but also in contrast to the all-carbon analogs,

Reprinted with permission from Huang et al. (2018). Copyright (2018) Elsevier Ltd.

the B←N embedded π-units also exhibit significantly lowered LUMOs, as illustrated by **88** and **89** (Liu K. et al., 2017). These results indicate that the B←N embedded π-units usually have depressed LUMOs and expanded absorption bands in contrast to their precursors and all-carbon analogs due to the introduction of electron-deficient center B.

Other properties that essentially affect the PCEs of OPV devices are the solid packing order and charge carrier mobility. Single crystal data indicate the B←N embedded π-units, e.g., **8** also have good co-planarity, rigidity, and ordered π-π packing, as shown in **Figure 10B** (Job et al., 2010). It's beneficial to the charge transport in solid state, which is critical to the OPV performances. These features of depressed LUMOs, namely, strong electron-affinity and good molecular planarity and π-π packing order make the B←N embedded π-units suitable for the electron-transporting materials. For example, the dimeric B←N embedded CPT showed electron mobility of 1.5×10−<sup>4</sup> cm<sup>2</sup> /V•s tested by time-off-light (TOF) carrier-mobility measurement (Wakamiya et al., 2006). Comprehensively, the B←N embedded π-units are excellent electron-deficient moieties with good co-planarity, depressed LUMOs, broadened absorption bands, and high electron mobility, which are promising for the application of photovoltaic materials, especially for acceptor materials. However, the contributions to exploit the potential

TABLE 1 | Frontier orbital energy levels and optical bandgaps of B←N embedded π-units.


*<sup>a</sup>Obtained by theoretical calculations. Other HOMOs and LUMOs were estimated by electrochemistry method. <sup>b</sup>Calculated by 1240/*λ*onset.* λ*onset is the absorption onset of UV-Vis absorption spectra.*

of B←N embedded π-units for OPV application are scarcely revealed.

Until recently, Liu and co-workers' pioneering work demonstrated the great potential of B←N embedded π-units for the construction of OPV materials. They selected B←N embedded CPT (BNCPT), which was developed in 2006 by Yamaguchi et al. (Wakamiya et al., 2006), as co-monomer to copolymerize with thieno[3,4-c]pyrrole-4,6-dione-1,3-diyl (TPD) unit, obtaining a novel conjugated polymer P-BN (Dou et al., 2015). For comparison, the all-carbon analog CPT was also copolymerized with TPD, leading to P-CC. The HOMO and LUMO of P-BN were significantly depressed by 0.65 and 0.53 eV, respectively, in contrast to the values of P-CC (**Figure 11A**), indicating the electron acceptor property of P-BN, which was further confirmed by the fluorescence quenching of P-BN with P3HT in solutions. These results demonstrated the B←N based copolymers are suitable for electron acceptor in OPV devices. Then, they synthesized another copolymer by combining BNCPT with isoindigo (IID), affording P-BN-IID with HOMO = −3.80 eV and LUMO = −5.84 eV (Zhao et al., 2016). Using PTB7-Th as electron donor, the all-polymer solar cells based on P-BN-IID exhibited a competitive PCE of 5.04% (**Figure 11B**). On the other hand, the BNCPT was also adopted to construct electron donor polymers by copolymerizing with its all-carbon analog CPT. This polymer displayed suitable HOMO and LUMO levels and exhibited a PCE of 3.74% by using PC71BM as electron acceptor (Zhang et al., 2015). These pioneering studies on the application of B←N based units to the OPV materials open a new window for the design of novel and highly efficient

photovoltaic materials, not only for the polymers, but also for the small molecules.

## π-Electronic Units Containing B-N Bonds

As the B-N covalent bond is isosterism of C=C bond, replacing the C=C unit in a PAH with the isosteric B-N has emerged as a useful strategy to enlarge the library of π-conjugated units. The B-N embedded PAHs usually have similar geometric parameters but rather distinct electronic structures to its allcarbon analogs. As for the OPV materials, the energy levels, absorption spectra and solid state packing ability are extensively concerned. Herein, we summarized recent progress in some typical B-N embedded PAHs, emphasizing the comparison of B-N doped π-conjugated units to their all-carbon analogs in terms of frontier orbital energy levels, absorption spectra, bandgaps, as well as single crystal packing order. Liu et al.

conducted systematic studies on the B-N embedded acenes, e.g., naphthalene, anthracene, and tetracene (Ishibashi et al., 2014, 2017; Liu Z. et al., 2017). In 2014, they revealed two B-N isosteres of anthracene, i.e., BN anthracene and bis-BN anthracene (**Figure 12A**) (Ishibashi et al., 2014). HOMO level tested from UV-photoelecton spectroscopy was −7.4, −7.7 eV, and −8.0 eV, respectively for anthrecene, BN anthracene, and bis-BN anthracene, indicating that the replacement of C=C with B-N gave rise to stabilized HOMO levels. Optical bandgaps estimated from the onset of the absorption spectra were similar for the three molecules. Comparing to anthracene, BN anthracene and bis-BN anthracene appeared a new absorption band at 310 nm, with relatively stronger oscillator strength, which mainly originated from the HOMO−1 to LUMO transition. Recently, they extended the reach of BN/CC isosterism to the tetracene, obtaining B-N perturbed tetracene (**Figure 12B**) (Ishibashi et al., 2017). In contrast to the all-carbon analog, the B-N perturbed tetracene showed slightly depressed HOMO and larger optical bandgap. Upon embedding B-N to tetracene, a blue-shift of HOMO to LUMO transition from 446 to 427 nm occurred and new absorption band originated from HOMO-1 to LUMO with stronger oscillator strength appeared around 380 nm. Very recently, they also disclosed that the orientation and location of B-N in the naphthalene exerted critical influence on the energy levels, bandgaps and absorption properties (Liu Z. et al., 2017).

The B-N embedded dibenzo[g,p]chrysene (BN-DBC) reported by Nakamura et al. showed negative-shifted redox potential in comparison to the all-carbon analog, dibenzo[g,p]chrysene (DCB) (**Figure 13A**) (Hatakeyama et al., 2011), indicating the synergistic depression of HOMO and LUMO levels and unchanged electrochemical bandgaps. X-ray crystallography data revealed the twisted conformations and offset face-to-face stacking style with π-π distances of 3.3–3.6 Å for both DCB and BN-DCB. Although the similar molecular stacking style for DCB and BN-DCB, the hole mobilities were distinct, with 0.07 and 0.007 cm<sup>2</sup> /V•s, respectively for BN-DCB and DCB. The favorable hole mobility of BN-DCB are beneficial from the introduction of polar B-N unit into backbone

permission from Ishibashi et al. (2014, 2017). Copyright (2014, 2017) American Chemical Society.

leading to stronger electronic coupling between neighboring molecules. Pei and co-workers developed two BN-substituted tetrathienonaphthalene derivatives, i.e., BN-TTN-C3 and BN-TTN-C6 (Wang X. et al., 2013). Because of the different side chains, the two BN embedded units exhibited distinct packing mode, with helical and layered packing style for BN-TTN-C3 and BN-TTN-C6, respectively. BN-TTN-C3 displayed close π-π stacking (3.44 Å) in crystal state whereas BN-TTN-C6 showed CH-π interaction. Moreover, BN-TTN-C3 gave closer dipole-dipole interaction (6.763 Å) compared to that of BN-TTN-C6 (9.207 Å). Due to the higher ordered molecular packing, BN-TTN-C3 exhibited superior hole mobility of 0.12 cm<sup>2</sup> /V•s, elevated by one magnitude than the value of BN-TTN-C3 (0.03 cm<sup>2</sup> /V•s). Recently, they copolymerized the BN-TTN with thiophene units to afford the conjugated polymers with lowered HOMO levels (−5.46 ∼ −5.67 eV) and strong intermolecular interactions (Wang et al., 2015b). OFET devices prepared from these azaborine-based polymers exhibited a champion hole mobility of 0.38 cm<sup>2</sup> /V•s. By changing the co-monomers, a vast of novel copolymers based on this azaborine unit can be obtained, predicting a great potential of this unit for electronic device applications.

Furthermore, they developed a straightforward strategy to produce the largest BN embedded heteroaromatic (**43**) to date (Wang et al., 2014). Single crystal X-ray diffraction indicated the significant distorted conformation of **43** due to the steric hindrance among peripheral rings and two different conformations (A and B) were found in the same crystal. The single crystal showed a columnar stacking style along the 011 direction and one-dimensional micro-ribbons can be obtained feasibly due to the strong π-π interactions. OFET devices based on the micro-ribbons gave a hole mobility of 0.23 cm<sup>2</sup> /V•s, a low threshold of −3 V, and a current on/off ratio of > 10<sup>4</sup> . Theoretical calculations indicated a depressed HOMO and unchanged LUMO of **43** in contrast to its all-carbon analog. HOMO estimated from electrochemistry and optical bandgap calculated from the absorption onset was −5.07 and −2.59 eV, respectively for **43**. OPV devices prepared with **43** as donor and PC71BM as acceptor gave a PCE of 3.12% and a Voc of 0.96 eV (Zhong et al., 2016). Moreover, when it was added to PTB7/PC71BM system as additive, the ternary solar cells displayed improved PCE (4.75%) in comparison to the PCE of binary devices (3.91%) (**Figure 13B**). This is the first example of applying B-N embedded heteroaromatics to the OPV devices.

From the aforementioned discussion on the B-N embedded units, the following conclusions can be deduced. Different from the B←N bond, which amends the energy levels and optical absorption of π-units significantly, the B-N embedded π-units usually have similar or slightly different energy levels and optical bandgaps to their all-carbon analogs. Most of the B-N embedded units have large optical bandgaps and narrow absorption bands. Extending the conjugation of the B-N embedded π-units would broaden the absorption bands. Introducing the dipolar B-N bond into the conjugated backbone would enhance the intermolecular interaction, facilitating the ordered π-π stacking and enhancing the hole mobility in contrast to their all-carbon analogs. Accordingly, B-N embedded π-units are excellent electron-rich units and promising candidates for construction of OPV materials. Otherwise, to now, OPV applications involving the B-N embedded π-units are scarcely revealed. From my point of view, these B-N embedded PAHs are promising for OPV applications and will represent an important direction of OPV materials.

## π-Electronic Units Containing N–B←N Groups

The typical π-electronic unit containing N–B←N group is the 4,4′ -difluoro-4-bora-3a,4a-diaza-s-indacene (BODIPY), possessing unique optoelectronic properties, e.g., strong molar extinct coefficient (10<sup>5</sup> M−1•cm−<sup>1</sup> ), low-lying HOMO (−5.5 eV) and LUMO (−3.5 eV), strong electron affinity, and high fluorescence quantum yield, which has drawn much attention in the field of labeling and chemical sensors (Sekiya et al., 2009; Lu et al., 2014). For OPVs, the BODIPY also plays an important role, either in the electron donor or acceptor materials. In general, α, β, and meso-positions are readily available for chemical modification to adjust the properties of BODIPY for OPV applications (**Figure 14**). **Table 2** summarizes the optoelectronic and photovoltaic parameters for BODIPYbased molecules. In 2009, Roncali et al. initially reported the BODIPY-based small molecules modified at α-position with styryl and meso-position with iodobenzene, i.e., **90a** and **90b** as electron donor materials, affording an optimum PCE of 1.17 and 1.34%, respectively, by using PC61BM as electron acceptor material (Rousseau et al., 2009a). Interestingly, ternary device prepared by blending **90a**, **90b**, and PC61BM as active layer showed a promoted PCE of 1.70% (Rousseau et al., 2009b). Then, the further modification of **90b** at meso-position with oligothiphene gave **90c**, leading to an improved PCE of 2.17% due to the enhanced hole mobility (Rousseau et al., 2010). In 2012, Ziessle et al. substituted the α-position of BODIPY with vinylthiophene to obtain **91a**, **91b**, **91c**, and **91d**, exhibiting a maximum PCE of 1.40, 4.70, 0.90, and 1.50%, respectively (Bura et al., 2012). The highest PCE for **91b** was interpreted by its depressed HOMO levels, broad and strong external quantum response and high hole mobility. Replacing at α and mesopositions with triphenylamine produced **92a** and **92b**, giving a moderate PCE of 1.50 and 0.51%, respectively (Kolemen et al., 2014). However, tailoring the meso-substituents with carbazole units along with device technique optimization by thermal annealing and solvent vapor annealing, **93a**, **93b**, and **93c** yielded a superior PCE of 5.05, 3.99, and 4.80%, respectively, by using PC71BM as electron acceptor material (Jadhav et al., 2015). Zhan and coworkers synthesized the dimeric BODIPY bridged with oligothiophene at the meso-positions (**94c**) (Liu et al., 2014). Compared with the single BODIPY cores **94a** and **94b**, the dimeric molecule **94c** showed improved packing order when blended with PC71BM and enhanced hole mobility, leading to a higher PCE of 3.13%. Mueller et al. revealed the BODIPY analogs of **95a** and **95b**, giving a PCE of 1.2 and 1.1% by using C60 as electron acceptor to prepare the vacuum-processed solar cells (Mueller et al., 2012). Similarly, Kraner et al. studied the influence of side groups on the OPV performance of BODIPY analogs, **95b**, **95c**, and **95d** (Kraner et al., 2015).

In order to further expand the absorption band of BODIPY, covalently combining it with other dye molecules emerged as an effective strategy. For example, the DPP unit was introduced to link with BODIPY, resulting into **96a** and **96b** (Cortizo-Lacalle et al., 2014). By using PC71BM as the acceptor, **96a** and **96b** showed a moderate PCE of 0.65 and 0.64%, respectively, due to the over-strong aggregating ability of these dye molecules, leading to poor film morphology. The porphyrin moiety was also selected to connect with BODIPY, producing **97**, which exhibited a competitive PCE of 5.29%

(Sharma et al., 2015). Moreover, the BODIPY was also utilized as building block for construction of conjugated polymers. By copolymerizing with acetylene and thiophene, conjugated copolymers of **98a** and **98b** can be obtained, showing low bandgap of 1.61 and 1.65 eV, respectively. By using PC61BM as acceptor, moderated PCEs around 2.0% can be obtained (Kim et al., 2010).

Except for using as donor materials, the BODIPY-based molecules are also qualified for the acceptor materials due to its strong electron affinity. BODIPY dimers bridged with BDT (**99a**), CPDT (**99b**), and DTP (**99c**) were synthesized to be used as electron-acceptor materials. By selecting P3HT as electron-donor material, fullerene-free devices based on these BODIPY dimers showed PCEs from 1.18 to 1.51% (Poe

et al., 2014). Zhan et al. reported the DPP bridged BODIPY dimers (**100**), exhibiting an competitive PCE of 2.84% by using PTB7-Th/p-DTS(FBTTh2)<sup>2</sup> (0.5:0.5) as donor (Liu W. et al., 2017).

Recently, a novel π-electronic unit containing N-B←N group, namely BNBP, with low-lying LUMO, bathochromic absorption, good co-planarity, and strong π-π interaction has been developed by Liu and coworkers. Owing to its strong electron-affinity, it's suitable to build conjugated copolymers for acceptor materials (**Figure 15**). **Table 3** summarizes the optoelectronic and photovoltaic parameters for BNBP-based materials. The primary attempt to copolymerize with thiophene produced a typical D-A copolymer **101a**, showing the HOMO and LUMO of −5.77 and −3.50 eV, respectively. By selecting PTB7 as donor materials, all-polymer solar cells were fabricated, affording a high Voc of 1.09 V and an impressive PCE of 3.38% (Dou et al., 2016). The high Voc was mainly originated from the large offset between HOMO of PTB7 and LUMO of **101a**. By selecting PCDTBT, a donor polymer with low-lying HOMO of −5.42 eV, to prepare all-polymer solar cells with **101a**, a recorded Voc of 1.3 V can be obtained (Ding et al., 2017). Using small molecule donor p-DTS(FBTTh2)<sup>2</sup> to match with **101a** also gave a Voc of 1.08 V and a PCE of 3.5% (Zhang Z. et al., 2017). Replacing the co-monomer from thiophene to selenophene led to the copolymer **101b**, which displayed depressed HOMO (−5.77 eV) and LUMO (−3.66 eV) compared to the values of **101a**. By using PTB7-Th as donor material, the all-polymer solar cells gave an improved PCE of 4.26%, which were interpreted by the enhanced driving force for the charge dissociation between PTB7-Th and **101b**, resulting from



the deeper LUMO of **101b** in contrast to the value of **101a** (Ding et al., 2016). Furthermore, when the 3,3′ -difluoro-2,2′ bithiophene (fBT) unit was utilized to copolymerize with BNBP, the resulting polymer **102** exhibited a high electron mobility of 2.4×10−<sup>4</sup> cm<sup>2</sup> /V•s due to the intra-molecular F. . . S interaction that locked the conformation, enhanced the co-planarity, and facilitate the ordered molecular packing. Therefore, all-polymer solar cells based on PTB7-Th/**102** afforded a recorded PCE of 6.26% for N-B←N based acceptor materials (Long et al., 2016a). In order to further broadened the absorption band to lower energy, the DPP unit was selected as co-monomer to obtain copolymer **103**, showing a small optical bandgap of 1.56 eV and high electron mobility of 2.1×10−<sup>4</sup> cm<sup>2</sup> /V•s. All-polymer solar cells based on PTB7/**103** gave a PCE of 2.69% (Long et al., 2016b).

Additionally, the electron-rich CPT unit was also copolymerized with BNBP leading to **104a**, showing the HOMO and LUMO of −5.64 and −3.45 eV, respectively, which was matched well with the energy levels of P3HT (HOMO/LUMO = −5.20/−3.20 eV). As such, all-polymer solar cells based on P3HT/**104a** were prepared, affording a moderate PCE of 1.76% with a high Voc of 1.01 V (Long et al., 2017a). When the alkyl groups on the CPT were replaced by F atoms, a new electron-rich unit, namely, 4,4-Difluoro-4Hcyclopenta[2,1-b:3,4-b′ ]dithiophene (fCPT) can be obtained. By copolymerizing with BNBP, the novel polymer **104b** exhibited depressed HOMO and LUMO in comparison with the values of **104a**. PTB7-Th was selected as donor material to afford a competitive PCE of 3.76% (Zhao et al., 2017b). Moreover, conjugated side groups were introduced to BNBP unit and copolymer **105a** and **105b** were produced. The conjugated side chains were considered to improve the electron mobility and a good PCE of 3.77 and 4.46% were demonstrated for **105a** and **105b**, respectively (Zhao et al., 2017a). Recently, they revealed an effective method to finely tune the HOMO and LUMO of BNBP-based polymers by varying the side chains, as evident by **106a**—**106h**, whose LUMOs were decreased gradually. **106a** and **106e** were suitable for the donor materials due to their high-lying energy levels whereas **106d** and **106h** were applicable for acceptor materials owing to their low-lying energy levels (Long et al., 2017b).

In general, these BNBP-based copolymers are suitable for acceptor materials due to the strong electron-affinity properties. Although several polymers based on BNBP were reported for


TABLE 3 | Optoelectronic and photovoltaic parameters for BNBP-based materials.

OPV applications, small molecules related to this unit are scarcely revealed to now. Very recently, a small-molecule acceptor (**107**) built by end-capping BNBP with IC unit was reported, exhibiting a low-lying LUMO of −3.93 eV and wide absorption band (Liu F. et al., 2017). By selecting PTB7-Th as donor material, OPV devices gave a Jsc of 14.62 mA/cm<sup>2</sup> , a Voc of 0.78 V, an FF of 62%, and a competitive PCE of 7.06%. This attempt points out the great potential of BNBP as a building block for small-molecule acceptor materials. We anticipate that the small molecules based on BNBP and its derivatives are also interesting and promising for non-fullerene acceptor materials. This field is blank to now and we strongly perceive that the small molecules based on novel π-electronic units containing N-B←N groups for OPV applications leave a large space and will be a research hot drawing great attention.

### PROSPECT

In summary, the OPV applications of BN embedded πconjugated electronic units are in infancy. From our point, the following aspects will presumably be the potential research interests concerning the BN perturbed π-conjugated units.

For B-N covalent bond embedded π-conjugated units, they are suitable to act as electron-rich units to construct electron-donor materials owing to their high-lying energy levels, good backbone co-planarity, and high hole mobility. However, their absorption bands should be further broadened to lower energy (600–800 nm) to enhance the light harvesting ability. On the one hand, novel π-units containing B-N bond with extended conjugation should be developed to further red-shift the absorption bands. On the other hand, linking the B-N embedded π-conjugated units with low bandgap electrondeficient units, e.g., PDI, NDI, DPP, IID, IC, TPD, and BT can be an effective method to decrease the optical bandgaps and shift the absorption bands to longer wavelength. To this end, fine-tuning the energy levels and optical absorption via D-A combination of B-N embedded π-units and electron-deficient units will be a systematical job.

For B←N and N-B←N embedded π-units, the low-lying energy levels, red-shifted absorption bands, and good electron mobility make them promising electron-deficient units for construction of acceptor materials. To now, most studies have been focusing on the BODIPY-based donor materials and a few novel units containing B←N or N-B←N, e.g., BNCTP and BNBP were used to construct polymer acceptor materials. However, the currently revealed structures are still limited in terms of electron affinity, conjugation degree and light absorption. Great research space remains in the development of novel fused π-units containing B←N or N-B←N groups with depressed energy levels and strong light-harvesting ability. Upon judicious molecular tailoring, these B←N or N-B←N embedded π-units possess high electron affinity and good rigidity, which may be comparable to those typical electron-deficient units, e.g., PDI, NDI, DPP, and IID. As such, small-molecule acceptors established with B←N or N-B←N embedded π-units will presumably be an important family of OPV materials, although few examples has been reported to now.

Challenges also exist in these structures, from synthesis routes, to material stability, and processability. The synthesis routes usually involve the usage of BX3, which is strong Lewis acid, toxic, and easily subjected to hydrolysis. Similarly, some of the BN embedded units are unstable in moisture and cannot be purified by typical chromatographic methods due to the Lewis acid property of B atom (Crossley et al., 2015; Zhu et al., 2016). Moreover, how to introduce bulk substituents into the backbone to ensure sufficient solution processability without sacrificing the molecular co-planarity and packing order is also a challenge. These challenges should be fully taken into account when designing BN embedded units for OPV applications. However, we expect that BN embedded units will draw great attention for the construction of OPV materials.

#### AUTHOR CONTRIBUTIONS

JH conceived, designed, and wrote the manuscript. YL retrieved the literature and edited sections of the manuscript. All authors approved it for publication.

#### REFERENCES


## ACKNOWLEDGMENTS

This work was financially supported by the National Natural Science Foundation of China (NSFC, No. 51603076) and Graphene Powder & Composite Research Center of Fujian Province (2017H2001). The Outstanding Youth Scientific Research Cultivation Plan of Colleges and Universities of Fujian Province and Promotion Program for Young and Middleaged Teacher in Science and Technology Research of Huaqiao University (ZQN-PY405) were appreciated.

small molecules with BODIPY end groups as novel donors for organic solar cells. Beilstein J. Org. Chem. 10, 2683–2695. doi: 10.3762/bjoc.10.283


modification of BDPPV backbones. J. Am. Chem. Soc. 137, 6979–6982. doi: 10.1021/jacs.5b00945


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Huang and Li. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

## Efficient Non-fullerene Organic Solar Cells Enabled by Sequential Fluorination of Small-Molecule Electron Acceptors

#### Ruihao Xie, Lei Ying\*, Hailong Liao, Zhongxin Chen, Fei Huang\* and Yong Cao

*State Key Laboratory of Luminescent Materials and Devices, Institute of Polymer Optoelectronic Materials and Devices, South China University of Technology, Guangzhou, China*

Three small-molecule non-fullerene electron acceptors containing different numbers of fluorine atoms in their end groups were designed and synthesized. All three acceptors were found to exhibit relatively narrow band gaps with absorption profiles extending into the near-infrared region. The fluorinated analog exhibited enhanced light-harvesting capabilities, which led to improved short-circuit current densities. Moreover, fluorination improved the blend film morphology and led to desirable phase separation that facilitated exciton dissociation and charge transport. As a result of these advantages, organic solar cells based on the non-fullerene acceptors exhibited clearly improved short-circuit current densities and power conversion efficiencies compared with the device based on the non-fluorinated acceptor. These results suggest that fluorination can be an effective approach for the molecular design of non-fullerene acceptors with near-infrared absorption for organic solar cells.

#### Edited by:

*Chuanlang Zhan, Institute of Chemistry (CAS), China*

#### Reviewed by:

*In Hwan Jung, Kookmin University, South Korea Chuluo Yang, Wuhan University, China*

#### \*Correspondence:

*Lei Ying msleiying@scut.edu.cn Fei Huang msfhuang@scut.edu.cn*

#### Specialty section:

*This article was submitted to Organic Chemistry, a section of the journal Frontiers in Chemistry*

Received: *13 May 2018* Accepted: *04 July 2018* Published: *26 July 2018*

#### Citation:

*Xie R, Ying L, Liao H, Chen Z, Huang F and Cao Y (2018) Efficient Non-fullerene Organic Solar Cells Enabled by Sequential Fluorination of Small-Molecule Electron Acceptors. Front. Chem. 6:303. doi: 10.3389/fchem.2018.00303* Keywords: organic solar cells, non-fullerene, small molecule electron acceptors, fluorination, near-infrared absorption

## INTRODUCTION

Bulk heterojunction organic solar cells (OSCs) are a promising technology for solar energy collection and have attracted much interest owing to their unique advantages for the fabrication of lightweight and flexible devices (Li et al., 2016b, 2017a, 2018a,b; Zhao et al., 2016; Zhang et al., 2017b, 2018; Cheng et al., 2018; Hou et al., 2018; Zhang, 2018). Over the past several years, although fullerene derivatives have been extensively used as electron-acceptor materials for OSCs, their various intrinsic limitations, such as poor absorption in the visible-light region, a difficult-to-adjust molecular structure and morphological instability, have impeded the further development of OSCs (He and Li, 2011). To circumvent this constraint, considerable progress has been achieved recently due to the development of non-fullerene acceptors (NFAs) for high-performance non-fullerene OSCs, as NFAs have high absorption coefficients and suitable frontier molecular orbital energy levels that facilitate both the harvesting of solar photons and charge separation (Bin et al., 2016; Du et al., 2017; Fan et al., 2017; Kan et al., 2017b; Xu et al., 2017; Cui et al., 2018; Gao et al., 2018; Luo et al., 2018; Zhang et al., 2018; Zhu et al., 2018).

Typically, to achieve high photovoltaic performance of OSCs based on novel NFAs, much effort has been devoted to the use of advanced device structures and sophisticated film-processing techniques (Li et al., 2016a, 2017c; Meng et al., 2016; Bao et al., 2017; Kan et al., 2017a; Zhang et al., 2017a; Wu et al., 2018). It is well established that the light-harvesting capability of OSCs plays a critical role in their photovoltaic performance, as their power conversion efficiencies can be enhanced by expanding the absorption spectrum of the photoactive layer in the near-infrared (NIR) region. Therefore, NFAs with absorption spectra extending into the NIR region have been explored, especially with respect to their potential applications in semitransparent organic photovoltaics and tandem OSCs (Li et al., 2017b; Yao et al., 2017, 2018). Furthermore, the fluorination of conjugated semiconductors has proved to be an effective synthetic strategy for developing efficient photoactive-layer materials for OSC applications. The introduction of fluorine atoms into small molecules or polymers can not only optimize their optical and electrical properties but also promote intermolecular interactions via the formation of non-covalent F···S and F···H bonds, resulting in enhanced charge mobility (Wang et al., 2013; Jo et al., 2015; Dai et al., 2017). More importantly, fluorination can be used to fine-tune the hydrophobicity and polarity of conjugated semiconductors, thus permitting control over the interfacial interactions in blend films (Pagliaro and Ciriminna, 2005). The combination of these advantages leads to improved film morphology with appropriate phase domains and larger interfacial areas, which facilitates exciton dissociation and charge transport and thus enhances the overall photovoltaic performance of OSCs.

In this work, we designed and synthesized a series of nonfullerene electron acceptors (BT-IC, BT-F, and BT-2F) with different numbers of fluorine atoms on their end groups. The strong intramolecular charge transfer between the electronrich cores and electron-deficient end groups of these acceptors was found to result in intense absorption in the NIR region. The sequential fluorination of the end groups not only enhanced the light-harvesting capabilities of BT-F and BT-2F but also simultaneously increased their electron mobilities, leading to a higher short-circuit current density (JSC). More importantly, the fluorinated acceptors exhibited more favorable phase separation after blending with a mediumband-gap conjugated polymer. The combination of these phenomena led to improved short-circuit current density and thus enhanced the photovoltaic performance of the resulting devices.

#### EXPERIMENTAL

#### Instrumentation

<sup>1</sup>H and <sup>13</sup>C NMR were characterized with Bruker-500 spectrometer in deuterated chloroform solution at 298 K. Chemical shifts were recorded as δ values (ppm) with the internal standard of tetramethylsilane (TMS). Mass spectra were collected on a MALDI Micro MX mass spectrometer, or an API QSTAR XL System. Number-average (Mn) and polydispersity index (PDI) were determined on a Polymer Laboratories PL-GPC 220 using 1,2,4-trichlorobenzene as eluent at 150◦C vs. polystyrene standards. Thermogravimetric analyses (TGA) were performed on a Netzsch TG 209 under nitrogen at a heating rate of 10◦C min−<sup>1</sup> . Differential scanning calorimetry (DSC) was performed on a Netzsch DSC 204 under nitrogen flow at heating/cooling rates of 10/10◦C min−<sup>1</sup> . The absorption coefficients of films are calculated by dividing the film thickness with the maximum absorption peak. The thin films with thickness of about 100 nm (measured by the profilometer) is spin-coated from chloroform solution on the top of quartz. Then the absorption spectra of these films were recorded by a HP 8453 spectrophotometer. Cyclic voltammetry (CV) was performed on a CHI600D electrochemical workstation with a glassy carbon working electrode and a Pt wire counter electrode at a scanning rate of 50 mV s−<sup>1</sup> against an Ag/Ag<sup>+</sup> reference electrode with a nitrogen saturated anhydrous solution of tetra-n-butylammonium hexafluorophosphate in acetonitrile (0.1 mol L−<sup>1</sup> ). Atomic force microscopy (AFM) measurements were carried out using a Digital Instrumental DI Multimode Nanoscope III in a taping mode. TEM images were characterized with a JEM-2100F instrument.

#### Photovoltaic Device Fabrication

The non-fullerene organic solar cells with a conventional device structure of ITO/PEDOT:PSS/active layer/PFN-Br/Ag were fabricated. Here PFN-Br represents poly[(9,9-bis(3′ -((N,Ndimethyl)-N-ethylammonium)-propyl)-2,7-fluorene)-alt-2,7- (9,9-dioctylfluorene)] dibromide, which functioned as the cathode interlayer to facilitate electron extraction from the active layer. Before fabrication of the device, the indium tin oxide (ITO)-coated glass substrates were cleaned by ultrasonic treatment in deionized water, acetone, isopropyl alcohol, and dried in oven at 80◦C for 12 h before used. After PEDOT:PSS (30 nm) layer was spin coated onto the substrate, and dried at 150◦C for 15 min in air. Then, the ITO substrates were transferred into a nitrogen protected glovebox where the H2O concentration is ≤ 0.5 ppm and O<sup>2</sup> concentration is ≤ 20 ppm. The thin film of active layer was spin-coated from a solution of PTZPF:non-fullerene acceptor blend in chlorobenzene. A thin PFN-Br layer (5 nm) was then spin coated onto the active layer as the cathode interface layer. The substrates were then transferred to a vacuum thermal evaporator, followed by deposition of the Ag cathode at a pressure of 2 × 10−<sup>7</sup> Torr through a shadow mask. Before the J-V test, a physical mask with an aperture with precise area of 0.04 cm<sup>2</sup> was used to define the device area. The J-V curves were measured on a computer-controlled Keithley 2,400 source meter under 1 sun, the AM 1.5 G spectra came from a class solar simulator (Enlitech, Taiwan), and the light intensity was 100 mWcm−<sup>2</sup> as calibrated by a China General Certification Center-certified reference monocrystal silicon cell (Enlitech). The external quantum efficiency (EQE) spectra measurements were performed on a commercial QE measurement system (QE-R3011, Enlitech).

#### Materials

The monomers of thieno[3′ ,2′ :4,5] cyclopenta[1,2-b] thieno[2′′ , 3 ′′:3′ ,4′ ] cyclopenta[1′ ,2′ :4,5] thieno[2,3-f][1] benzothiophene-2,8-dicarboxaldehyde, 5,11-bis[(2-ethylhexyl)oxy] -4,4,10,10 tetrakis(4-hexylphenyl)-4,10-dihydro **(1)** were synthesized according to the reported procedures (Li et al., 2017b). And the donor polymer PTZPF was synthesized via Stille polymerization (Scheme S1, supporting information, SI, with molecular structure shown in **Figure 1A**). 2-(3-Oxo-2,3-dihydro-1Hinden-1-ylidene)malononitrile (**2**), a mixture of 2-(5-fluoro-3 oxo-2,3-dihydro-1H-inden-1-ylidene)malononitrile and 2-(6 fluoro-3-oxo-2,3-dihydro-1H-inden-1-ylidene)malononitrile (**3**) and 2-(5,6-difluoro-3-oxo-2,3-dihydro-1H-inden-1-ylidene) malononitrile (**4**) were obtained from commercial sources and used without further purification. The small-molecule acceptors were prepared as the following procedures as below.

#### Synthesis of BT-IC

2-(3-Oxo-2,3-dihydro-1H-inden-1-ylidene)malononitrile (**2**) (194.2 mg, 1.0 mmol) was added into the mixture of thieno[3′ , 2 ′ :4,5] cyclopenta[1,2-b] thieno[2′′,3′′:3′ ,4′ ] cyclopenta[1′ ,2′ : 4,5] thieno[2,3-f][1] benzothiophene-2,8-dicarboxaldehyde, 5,11-bis[(2-ethylhexyl)oxy] -4,4,10,10-tetrakis(4-hexylphenyl)- 4,10-dihydro (**1**) (133.2 mg, 0.1 mmol) in chloroform (50 mL) with pyridine (1 mL). The reactant was refluxed for 6 h under nitrogen atmosphere. After cooling to room temperature, the reactant was poured into methanol and the precipitate was filtered off. The crude product was purified by silica gel using a mixture of hexane/dichloromethane as the eluent to give a blue black powder (121.2 mg, 73%). <sup>1</sup>H NMR (400 MHz, CDCl3, δ ): 8.79 (s, 2H), 8.64 (m, 2H), 7.87 (m, 2H), 7.70 (m, 4H), 7.48 (s, 2H), 7.31 (m, 8H), 7.08 (m, 8H), 3.48 (t, 4H), 2.57 (t, 8H), 1.60-1.53 (m, 2H), 1.35-1.28 (m, 48H), 0.96 (t, 6H), 0.87 (m, 18H). <sup>13</sup>C NMR (100 MHz, CDCl3, δ): 188.56, 164.46, 160.31, 157.20, 153.89, 146.22, 142.16, 142.14, 140.67, 139.93, 138.69, 138.45, 138.23, 136.81, 135.89, 135.01, 134.30, 128.51, 128.50, 128.31, 125.25, 123.61, 121.32, 114.81, 68.40, 63.94, 39.38, 35.56, 31.71, 31.26, 31.24, 29.54, 29.21, 28.77, 23.34, 22.68, 22.59, 14.21, 14.10, 10.78. MS (MALDI-TOF) calcd for C110H114N4O4S4, 1684.386; found, 1683.622.

#### Synthesis of BT-F

A similar procedure was followed as that described for **BT-IC**, a mixture of 2-(5-fluoro-3-oxo-2,3-dihydro-1H-inden-1-ylidene) malononitrile and 2-(6-fluoro-3-oxo-2,3-dihydro-1H-inden-1 ylidene)malononitrile (**3**) (212.2 mg, 1.0 mmol) and thieno [3′ ,2′ :4,5] cyclopenta[1,2-b] thieno[2′′,3′′:3′ ,4′ ] cyclopenta[1′ , 2 ′ :4,5] thieno[2,3-f] [1]benzothiophene-2,8-dicarboxaldehyde, 5,11-bis[(2-ethylhexyl)oxy] -4,4,10,10-tetrakis(4-hexylphenyl)-4, 10-dihydro (**1**) (133.2 mg, 0.1 mmol) were used. **BT-F** was obtained as a blue black solid (149.2 mg, 88.0 %). <sup>1</sup>H NMR (400 MHz, CDCl3, δ): 8.79 (s, 2H), 8.68 (m, 0.5H), 8.36 (m, 1.5H),7.88 (m, 1.5H), 7.52 (m, 2.5H), 7.40 (m, 2H), 7.30 (m, 8H), 7.08 (m, 8H), 3.48 (m, 4H), 2.53 (m, 8H), 1.60-1.54 (m, 2H), 1.35-1.28 (m, 48H), 0.96 (t, 6H), 0.87 (m, 18H). <sup>13</sup>C NMR (100 MHz, CDCl3, δ): 187.10, 167.69, 165.65, 165.19, 164.64, 164.61, 159.23, 158.89, 157.55, 157.50, 154.55, 154.39, 146.29, 142.28, 142.22, 142.19, 140.66, 140.62, 139.94, 139.88, 138.74, 138.69, 138.42, 138.29, 135.97, 135.79, 133.01, 128.50, 128.48, 128.39, 128.33, 127.76, 125.75, 125.67, 122.13, 121.67, 121.48, 121.15, 121.08, 114.81, 114.65, 114.50, 114.36, 112.89, 112.68, 110.75, 76.77, 69.09, 68.18, 63.95, 39.37, 35.55, 31.70, 31.25, 31.23, 29.56, 29.22, 28.78, 23.38, 22.71, 22.60, 14.22, 14.11, 10.80. MS (MALDI-TOF) calcd for C110H112F2N4O4S4, 1720.367; found, 1719.595.

#### Synthesis of BT-2F

A similar procedure was followed as that described for **BT-IC**, 2-(5,6-difluoro-3-oxo-2,3-dihydro-1H-inden-1-ylidene) malononitrile (**4**) (230.2 mg, 1.0 mmol) and thieno[3′ ,2′ :4,5] cyclopenta[1,2-b] thieno[2′′,3′′:3′ ,4′ ] cyclopenta[1′ ,2′ :4,5] thieno [2,3-f][1] benzothiophene-2,8-dicarboxaldehyde, 5,11-bis[(2 ethylhexyl)oxy] -4,4,10,10-tetrakis(4-hexylphenyl)-4,10-dihydro (**1**) (133.2 mg, 0.1 mmol) were used. **BT-2F** was obtained as a black solid (138.6 mg, 80.0 %). <sup>1</sup>H NMR (400 MHz, CDCl3, δ): 8.78 (s, 2H), 8.52 (m, 2H), 7.64 (t, 2H), 7.50 (s, 2H), 7.31 (dd, 8H), 7.08 (d, 8H), 3.49 (t, 4H), 2.57 (t, 8H), 1.60-1.54 (m, 2H), 1.36-1.28 (m, 48H), 0.96 (t, 6H), 0.86 (m, 18H). <sup>13</sup>C NMR (100 MHz, CDCl3, δ): 186.22, 164.72, 160.58, 158.12, 155.49, 146.38, 142.47, 142.28, 140.62, 139.08, 138.95, 138.75, 138.35, 136.53, 136.03, 135.70, 134.62, 128.56, 128.48, 128.36, 120.53, 115.13, 114.39, 112.46, 68.67, 63.78, 39.40, 35.56, 31.72, 31.26, 31.24, 29.54, 29.21, 28.77, 23.34, 22.68, 22.59, 14.21, 14.10, 10.76. MS (MALDI-TOF) calcd for C110H110F4N4O4S4, 1756.348; found, 1755.581.

## RESULTS AND DISCUSSION

#### Synthesis and Characterization

The synthesis of the target compounds BT-IC, BT-F, and BT-2F is outlined in **Scheme 1**. These small-molecule acceptors were prepared via Knoevenagel condensation between thieno[3′ ,2′ : 4,5] cyclopenta[1,2-b] thieno[2′′,3′′:3′ ,4′ ] cyclopenta[1′ ,2′ :4,5] thieno[2,3-f][1] benzothiophene-2,8-dicarboxaldehyde, 5,11-bis [(2-ethylhexyl)oxy] -4,4,10,10-tetrakis(4-hexylphenyl)-4,10 dihydro (**1**) and 2-(3-oxo-2,3-dihydro-1H-inden-1-ylidene) malononitrile (**2**) or its fluorinated derivatives at 60◦C in the presence of a catalytic amount of pyridine. It is worth noting that the monofluorinated compound **3** consisted of two regioisomers, namely, 2-(5-fluoro-3-oxo-2,3-dihydro-1H-inden-1-ylidene)malononitrile and 2-(6-fluoro-3-oxo-2,3-dihydro-1H-inden-1-ylidene)malononitrile. As these two isomers have very similar molecular structures and polarities, they could not be separated, and thus the resulting BT-F was obtained as a mixture of isomers. All three acceptors exhibited good solubility in typical organic solvents, such as chloroform, chlorobenzene and ortho-dichlorobenzene, at room temperature. The chemical structures of the three acceptors were confirmed by nuclear magnetic resonance spectroscopy and mass spectrometry (Figures S7–S12).

The thermal properties of these resulting NFAs were evaluated by thermogravimetric analysis and differential scanning calorimetry (DSC) under a nitrogen atmosphere (Figures S1, S2,). All of these NFAs exhibited excellent thermal stabilities with onset decomposition temperatures (Td) higher than 310◦C. The DSC curves were obtained by heating from 30 to 250◦C in the second heating/cooling cycle. It was found that BT-IC exhibited a melting peak at 127◦C, whereas no phase-transition signals were observed during the DSC measurements of the other two materials.

## Optical, Electrochemical, and Electron-Transport Properties

**Figure 1B** shows the UV–vis absorption spectra of thin films of the donor polymer and acceptor molecules. All three NFAs showed similar absorption profile cut-offs in the NIR region (up to 866 nm) in the solid state. Such absorption profiles are complementary with a medium band gap conjugated polymer, namely PTZPF, which has the absorption onset of 620 nm (**Figure 1D**). Note that the complementary absorption is beneficial for the harvesting of solar photons to achieve a high

PTZPF:NFA blend films.


TABLE 1 | Optophysical and electrochemical properties of active layer materials.

*<sup>a</sup>Calculated from the onset of UV-vis absorption as pristine thin films; <sup>b</sup>EHOMO* = *–e (Eox* + *4.50) (eV); <sup>c</sup>ELUMO* = *–e (Ered* + *4.50).*

short-circuit current density. The fluorinated small molecules BT-F and BT-2F exhibited slightly red-shifted absorption edges compared with BT-IC. An optical band gap of 1.43 eV was obtained for BT-IC, which slightly decreased to 1.42 and 1.41 eV for BT-F and BT-2F, respectively. Importantly, the absorption coefficients were slightly enhanced from 1.05 × 10<sup>5</sup> cm−<sup>1</sup> for BT-IC to 1.12 × 10<sup>5</sup> cm−<sup>1</sup> and 1.22 × 10<sup>5</sup> cm−<sup>1</sup> for BT-F and BT-2F, respectively (**Figure 1B**), indicating that the introduction of fluorine atoms into these acceptors enhanced their light-harvesting ability through improved intermolecular interactions (Yang et al., 2013; Wolf et al., 2015).

The electrochemical properties and energy levels of the polymer acceptors were investigated by cyclic voltammetry. Here we used the potential ferrocene/ferrocenium (Fc/Fc+) redox couple as the standard. Under the current measurement conditions, the potential of Fc/Fc<sup>+</sup> couple was measured as 0.30 V regarding to the reference electrode. Assuming that the Fc/Fc<sup>+</sup> redox couple has an absolute potential of −4.80 V to vacuum, the highest occupied molecular orbital energy levels (EHOMO) is calculated as EHOMO = –e (Eox + 4.80 – 0.30) (eV), and the lowest unoccupied molecular orbital energy levels (ELUMO) is calculated as ELUMO = –e (Ered + 4.80 – 0.30) (eV). The energy level diagrams of all of the materials are depicted in **Figure 1C** and the corresponding electrochemical data are summarized in **Table 1**. The calculated ELUMO/EHOMO values of BT-IC, BT-F and BT-2F were −3.88/−5.55 eV, −3.97/−5.57 eV and −4.00/−5.60 eV, respectively. Both the ELUMO and EHOMO levels of these acceptor molecules gradually decreased with the increasing number of fluorine atoms, indicating that fluorination of the end groups of small-molecule NFAs can effectively decrease the ELUMO and EHOMO levels owing to the strongly electron-withdrawing characteristics of fluorine atoms (Dutta et al., 2014). The ELUMO and EHOMO levels of the polymer donor PTZPF were −3.42 and −5.41 eV, respectively, which ensures an adequate driving force for efficient exciton dissociation (Scharber et al., 2010). The charge carrier mobilities of the pure films of acceptors were measured by the space-charge-limited current (SCLC) method using the Mott–Gurney equation (Figure S5). The measurements are carried out by fabricating electron-only devices with architecture of ITO/Ag/active layer/Ag structure. The pure film based on BT-2F exhibited the electron mobility (µe) of 9.64 × 10−<sup>5</sup> cm<sup>2</sup> V −1 s −1 , which is higher than those of BT-F (µ<sup>e</sup> = 5.53 × 10−<sup>5</sup> cm<sup>2</sup> V −1 s −1 ) and BT-IC (µ<sup>e</sup> = 2.28 × 10−<sup>5</sup> cm<sup>2</sup> V −1 s −1 ).

#### Photovoltaic Performances

To elucidate the effects of fluorination on the photovoltaic properties, OSC devices were fabricated using PTZPF as the electron-donor material and BT-IC, BT-F, or BT-2F as the electron-acceptor material. The devices were fabricated with the conventional configuration of ITO/PEDOT:PSS/active layer/PFN-Br/Ag, and the device performances were measured under simulated AM 1.5 G illumination at 100 mW cm−<sup>2</sup> . Poly[(9,9-bis(3'-((N,N-dimethyl)-N-ethylammonium)propyl)-

2,7-fluorene)-alt-2,7-(9,9-dioctylfluorene)] dibromide (PFN-Br) was used as the cathode interfacial layer to facilitate charge carrier collection (Zhang et al., 2017c). The initial optimisation of device performance was carried out by screening the weight ratios of the donor:acceptor (D:A) blend, film thickness of the photoactive layers and the effects of additives to the processing solvents (Figure S6 and Table S1). All of the photoactive layers of the devices were processed under the optimized conditions, which consisted of a D:A weight ratio of 1:1, spin casting of the films from chlorobenzene containing 0.5 v/v% of 1-chloronaphthalene (CN) as additive, and annealing of the fabricated films at 120◦C for 10 min. The current density–voltage (J–V) curves are presented in **Figure 2A** and the corresponding data are summarized in **Table 2**.

Interestingly, the photovoltaic parameters of the resulting devices were strongly dependent on the number of fluorine substituents. The device based on BT-IC as the acceptor exhibited a moderate power conversion efficiency (PCE) of 5.63%, with an open-circuit voltage (VOC) of 0.93 V, a JSC of 12.27 mA cm−<sup>2</sup> and a fill factor (FF) of 49.0%. In contrast, the devices based on the fluorinated acceptors BT-F and BT-2F exhibited clearly enhanced PCE values of 7.27% (VOC = 0.88 V, JSC = 16.64 mA cm−<sup>2</sup> and FF = 49.0%) and 8.54% (VOC = 0.84 V, JSC = 19.29 mA cm−<sup>2</sup> and FF = 53.0%), respectively. It should be noted that despite the decrease in the VOC of the resulting devices upon the incorporation of fluorine substituents into the acceptors, which is consistent with the down-shifted ELUMO levels observed for BT-F and BT-2F (Brabec et al., 2010; Kang et al., 2012), the JSC values were dramatically enhanced. The combination of these effects led to the clearly enhanced PCE values of the devices based on the fluorinated acceptors.

To investigate the obvious enhancement of JSC, we analyzed the absorption of PTZPF:BT-IC, PTZPF:BT-F and PTZPF:BT-2F blend films (**Figure 1D**). Similar to the absorption of neat films of BT-IC, BT-F, or BT-2F, the absorption coefficients of the PTZPF:BT-F and PTZPF:BT-2F blend films were both slightly

TABLE 2 | Photovoltaic parameters of OSCs measured under AM1.5 Illumination at 100 mW cm−<sup>2</sup> .


*<sup>a</sup>All of the blend films are processed by CB with 0.5 vol % CN and treated with 120*◦*C for 10 min; <sup>b</sup>Obtained from J–V measurements; <sup>c</sup>Obtained from the integration of EQE spectra; <sup>d</sup>Average values across more than 6 devices. Device structure: ITO/PEDOT:PSS/active layer/PFN-Br/Ag.*

higher than that of the PTZPF:BT-IC blend film, which can be correlated to the improved JSC of the devices based on fluorinated acceptors. Furthermore, to confirm the accuracy of the observed JSC, we measured the external quantum efficiencies (EQEs) of the devices. It should be noted that the calculated JSC values from the EQE spectra (**Figure 2B**) matched well with the JSC values obtained from the J–V curves. The device based on BT-2F exhibited a stronger photocurrent response from 400 to 870 nm, with a maximum EQE of 75%, which was higher than those observed for the devices based on BT-F and BT-IC (**Figure 2B**).

## Charge Generation, Transport, and Recombination

To study the charge generation process in the resulting bulkheterojunction films, we measured the photoluminescence (PL) spectra of the neat and D:A blend films. The peak emission of the pure PTZPF film was located at 640 nm upon excitation at 500 nm, whereas the acceptor molecules BT-IC, BT-F, and BT-2F exhibited similar emission peaks at ∼845 nm upon excitation at 720 nm. As shown in **Figure 3A**, the strong emission peak of BT-IC was clearly observed in the PTZPF:BT-IC blend film, indicating the low charge separation efficiency of the device based on BT-IC. In contrast, the PL emission of the neat films was effectively quenched in the PTZPF:BT-F and PTZPF:BT-2F blend films, indicating that exciton dissociation and charge transfer were remarkably enhanced by the introduction of fluorine atoms into the acceptor moiety. A similar phenomenon can be observed in **Figure 3B**, where the PL of BT-2F was quenched by 92.7% in the blend film, which was more pronounced than the quenching observed for the blend films based on BT-F (90.9%) or BT-IT (86.4%).

The JSC and VOC values of the devices were measured as a function of the light intensity (Plight) to elucidate the charge recombination dynamics in the photoactive layer, as shown in **Figures 3C**,**D**, respectively. For organic solar cells, the powerlaw dependence of JSC on the illumination intensity can generally be expressed as JSC <sup>∝</sup>(Plight) S , where S is the exponential factor, which is close to unity when the bimolecular recombination in the device is weak (Kyaw et al., 2013; Lu et al., 2015). The extracted values of S were 1.058, 1.060 and 1.043 for the devices based on BT-IC, BT-F, and BT-2F, respectively, all of which were close to unity, indicating the weak bimolecular recombination in these devices (Yang et al., 2016). Based on the VOC-Plight plot, VOC was plotted against the natural logarithm of Plight and the slope of nkT/q was calculated, where an n value of unity implies predominantly bimolecular recombination and an enhanced dependence of VOC on Plight (2kT/q) indicates trap-assisted monomolecular recombination (Gasparini et al., 2016). The calculated slopes were 1.32, 1.23, and 1.21 kT/q for the devices based on BT-IC, BT-F and BT-2F, respectively. The smaller slope value for BT-2F than the others suggests less trap-assisted recombination, thus resulting in a higher FF value.

#### Film Morphology

Tapping-mode atomic force microscopy (AFM) and transmission electron microscopy (TEM) measurements were performed to determine the influence of fluorination on the film morphology. **Figures 4a–c** shows topographical AFM images of active layers of the different NFAs. The PTZPF:BT-IC blend film contained large granular aggregates across the entire film with

PTZPF:BT-2F.

a root-mean-square (RMS) roughness of 6.32 nm (**Figure 4a**), whereas the blend films gradually became smoother as the number of fluorine atoms increased, with RMS roughness values of 3.50 nm and 2.19 nm for BT-F and BT-2F, respectively, suggesting that the incorporation of fluorine atoms led to a smoother film morphology. The phase separation of the blend films with different electron acceptors was also readily apparent from the TEM images. **Figure 4d** shows that the PTZPF:BT-IC blend film exhibited large-scale phase-separation features across the entire film, which is unfavorable for charge transfer at the donor–acceptor interface. Interestingly, the degree of phase separation of the blend films gradually decreased as the number of fluorine atoms increased. Consequently, the PTZPF:BT-2F blend film exhibited a smoother surface morphology with favorable phase separation (**Figure 4f**), which induced desirable exciton dissociation and thus simultaneously enhanced the JSC and FF values.

#### CONCLUSIONS

In summary, three NIR-absorbing electron acceptors containing different numbers of fluorine atoms were designed and synthesized. The results revealed that the fluorinated acceptors outperformed their non-fluorinated counterpart BT-IC. Sequentially increasing the number of fluorine atoms on the end groups of the acceptor molecules led to a dramatic improvement in the JSC of the resulting photovoltaic devices. Non-fullerene OSCs based on the fluorinated acceptor BT-2F exhibited an improved PCE of 8.54% with a high JSC of 19.29 mA cm−<sup>2</sup> , regarding to the device based on BT-IC (PCE = 5.63%, JSC = 12.27 mA cm−<sup>2</sup> ) that does not contain fluorine atom. The improved photovoltaic performances of devices based on fluorinated acceptors can be correlated to the broad absorption profile extending into the NIR

### REFERENCES


region, favorable film morphology and efficient charge transfer. These results demonstrate that fluorination can be an effective technique in the design of efficient electron-acceptor materials.

#### AUTHOR CONTRIBUTIONS

RX, LY and FH conceived the ideas and coordinated the work. RX and LY designed the donor polymer and the acceptor molecules. RX synthesized the polymer PTZPF and conducted the DSC, TGA, UV-vis, PL and cyclic voltammetric measurements. HL performed the device fabrication, the light intensity-dependent J-V characterization, and analyzed the data. ZC synthesized the acceptor molecules of BT-IC, BT-F, and BT-2F. RX and HL conducted the AFM and TEM measurements. RX, LY, FH, and YC contributed to manuscript preparation. All authors commented on the manuscript and approved for submission.

#### FUNDING

This work was financially supported by the Ministry of Science and Technology (No. 2014CB643501), the National Natural Science Foundation of China (No. 91633301, 21490573, 51673069), the Natural Science Fund of Guangdong (No. 2017A030306011 and 2015A030313229), and the Science and Technology Program of Guangzhou, China (No. 201710010021, 201707020019 and 2017A050503002).

#### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem. 2018.00303/full#supplementary-material


any processing additives or post-treatments. J. Am. Chem. Soc. 135, 17060–17068. doi: 10.1021/ja409881g


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Xie, Ying, Liao, Chen, Huang and Cao. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

## Two Novel Small Molecule Donors and the Applications in Bulk-Heterojunction Solar Cells

Xin Qi <sup>1</sup> , Yuan-Chih Lo<sup>2</sup> , Yifan Zhao<sup>1</sup> , Liyang Xuan<sup>1</sup> , Hao-Chun Ting<sup>2</sup> , Ken-Tsung Wong<sup>2</sup> \*, Mostafizur Rahaman<sup>3</sup> , Zhijian Chen1,4, Lixin Xiao1,4 and Bo Qu1,4 \*

<sup>1</sup> State Key Laboratory for Artificial Microstructures and Mesoscopic Physics, Department of Physics, Peking University, Beijing, China, <sup>2</sup> Department of Chemistry, National Taiwan University, Taipei, Taiwan, <sup>3</sup> Department of Chemistry, King Saud University, Riyadh, Saudi Arabia, <sup>4</sup> New Display Device and System Integration Collaborative Innovation Center of the West Coast of the Taiwan Strait, Fuzhou, China

Two novel small molecules DTRDTQX and DTIDTQX, based on ditolylaminothienyl group as donor moiety and quinoxaline as middle acceptor moiety with different terminal acceptor groups were synthesized and characterized in this work. In order to study the photovoltaic properties of DTRDTQX and DTIDTQX, bulk-heterojunction solar cells with the configuration of FTO/c-TiO2/DTRDTQX(or DTIDTQX):C70/MoO3/Ag were fabricated, in which DTRDTQX and DTIDTQX acted as the donors and neat C<sup>70</sup> as the acceptor. When the weight ratio of DTRDTQX:C<sup>70</sup> reached 1:2 and the active layer was annealed at 100◦C, the optimal device was realized with the power conversion efficiency (PCE) of 1.44%. As to DTIDTQX:C70-based devices, the highest PCE of 1.70% was achieved with the optimal blend ratio (DTIDTQX:C<sup>70</sup> = 1:2) and 100◦C thermal annealing treatment. All the experimental data indicated that DTRDTQX and DTIDTQX could be employed as potential donor candidates for organic solar cell applications.

Keywords: bulk-heterojunction, small molecule, donor, solar cell, ditolylaminothienyl, quinoxaline

## INTRODUCTION

Recently, organic solar cells (OSCs) based on bulk-heterojunction structure have attracted much attention due to the distinctive characteristics of low cost, easy fabrication, flexibility and light weight, etc. (Gustafsson et al., 1992; Shaheen et al., 2001; Chen and Cao, 2009). Compared with polymers employed in solar cells, small molecule donors have the advantage of less batch-to-batch variation, well-defined molecular structure, easier purification, etc. (You et al., 2013; Chen et al., 2014, 2015; He et al., 2015; Zhou et al., 2015). Therefore, much work focused on small molecule donors and the photovoltaic performance of OSCs was improved accordingly (Sun et al., 2011; Liu et al., 2013; Love et al., 2013; Coughlin et al., 2014). In general, the active layers of the solar cells consisted of small molecule donors and fullerene/fullerene derivative acceptors (Chen et al., 2012; Huang et al., 2016). In order to optimize the photovoltaic characteristics of OSCs, narrow band-gap and deep highest occupied molecular orbital (HOMO) of small molecule donors should be considered, which resulted in broad absorption and high open-circuit voltage (Voc) of devices. Then, various small molecules composed of electron rich moieties (donor, "D") and electron deficient moieties (acceptor, "A"), have been reported with the molecular configuration such as D-A (Roquet et al., 2006), A-D-A (Schulze et al., 2006), D-A-A (Lin et al., 2011) and D-A-D conjugated structures. In this regard, the HOMO and lowest unoccupied molecular orbital (LUMO) of the small molecules were effectively tuned, mainly due to the intramolecular charge transfer (ICT) between donors and acceptors (Zhang et al., 2011).

#### Edited by:

Chuanlang Zhan, Institute of Chemistry (CAS), China

#### Reviewed by:

Xiaozhang Zhu, Institute of Chemistry (CAS), China Daniel Glossman-Mitnik, Centro de Investigación en Materiales Avanzados, Mexico

#### \*Correspondence:

Ken-Tsung Wong kenwong@ntu.edu.tw Bo Qu bqu@pku.edu.cn

#### Specialty section:

This article was submitted to Organic Chemistry, a section of the journal Frontiers in Chemistry

Received: 09 March 2018 Accepted: 11 June 2018 Published: 02 July 2018

#### Citation:

Qi X, Lo Y-C, Zhao Y, Xuan L, Ting H-C, Wong K-T, Rahaman M, Chen Z, Xiao L and Qu B (2018) Two Novel Small Molecule Donors and the Applications in Bulk-Heterojunction Solar Cells. Front. Chem. 6:260. doi: 10.3389/fchem.2018.00260

Herein, the photovoltaic properties of two novel small molecule donors (named **DTRDTQX** and **DTIDTQX**, **Figure 1**) based on D-A-A structure were studied in this work. **DTIDTQX** or **DTRDTQX** consisted of ditolylaminothienyl group as the donor moiety, quinoxaline as middle acceptor moiety with different terminal acceptor groups such as 1,3-indandione or 3-ethylrhodanine, respectively. To investigate the photovoltaic properties of the small molecules, bulk-heterojunction (BHJ) solar cells based on **DTRDTQX** or **DTIDTQX** as the donor together with C<sup>70</sup> as the acceptor were fabricated and the optimal cells showed PCE of 1.44 and 1.70%, respectively.

## EXPERIMENTAL

### Materials and Characterization

All materials in this work were purchased commercially, except for the tailor made **DTRDTQX** and **DTIDTQX** donors. The commercial materials were used without further purification.

**Scheme 1** depicts the synthesis of **DTIDTQX** and **DTRDTQX**. By following the protocols established by Krebs et al. (Jorgensen and Krebs, 2005) and Janssen et al. (Bijleveld et al., 2009), we could get 4-bromo-7-methyl-2,1,3-benzothiadiazole (**3**). Then the hetereocyclic **3** was converted to diamine intermediate **4** by treating Fe/HCl, which was then followed by condensation with glyoxal to afford 5-bromo-8-methylquinoxaline (**5**) without further purification. The 8-bromoquinoxaline-5-carbaldehyde (**7**) was synthesized by benzylic bromination with N-bromosuccinimide (NBS) initiated by azobisisobutyronitrile (AIBN) and followed by hydrolysis with CaCO<sup>3</sup> in H2O/acetonitrile (Lin et al., 2011). Aldehyde **7** was reacted with N,N-di-p-tolyl-5-(tri-n-butylstannyl) thiophen-2-amine (**8**) through Stille coupling reaction and gave key intermediate **9**. Finally, the condensation of **9**

with 1,3-indandione and 3-ethylrhodanine via Knöevenagel reaction afforded **DTIDTQX** and **DTRDTQX**, respectively. The absorption spectra were measured with JASCO V-670 spectrophotometer. Themogravimetric analysis (TGA) was determined on a TA Instruments Model TGA Q500 V20.13 (build 39) with a heating rate of 10◦C/min. Differential Scanning Calorimeter (DSC) was carried out at a heating rate of 10◦C/min on a TA Instruments Model DSC Q100 V9.9 (build 303). The thickness of the films was evaluated using a surface profilometer. The electrochemical cyclic voltammetry (CV) was recorded by a CHI619B potentiostat with glassy carbon electrode, Pt wire and Ag/AgCl which were used as the working electrode, counter electrode, and reference electrode, respectively, further calibrated with the ferrocene/ferrocenium (Fc/Fc+) redox couple. The oxidation waves were recorded in CH2Cl<sup>2</sup> (for 1.0 mM) with 0.1 M tetrabutylammonium hexafluorophosphate ( <sup>n</sup>BuNPF6) as supporting electrolyte, while reductive waves were recorded in THF (for 1.0 mM) with 0.1 M tetrabutylammonium perchlorate (nBuNClO4) as supporting electrolyte.

#### Solar Cell Fabrication and Characterization

In order to investigate the photovoltaic properties of **DTRDTQX** and **DTIDTQX**, the OSCs with the configuration of FTO/c-TiO2/**DTRDTQX**(or **DTIDTQX**):C70/MoO3/Ag were fabricated as shown in **Figure 1**. The compact TiO<sup>2</sup> layer in OSCs acted as the electron transporting layer (Heo et al., 2015) and MoO<sup>3</sup> as the hole buffer layer. As to the photoactive layers, **DTRDTQX** and **DTIDTQX** served as the donors and C<sup>70</sup> as the acceptor, respectively. The FTO cathode was pre-cleaned in an ultrasonic cleaner with deionized water, acetone and alcohol for 15 min respectively and then treated with oxygen plasma for 15 min. The TiO<sup>2</sup> films were fabricated according to the literatures (Kim et al., 2012; Zhang et al., 2016) and sintered at 500◦C for 15 min in a muffle furnace. And then, the TiO<sup>2</sup> films were naturally cooled to room temperature. Blended solutions (total concentration: 20 mg/ml) of **DTRDTQX**(or **DTIDTQX**):C<sup>70</sup> in ortho-dichlorobenzene (oDCB) were spin-coated (700 rpm, 18 s) onto FTO/TiO<sup>2</sup> substrates in a glove box and then thermal annealed at 100◦C or 150◦C. The effect of thermal annealing on the photovoltaic properties of the active layers was also studied in this work. Finally, 7 nm MoO<sup>3</sup> buffer layers and 100 nm Ag anodes were thermal evaporated successively below 10−<sup>6</sup> Torr. The photovoltaic performance of the OSCs were evaluated by current density-bias voltage (J-V) measurement (using a Keithley 2400 source meter) under AM 1.5G simulated solar illumination (Newport model 94021A, 100 mW cm−<sup>2</sup> ).

## RESULTS AND DISCUSSION

#### Thermal Property

Thermal properties of the two small molecules were investigated by TGA measurement as shown in **Figure 2** and the thermal decomposition temperatures (Td, 5% weight loss) were evaluated to be 362◦C and 312◦C for DTRDTQX and DTIDTQX respectively, indicating the good thermal stability of the small molecules. According to the DSC plots shown in **Figure 3**, the melting temperatures (Tm) were evaluated to be 187.8◦C and

263.3◦C for DTRDTQX and DTIDTQX, respectively. Moreover, the glass transition temperatures (Tg) were measured to be 94.0◦C and 149.7◦C for DTRDTQX and DTIDTQX, respectively. Therefore, both DTRDTQX and DTIDTQX were stable donors for OSCs due to their decent thermal stability.

### Absorption Properties

The UV-Vis absorption of **DTIDTQX** and **DTRDTQX** in CH2Cl<sup>2</sup> were shown in **Figure 4** and the corresponding data were summarized in **Table 1**. The compounds showed broad band absorption from 480 to 750 nm with high extinction coefficient (3.3–3.5 × 10<sup>4</sup> M−<sup>1</sup> cm−<sup>1</sup> ) in the visible range (450–700 nm). **DTIDTQX** absorbed longer wavelength than **DTRDTQX** (631 vs. 588 nm), mainly due to the stronger electron withdrawing ability of 1,3-indanedione group than that of N-ethylrhodanine group.

### Electrochemical Properties

The electrochemical properties of **DTRDTQX** and **DTIDTQX** were studied with cyclic voltammetry (CV) as shown in

**Figure 5**. In addition, the energy levels as well as the band gaps of **DTRDTQX** and **DTIDTQX** were summarized in **Table 1**. With the oxidation and reduction potentials recorded, the HOMO and LUMO levels of the two materials could be calculated (HOMO = −5.1 eV – Eox onset, LUMO = −5.1 eV – Ered onset), which were −5.33 eV, −3.96 eV for **DTIDTQX** and −5.29 eV, −3.59 eV for **DTRDTQX** respectively. Interestingly, the HOMO and LUMO levels of **DTIDTQX** were both deeper than those of **DTRDTQX**. The phenomenon implied that the electron withdrawing ability of 1,3-indanedione group was stronger than that of N-ethylrhodanine group, which was consistent with the observation of UV-Vis absorption. The energy levels of the materials used in the OSCs were depicted in **Figure 6**. The large gap between the low-lying HOMO level (−5.33 eV) of **DTIDTQX** and LUMO (−4.20 eV) of C<sup>70</sup> was evaluated to be 1.13 eV, which resulted in the large Voc (0.71 V) of the optimal **DTIDTQX**-based OSCs in this work. Furthermore, the electrochemical energy band gap (1E CV) of **DTIDTQX** was 0.33 eV lower than that of **DTRDTQX** and strong absorption of **DTIDTQX** active layer in red region could be realized, which was matched well with the UV-Vis absorption spectrum shown in **Figure 4**. Therefore, the lightharvesting capability as well as the photovoltaic performance of **DTIDTQX-**based devices could be superior to that of **DTRDTQX**-based counterparts, which will be discussed further in following.

#### Photovoltaic Properties

To study the photovoltaic properties of the small molecules, OSCs with the structure of FTO/c-TiO2/donor:C70/MoO3/Ag were fabricated. The weight ratios of **DTRDTQX**:C<sup>70</sup> and **DTIDTQX**:C<sup>70</sup> varied from 1:1 to 1:3 and the corresponding J-V curves of the OSCs were shown in **Figures 7**, **8**. All the photovoltaic data of OSCs were summarized in **Table 2**. When the weight ratio of **DTRDTQX**:C<sup>70</sup> reached 1:2 and the photoactive layer was thermal annealed at 100◦C, the


#### TABLE 1 | Physical properties of DTIDTQX and DTRDTQX.

<sup>a</sup>Measured in CH2Cl<sup>2</sup> solution (10−<sup>5</sup> M) and the value was estimated from the onset. <sup>b</sup>Estimated from the HOMO (−5.1 eV) (Cardona et al., 2011) of Fc+/Fc as reference. <sup>c</sup>Temperature corresponding to 5% weight loss obtained from TGA analysis.

best **DTRDTQX**-based OSC was realized with the shortcircuit current density (Jsc) and PCE of 5.66 mA/cm<sup>2</sup> and 1.44%, respectively. The champion **DTRDTQX**-based OSC exhibited almost the same open-circuit voltage (Voc) of ∼0.65 V as other OSCs with different weight ratios (1:1 and 1:3) of **DTRDTQX**:C70. Moreover, for the devices based on **DTRDTQX**:C<sup>70</sup> with the weight ratios of 1:1 and 1:3, the decreased Jsc was mainly ascribed to the imbalanced electron and hole diffusion in the OSCs (Kim et al., 2009). The photovoltaic data in **Table 2** implied that the weight ratio (**DTRDTQX**:C70) of 1:2 was advantageous to the photovoltaic performance of

**DTRDTQX**:C70-based OSCs. The photovoltaic properties of **DTRDTQX**:C70(1:2)-based OSCs with 150◦C thermal annealing and without thermal annealing were also studied and compared. The Voc and PCE of the OSC with 150◦C thermal annealing were decreased to 0.51 V and 1.19%, respectively. As to the OSC without thermal annealing, the PCE was decreased to 1.14% and Voc (∼0.66V) was almost unchanged compared with the champion **DTRDTQX**-based OSC. Therefore, 100◦C thermal annealing treatment was necessary for the reasonable photovoltaic performance of **DTRDTQX**:C70(1:2)-based OSCs according to the experimental data.

As to **DTIDTQX**-based OSCs, the photovoltaic performance was modulated by the weight ratios of **DTIDTQX**:C<sup>70</sup> from 1:1 to 1:3. When the blend ratio of **DTIDTQX**:C<sup>70</sup> reached 1:2, the best **DTIDTQX**-based OSC was realized as shown in **Table 2**. The Voc, Jsc, FF, and PCE of the champion device were 0.71V, 6.24 mA/cm<sup>2</sup> , 0.38 and 1.70%, respectively. It was worthy to note that the Voc of **DTIDTQX**:C70(1:2)-OSC was 0.06 V higher than that of **DTRDTQX**:C70(1:2)-OSC, mainly due to the low-lying HOMO (−5.33 eV) of **DTIDTQX** as shown in **Figure 6**. Moreover, the Jsc and PCE of **DTIDTQX**:C70(1:2)- OSC were both higher than those of **DTRDTQX**:C70(1:2)- OSC. Therefore, the photovoltaic properties of **DTIDTQX**based devices were superior to those of **DTRDTQX**-based counterparts, which was mainly ascribed to the narrow band gap (∼1.37 eV) of **DTIDTQX** and the consequent effective absorption in solar spectrum. The photovoltaic performance of **DTIDTQX**:C70(1:2)-OSC was deteriorated when the active layer

TABLE 2 | Photovoltaic data of the OSCs. DTRDTQX: C70 Thermal annealing Voc (V) Jsc (mA/cm<sup>2</sup> ) FF PCE (%) 1:1 100◦C 0.66 4.27 0.36 1.01 1:2 100◦C 0.65 5.66 0.39 1.44 1:3 100◦C 0.64 5.00 0.33 1.05 1:2 150◦C 0.51 6.09 0.38 1.19 1:2 w/o 0.66 4.86 0.36 1.14 DTIDTQX: C70 1:1 100◦C 0.67 5.13 0.30 1.02 1:2 100◦C 0.71 6.24 0.38 1.70 1:3 100◦C 0.67 6.31 0.34 1.43 1:2 150◦C 0.40 4.71 0.35 0.66 1:2 w/o 0.67 5.51 0.34 1.26

was treated with 150◦C thermal annealing as shown in **Table 2**. And when **DTIDTQX**:C70(1:2)-OSC was fabricated without thermal annealing, the PCE decreased to 1.26%. Therefore, 100◦C thermal annealing was favorable to **DTIDTQX**:C70(1:2)-OSC and a decent PCE of 1.70% was obtained accordingly. However, the FF values of the OSCs were relatively low in this work and much work should be required to further increase FF as well as PCE of the OSCs, such as inserting buffer layers (Ji et al., 2016; Li et al., 2016; Mbuyise et al., 2016), introducing optical spacers (Ben Dkhil et al., 2014), employing solvent annealing (Sun et al., 2014; Li et al., 2015), chemical treatments (Bai et al., 2015), etc.

The morphology of **DTRDTQX**:C70(1:2) and **DTIDTQX**:C70(1:2) films was studied by atomic force microscopy (AFM) (Agilent Series 5500) as shown in **Figure 9**. The root-mean-square roughness (RMS) of **DTIDTQX**:C70

**163**

(1:2) film was 2.94 nm, which was a little higher than that of **DTRDTQX**:C70 (1:2) film (2.58 nm), The relatively low RMS of **DTRDTQX**:C70(1:2) and **DTIDTQX**:C70(1:2) facilitated the reasonable photovoltaic performance of the corresponding devices. Besides, the external quantum efficiency (EQE) spectra of the champion devices were measured with a lock-in amplifier (model SR830 DSP) as shown in **Figure 10**. The EQE of **DTIDTQX**-based device was higher than that of **DTRDTQX**based counterpart and the integrated photocurrent was 5.47 and 4.71 mA/cm<sup>2</sup> , respectively, which was consistent with the photovoltaic properties of the corresponding OSCs. In order to further study the charge transporting properties of the p-type small molecules, hole mobility was measured by using the spacecharge-limited current (SCLC) method and the structure of the hole-only devices was ITO/PEDOT:PSS/donor/Au. The J1/<sup>2</sup> -V curves were measured as shown in Supplementary Material. The relation of J and V could be described by J = 9ε0εµ(Vapp-Vs-Vbi) 2 /8L<sup>3</sup> , where J was the current density, ε<sup>0</sup> was the permittivity of free space, ε was the relative permittivity of the p-type small molecules, µ was the hole mobility, Vapp was the applied voltage, V<sup>s</sup> was the voltage drop from series resistance of the substrate, Vbi was the built-in voltage and L was the thickness of the active layers (Qu et al., 2017). The hole mobilities were calculated with the fitted slope of the J1/<sup>2</sup> -V curves, which were 3.62<sup>∗</sup> 10−<sup>6</sup> cm<sup>2</sup> V −1 s −1 and 2.27<sup>∗</sup> 10−<sup>5</sup> cm<sup>2</sup> V −1 s −1 for **DTRDTQX** and **DTIDTQX**, respectively. The hole mobility of **DTIDTQX** was higher than that of **DTRDTQX**, which contributed to the decent photovoltaic performance of **DTIDTQX**-based OSCs. All the experimental data showed that **DTIDTQX** and **DTRDTQX** were promising donor candidates for small molecule OSCs and improved photovoltaic performance of OSCs based on **DTIDTQX** and **DTRDTQX** would be foreseen in the future.

### REFERENCES


## CONCLUSIONS

Two small molecules **DTRDTQX** and **DTIDTQX** with the D-A-A structure were studied in this work. **DTRDTQX** and **DTIDTQX** were used as the donors in bulk-heterojunction solar cells. The optimal OSCs based on **DTRDTQX**:C70(1:2) and **DTIDTQX**:C70(1:2) were achieved with the PCE of 1.44% and 1.70%, respectively. The photovoltaic properties of **DTIDTQX** were superior to those of **DTRDTQX**, which was attributed to the narrow band gap (1.37 eV) and the high hole mobility (2.27<sup>∗</sup> 10−<sup>5</sup> cm<sup>2</sup> V −1 s −1 ) of **DTIDTQX.** Therefore, **DTRDTQX** and **DTIDTQX** would be promising donor materials for organic solar cells in future.

### AUTHOR CONTRIBUTIONS

All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.

### FUNDING

This work was supported by the National Natural Science Foundation of China under grant Nos 11574013, U1605244, and 11527901, the National Fund for Fostering Talents of Basic Science (NFFTBS) with grant No. J1030310 and J1103205, the authors also extend their appreciation for the support from the International Scientific Partnership Program ISPP at King Saud University, ISPP#0112.

## SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem. 2018.00260/full#supplementary-material


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The reviewer XZ and handling Editor declared their shared affiliation.

Copyright © 2018 Qi, Lo, Zhao, Xuan, Ting, Wong, Rahaman, Chen, Xiao and Qu. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

## Insight Into the Role of PC71BM on Enhancing the Photovoltaic Performance of Ternary Organic Solar Cells

Bei Wang1,2, Yingying Fu<sup>1</sup> , Chi Yan1,3, Rui Zhang1,3, Qingqing Yang<sup>1</sup> , Yanchun Han<sup>1</sup> and Zhiyuan Xie<sup>1</sup> \*

*<sup>1</sup> State Key Laboratory of Polymer Physics and Chemistry, Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, Changchun, China, <sup>2</sup> University of Science and Technology of China, Hefei, China, <sup>3</sup> University of Chinese Academy of Sciences, Beijing, China*

The development of non-fullerene acceptor molecules have remarkably boosted power conversion efficiency (PCE) of polymer solar cells (PSCs) due to the improved spectral coverage and reduced energy loss. An introduction of fullerene molecules into the non-fullerene acceptor-based blend may further improve the photovoltaic performance of the resultant ternary PSCs. However, the underlying mechanism is still debatable. Herein, the ternary PSCs based on PBDB-T:ITIC:PC71BM blend were fabricated and its PCE was increased to 10.2% compared to 9.2% for the binary PBDB-T:ITIC devices and 8.1% for the PBDB-T:PC71BM PSCs. Systematic investigation was carried out to disclose the effect of PC71BM on the blend morphology and charge transport behavior. It is found that the PC71BM tends to intermix with the PBDB-T donor compared to the ITIC counterpart. A small amount of PC71BM in the ternary blend is helpful for ITIC to aggregate and form efficient electron-transport pathways. Accordingly, the electron mobility is increased and the density of electron traps is decreased in the ternary blend in comparison with the PBDB-T:ITIC blend. Finally, the suppressed bimolecular recombination and enhanced charge collection lead to high PCE for the ternary solar cells.

Keywords: ternary organic solar cells, morphology, aggregation, charge transport, trap density

## INTRODUCTION

Non-fullerene acceptors have drawn great research interests in the community of polymer solar cells (PSCs) in recent years. The power conversion efficiencies (PCE) of PSCs employing nonfullerene acceptors have increased rapidly as compared to the PSCs using fullerene derivatives as acceptors (Lin et al., 2015; Cao et al., 2017; Fan et al., 2017; Li et al., 2017; Xiao et al., 2017; Dai et al., 2018). Non-fullerene acceptors possess some advantages such as strong absorption in the visible region and tunable energy levels with regard to fullerene derivatives, and thus allows for suitable combination of donor/acceptor blend to improve the spectral coverage and reduce the energy loss (Holliday et al., 2016; Li et al., 2016, 2018; Qiu et al., 2017). In addition to the strong and complementary absorption and the matched energy levels of the donor/acceptor combination, the donor/acceptor morphology with a suitable phase separation is equally important for realizing a high PCE. More recently, addition of another type of donor or acceptor in the binary non-fullerene PSCs to fabricate so-called ternary PSCs have drawn great interests. This kind of ternary strategy

Edited by:

*Donghong Yu, Aalborg University, Denmark*

#### Reviewed by:

*Junwu Chen, South China University of Technology, China Zhishan Bo, Beijing Normal University, China*

> \*Correspondence: *Zhiyuan Xie xiezy\_n@ciac.ac.cn*

#### Specialty section:

*This article was submitted to Organic Chemistry, a section of the journal Frontiers in Chemistry*

Received: *09 April 2018* Accepted: *15 May 2018* Published: *05 June 2018*

#### Citation:

*Wang B, Fu Y, Yan C, Zhang R, Yang Q, Han Y and Xie Z (2018) Insight Into the Role of PC71BM on Enhancing the Photovoltaic Performance of Ternary Organic Solar Cells. Front. Chem. 6:198. doi: 10.3389/fchem.2018.00198* Wang et al. Ternary Organic Solar Cells

is to some extent powerful to enhance the photovoltaic performance of the devices (Cheng et al., 2014; Gasparini et al., 2016). As it is argued, ternary PSCs possess some features such as more complementary absorption (Jiang et al., 2018) and more appropriate microstructure relative to the binary counterparts (Wang et al., 2017) and easier fabrication compared with tandem solar cells. With these superiorities, ternary PSCs have developed very quickly and become a research focus in the field (Yu et al., 2017; Zhang G. et al., 2017; Zhao et al., 2017; Wang et al., 2018). Nonetheless, the morphology of ternary blends is more complex resulting in the underlying mechanism debatable. It is proposed that the ternary morphology can be divided into four types according to the relative position of third component to the donor phases and acceptor phases, namely the third component embedded in one phase, located at the interfaces, formed alloy structure with either the donor or acceptor material and parallel-like bulk heterojunction structure with the donor or acceptor (Lu et al., 2015). In fact, the ternary morphology is too complicated to be clearly identified using the current technology. Typical electron acceptor fullerene derivatives, such as PC71BM and Bis-PC71BM have been used as additives in non-fullerene PSCs. For example, Bo et al. have reported high-performance ternary PSCs employing non-fullerene and fullerene acceptors simultaneously for the first time, in which they found that a small amount of fullerene is in favor of boosting the photovoltaic performance of the devices (Lu et al., 2016). Hou et al. also reported the high-efficiency ternary PSCs using Bis-PC71BM as the third component (Zhao et al., 2017). They proposed that Bis-PC71BM mainly exists in the upper surface of active layer and promotes electron transport in their ternary blend PSCs. For the ternary PSCs, the third component such as PC71BM may have strong effect on the resultant ternary morphology and hence its photovoltaic properties. Although some researches on the function of third component were carried out (Yu et al., 2017; Zhang J. et al., 2017; Wang et al., 2018), the underlying mechanism is still debatable.

Herein, a reported wide bandgap polymer poly[(2,6- (4,8-bis(5-(2-ethylhexyl)thiophen-2-yl)-benzo[1,2-b:4,5-b′ ] dithiophene))-alt-(5,5-(1′ ,3′ -di-2-thienyl-5′ ,7′ -bis(2-ethylhexyl) benzo[1′ ,2′ -c:4′ ,5′ -c′ ]dithiophene-4,8-dione))] (PBDB-T) is used as the donor, organic molecule (3,9-bis(2-methylene-(3- (1,1-dicyanomethylene)-indanone))- 5,5,11,11-tetrakis(4 hexylphenyl)-dithieno[2,3-d:2′ ,3′ -d′ ]-s-indaceno[1,2-b:5,6-b′ ] dithiophene) (ITIC) is used as the acceptor and [6,6]-phenyl-C71-butyric acid methyl ester (PC71BM) is used as the third component to prepare the ternary PSCs. The optimized PSCs based on ternary PBDB-T:ITIC:PC71BM blend demonstrate a higher PCE of 10.2% than 9.2% of the binary PBDB-T:ITIC devices and 8.1% of the PBDB-T:PC71BM PSCs. Further studies are mainly focused on the effect of PC71BM on the blend morphology and charge transport behavior. It is found that the PC71BM tends to intermix with the PBDB-T donor compared to the ITIC counterpart. A small amount of PC71BM in the ternary blend is helpful for ITIC to aggregate and form efficient electron-transport pathways. Accordingly, the electron mobility is increased and the density of electron traps is decreased in the ternary blend in comparison with the PBDB-T:ITIC blend. Finally, the suppressed bimolecular recombination and enhanced charge collection lead to an enhanced PCE for the ternary solar cells.

## EXPERIMENTAL SECTION

Both the polymer PBDB-T donor and small molecule ITIC acceptor were bought from Solarmer Ltd. The PBDB-T has a molecular weight Mn of 21.5 kDa and a PDI of 1.9. PC71BM was bought from American Dye Source. The interfacial material PDINO was provided by Dr. Zhiguo Zhang in Institute of Chemistry, Chinese Academy of Sciences. All materials were used as received without further purification.

Polymer solar cells were fabricated with a structure of ITO/PEDOT: PSS/active layer/PDINO/Al. The ITO substrates were subject to routine cleaning procedure of detergent, acetone and deionized water. After drying in an oven for 30 min, the ITO substrates were treated with UV-ozone for 25 min. The PEDOT:PSS layer was first deposited via spin-coating and dried at 140◦C for 30 min in air. The subsequent active layer and buffer layer were spin-coated in a glove box. The active layers consisting of PBDB-T:PC71BM, PBDB-T:ITIC, or ternary PBDB-T:ITIC:PC71BM were spin-coated from their respective solutions in CB containing 0.5% DIO with a total concentration of 20 mg/mL. The spin-coating rate for the active layers was kept at 2,500 rpm for 1 min. The samples were annealed at 160◦C for 10 min. A cathode buffer layer of PDINO was spin-coated on the active layer from its methanol solution with a concentration of 1 mg/mL at 3,000 rpm for 30 s. Finally, the Al cathode with a thickness of 100 nm was thermally deposited in a vacuum chamber. The electron—and hole-only devices were fabricated under the same procedure with a structure of ITO/PEIE/active layer/PDINO/Al and ITO/PEDOT:PSS/active layer/MoO3/Al, respectively.

The current density-voltage (J-V) curves of the PSCs were traced by a computer-controlled Keithley 2400 Source Meter under simulated solar light illumination (AM 1.5G, 100 mW/cm<sup>2</sup> ). The EQE data were measured by solar cell spectral response measurement system (QE-R 3011, Enli Tech. Co.). The film absorption and fluorescence spectra were recorded on Agilent Cary 60UV-Vis spectrophotometer and Perkin-Elmer LS 55 spectrofluorometer, respectively. The thicknesses of individual layers were measured with a surface profilometer. The 2D-GIXD data were acquired at station 14B in Shanghai Synchrotron Radiation Facility.

### RESULTS AND DISCUSSION

The chemical structures of PBDB-T, ITIC and PC71BM are shown in **Figure 1A**. The normalized UV-vis absorption spectra of neat PBDB-T, ITIC and PC71BM films are plotted in **Figure 1B**. It is clearly seen from **Figure 1B** that the main peaks of PBDB-T and ITIC are located at 625 and 710 nm, respectively, and their complementary absorption can strongly improve the spectral coverage in the visible region. PC71BM exhibits absorption in short wavelength region but its absorption capability is relatively weak. In this study, the PSCs with a conventional device structure of ITO/PEDOT:PSS/active layer/PDINO/Al were fabricated, and their energy levels were plotted in **Figure 1C**. The polymer PDINO was used as cathode buffer layer referenced to the literature (Zhang Z. et al., 2014). The total ratio of donor component to acceptor component is kept at a fixed ratio of 1:1 (w/w). The ratio of PBDB-T:ITIC:PC71BM is marked as the D:A1:A2.

The photovoltaic performance of the PSCs based on the PBDB-T:ITIC:PC71BM (1:1-x:x) blend were firstly evaluated. The ratio of A1:A2 is varied from 1:0 to 0:1 while the ratio of D:(A1+A2) is fixed at 1:1 in order to keep a constant active layer thickness. The detailed photovoltaic parameters of the resultant PSCs with different ratio of two acceptors are listed in Table S1 in Supplementary Information. The dependence of VOC, JSC, FF, and PCE on the PC71BM content for the resultant PSCs are plotted in **Figure 2A**. In the case of VOC, it undergoes gradually decrease from 0.902 V for the binary PBDB-T:ITIC PSCs to 0.856 V for the binary PBDB-T: PC71BM PSCs with increasing PC71BM content from 0 to 1 in the ternary blend PSCs. The change is reasonable since the LUMO of ITIC is higher than that of PC71BM, and the electron transport and collection will occur in PC71BM phases when the PC71BM amount is larger. The interesting thing is that the JSC are initially increased and then decreased with the increase of PC71BM contents for the resultant ternary PSCs. The FF shows similar trend. Considering the intrinsic absorption properties between ITIC and PC71BM, the substitution of small amount of ITIC by PC71BM would decrease the total absorption of the ternary PBDB-T:ITIC:PC71BM blend active layer. Thus, the improved JSC and FF for the ternary PBDB-T:ITIC:PC71BM PSCs may imply that both the exciton dissociation and charge-collection efficiencies are enhanced in comparison to the binary PBDB-T:ITIC PSCs. It is plotted the illuminated J-V curves of the PBDB-T:ITIC PSCs, the PBDB-T:PC71BM PSCs and the optimized ternary PBDB-T:ITIC:PC71BM PSCs in **Figure 2B**, respectively. The non-fullerene PSCs based on binary PBDB-T:ITIC blend demonstrate a VOC of 0.902 V, a JSC of 15.06 mA/cm<sup>2</sup> , a FF of 0.69, respectively, leading to a PCE of 9.38%. Due to the limited spectral coverage, the PBDB-T:PC71BM PSCs show a lower JSC of 13.64 mA/cm<sup>2</sup> , and finally result in a PCE of 8.21% together with a VOC of 0.856 V and a FF of 0.703. The ternary PBDB-T:ITIC:PC71BM (1:0.8:0.2) PSCs demonstrate a VOC of 0.892 V, a JSC of 15.98 mA/cm<sup>2</sup> , a FF of 0.717 and an overall PCE of 10.22%. Although the VOC is a little lowered in comparison to the binary PBDB-T:ITIC PSCs, the increased JSC and FF boost the PCE enhancement of the ternary PSCs. The external quantum efficiency (EQE) curves of the three devices are shown in **Figure 2C**. The spectral response of the PBDB-T:PC71BM PSCs cover at a range of 300–700 nm with EQE higher than 70% at 450–650 nm. The PBDB-T:ITIC PSCs show extended spectral coverage of 300–800 nm due to the narrow bandgap of ITIC. The ternary PBDB-T:ITIC:PC71BM PSCs show similar spectral response profile but high EQE compared to the PBDB-T:ITIC PSCs. The internal quantum efficiency (IQE) of these devices are also measured to clarify the absolute quantum efficiency in these devices and the curves are plotted in **Figure 2D**. It indicates that the photon-to-electron conversion efficiency is really improved by adding some amount of PC71BM into the PBDB-T:ITIC blend. The absorption spectra of the active layers in three kinds of PSCs are plotted in **Figure 3A**.

The PBDB-T:ITIC:PC71BM (1:0.8:0.2) active layer shows a little low absorption at 550–750 nm due to decreased ITIC content compared to the PBDB-T:ITIC (1:1) film. Although the PBDB-T:PC71BM (1:1) film demonstrates enhanced absorption at 300– 500 nm, its absorption at 550–700 nm is remarkably decreased. Photoluminescence (PL) quenching experiments were carried out to check the charge transfer status in these films. As shown in **Figure 3B**, the introduction of PC71BM favors to quench the excitons dominated on the PBDB-T donor.

The relationship between photocurrent (Jph) and effective voltage (Veff) is investigated to judge the charge generation and collection status in the PSCs based on the PBDB-T:ITIC (1:1), PBDB-T:ITIC:PC71BM (1:0.8:0.2), and PBDB-T:PC71BM (1:1) blends (Yuan et al., 2017; Zhang G. et al., 2017). As shown in **Figure 4A**, Jph is given by Jph = J<sup>L</sup> − JD, where J<sup>L</sup> & J<sup>D</sup> are the current density under illumination and in the dark, respectively. Veff is defined as V<sup>0</sup> − Vappl, V<sup>0</sup> is the voltage when Jph = 0 and Vappl is the applied voltage during the measurement. The Jph is supposed to be saturated at a Veff of 3 V, which are 16.55 mA/cm<sup>2</sup> , 17.40 mA/cm<sup>2</sup> and 14.45 mA/cm<sup>2</sup> , respectively, for the PSCs based on the PBDB-T:ITIC

(1:1), PBDB-T:ITIC:PC71BM (1:0.8:0.2), and PBDB-T:PC71BM (1:1) blends. The device based on ternary PBDB-T:ITIC:PC71BM (1:0.8:0.2) blend demonstrates the highest saturation current, indicating its highest charge-generation capability. Charge dissociation probability P(E,T) is defined as Jph/Jsat. When a high bias is applied, the charge recombination is suppressed and most of photo-generated charges are extracted, leading to P(E,T) close to100%. It is calculated that the P(E,T) values under short-circuit condition (Vappl = 0 V) are 90.6, 91.9, and 94.9%, respectively, for the PSCs based on the PBDB-T:ITIC (1:1), PBDB-T:ITIC:PC71BM (1:0.8:0.2), and PBDB-T:PC71BM (1:1) blends. The PBDB-T:PC71BM (1:1) device shows the best chargeextraction capability, whereas the charge-extraction ability is a little poor for the PBDB-T:ITIC (1:1) device. The ternary PBDB-T:ITIC:PC71BM (1:0.8:0.2) shows improved exciton dissociation and charge extraction compared to the PBDB-T:ITIC (1:1) device. The charge recombination status in these devices is also investigated via the dependence of JSC on the incident light intensity as shown in **Figure 4B**. JSC is dependent on the incident light intensity with JSC∝I α , in which exponential factor α would be close to unity without bimolecular recombination (Gao et al., 2016; Xu et al., 2017, 2018). The PBDB-T:PC71BM (1:1) device has the highest α value of 0.964, whereas α value of the PBDB-T:ITIC (1:1) device is merely 0.935 implying severe bimolecular recombination within the film. Such kind of bimolecular recombination in the PBDB-T:ITIC (1:1) device is to some extent suppressed with α value of 0.955 in ternary PBDB-T:ITIC:PC71BM (1:0.8:0.2) device.

As discussed above, with the introduction of PC71BM into the PBDB-T:ITIC blend, the resultant ternary PSCs demonstrate an enhanced exciton-dissociating and charge-extracting property, leading to improved PCE compared to the PBDB-T:ITIC PSCs. The detailed mechanism accounting for the enhancement is further investigated. The electron and hole transport properties of the various blend films were firstly measured via space-charge limited current (SCLC) method (Bin et al., 2016). The J-V curves of the devices were plotted as in Figure S1 and the calculated electron and hole mobility are listed in Table S2. The PBDB-T:PC71BM (1:1) blend film shows relatively high electron and hole mobility of 8.45 × 10−<sup>4</sup> and 5.30 × 10−<sup>4</sup> cm2V −1 s −1 , respectively. However, the electron and hole mobility of the PBDB-T:ITIC (1:1) film is 3.05 × 10−<sup>4</sup> and 2.70 × 10−<sup>4</sup> cm2V −1 s −1 . After incorporating PC71BM, the electron and hole mobility of the resultant PBDB-T:ITIC:PC71BM (1:0.8:0.2) film is increased to 4.55 × 10−<sup>4</sup> and 3.37 × 10−<sup>4</sup> cm2V −1 s −1 . As it is known, the high and balanced charge transport may help to efficient charge extraction and hence low charge recombination. This is in agreement with the recombination status in these PSCs. It is speculated that the incorporation of PC71BM improves the PBDB-T and ITIC interpenetrating networks and thus the photovoltaic performance.

The morphology of the PBDB-T:ITIC (1:1), PBDB-T:ITIC:PC71BM (1:0.8:0.2), and PBDB-T:PC71BM (1:1) blend films are investigated and their AFM and TEM images are shown in Figures S2, S3. All the blend films are smooth with a root mean square (RMS) roughness of <2 nm. However, the morphology of the ternary blend is not changed obviously compared to the PBDB-T:ITIC (1:1) blend film. The structural information of the blended films are further investigated by grazing incidence X-ray diffraction (GIXD). The 2D-GIXD images of the pure PBDB-T and ITIC films are plotted in Figure S4. The pure PBDB-T film shows dominant peak at qxy = 0.29 Å−<sup>1</sup> in the in-plane direction, attributing to its strong (100) diffraction. The pure ITIC film demonstrates two featured peaks at q<sup>z</sup> = 0.26 Å−<sup>1</sup> and q<sup>z</sup> = 0.53 Å −1 in out-of-plane direction and one peak in in-plane direction at qxy = 0.36 Å−<sup>1</sup> . 2D-GIXD patterns and line-cut profiles of the PBDB-T:ITIC (1:1), PBDB-T:ITIC:PC71BM (1:0.8:0.2) and PBDB-T:PC71BM (1:1) films are shown in **Figure 5**. The diffractions originated from PBDB-T aggregation is observed both in PBDB-T:ITIC (1:1) and PBDB-T:PC71BM (1:1) films in contrast to the pure PBDB-T film. However, the diffraction signal originated from the ITIC component is not obviously presented in PBDB-T:ITIC blend film, implying its poor aggregation. In the PBDB-T:ITIC:PC71BM (1:0.8:0.2) blend film, the diffraction signal from PBDB-T is still presented but weakened. More importantly, two additional peaks marked 1 and 2 are presented in the ternary blend film compared to the PBDB-T:ITIC (1:1)

counterpart, which corresponds to the featured diffraction from ITIC in the in-plane and out-of-plane directions. Bo et al. observed similar phenomenon in their work (Lu et al., 2016). This indicates that the ITIC aggregation is enhanced in the PBDB-T:ITIC:PC71BM (1:0.8:0.2) blend film though the ITIC content is decreased. In other words, the small amount of PC71BM may serve as "lubricant" to favor ITIC molecules to aggregate out of the polymer PBDB-T matrix.

The prerequisite to get such a conclusion is that the compatibility between PBDB-T and PC71BM should be better than that between PBDB-T and ITIC. The electron mobility of the PBDB-T:ITIC (1:x) and PBDB-T:PC71BM (1:x) blend films are measured and its dependence on the acceptor ratio is plotted in **Figure 6**. It is supposed that the acceptor phase forms continuous tunnels when the electron mobility of the blend films exceeds 10−<sup>5</sup> cm2V −1 s −1 . Both the as-prepared film deposited in CB solution and the annealed film deposited in CB:DIO (0.5%, v/v) solution are tested. The results show that the continuous electron-transport tunnels is formed in the PBDB-T:ITIC film with the ITIC/PBDB-T ratio <20%. The DIO additive and thermal annealing donot change the electron transport remarkably. For the as-prepared PBDB-T:PC71BM film, the critical PC71BM content is increased to 25%. The DIO additive and thermal annealing lowers the critical PC71BM content. This indicates that the PBDB-T is inclined to intermix

with PC71BM better than with ITIC. In the ternary PBDB-T:ITIC:PC71BM (1:0.8:0.2) blend film, the small amount of PC71BM may mix with PBDB-T and does not form networks.

Moreover, the existence of PC71BM favors ITIC aggregating to form electron-transporting networks. This is confirmed by the increased electron mobility and the XRD results compared to the PBDB-T:ITIC (1:1) film.

The electron trap-state density in these blend films is further investigated. From the I–V curves of the electron-only devices based on the PBDB-T:ITIC (1:1), PBDB-T:ITIC:PC71BM (1:0.8:0.2) and PBDB-T:PC71BM (1:1) blend shown in **Figure 7**, the electron trap density N is calculated by the equation as below (Yang et al., 2016):

$$V\_{TFL} = \frac{eNL^2}{2\varepsilon\_0 \varepsilon\_r} \tag{1}$$

where VTFL is the trap-filled limit voltage, e is the elementary charge of electron, L is the thickness of film, ǫ<sup>0</sup> is the vacuum permittivity, and ǫ<sup>r</sup> is the relative dielectric constant (ǫ<sup>r</sup> = 3). The VTFL of three devices is 0.12, 0.05, and 0.02 V. The calculated deep trap density is 3.3 × 1015, 1.5 × 1015, and 7.3 × 10<sup>14</sup> cm−<sup>3</sup> , respectively. It confirms that the incorporation of PC71BM is helpful for ITIC molecules to move out of PBDB-T phases and thus the density of electron traps is reduced. The increased electron mobility and the decreased electron traps are

#### REFERENCES


helpful to enhance the photovoltaic performance of the PBDB-T:ITIC:PC71BM (1:0.8:0.2).

#### CONCLUSION

In summary, the ternary PSCs based on PBDB-T:ITIC:PC71BM (1:0.8:0.2) blend were fabricated and its PCE was increased to 10.2% compared to 9.2% for the PBDB-T:ITIC (1:1) devices. The mechanism accounting for the enhanced photovoltaic performance is discussed in detail. It is found that the PC71BM tends to intermix with the PBDB-T donor compared to the ITIC counterpart. A small amount of PC71BM in the ternary blend is helpful for ITIC to aggregate and form efficient electrontransport pathways. The electron mobility is increased and the density of electron traps is decreased in the ternary PBDB-T:ITIC:PC71BM (1:0.8:0.2) blend in comparison with the PBDB-T:ITIC blend. Finally, the suppressed bimolecular recombination and enhanced charge collection lead to a high PCE for the ternary solar cells.

#### AUTHOR CONTRIBUTIONS

BW and CY conceived and designed the experiments. BW performed the experiments. RZ performed 2D-GIXD measurements. QY performed the absorption and PL experiments. BW and YF analyzed data. BW wrote the manuscript. All authors discussed and commented on the paper.

#### ACKNOWLEDGMENTS

This work is supported by the National Key Basic Research and Development Program of China (Nos. 2014CB643504, 2015CB655001) and the National Natural Science Foundation of China (nos.21774122, 21334006, 51611530705, 51773195, 51703222). The financial support of the Strategic Priority Research Program of the Chinese Academy of Sciences (XDB12030200) is also acknowledged. We also thank Shanghai Synchrotron Radiation Facility (SSRF) beamline 14B for the grazing X-ray measurements.

#### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem. 2018.00198/full#supplementary-material


Cheng, P., Li, Y., and Zhan, X. (2014). Efficient ternary blend polymer solar cells with indene-C60 bisadduct as an electron-cascadeacceptor. Energy Environ. Sci. 7, 2005–2011. doi: 10.1039/c3ee44202k


all small molecule ternary solar cells. J. Am. Chem. Soc. 140, 1549–1556. doi: 10.1021/jacs.7b13054


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Wang, Fu, Yan, Zhang, Yang, Han and Xie. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

## Utilizing Benzotriazole and Indacenodithiophene Units to Construct Both Polymeric Donor and Small Molecular Acceptors to Realize Organic Solar Cells With High Open-Circuit Voltages Beyond 1.2 V

#### Edited by:

*Chuanlang Zhan, Institute of Chemistry (CAS), China*

#### Reviewed by:

*Qiang Peng, Sichuan University, China Zhan'Ao Tan, North China Electric Power University, China*

#### \*Correspondence:

*Erjun Zhou zhouej@nanoctr.cn*

#### Specialty section:

*This article was submitted to Organic Chemistry, a section of the journal Frontiers in Chemistry*

Received: *11 March 2018* Accepted: *16 April 2018* Published: *01 May 2018*

#### Citation:

*Tang A, Chen F, Xiao B, Yang J, Li J, Wang X and Zhou E (2018) Utilizing Benzotriazole and Indacenodithiophene Units to Construct Both Polymeric Donor and Small Molecular Acceptors to Realize Organic Solar Cells With High Open-Circuit Voltages Beyond 1.2 V. Front. Chem. 6:147. doi: 10.3389/fchem.2018.00147* Ailing Tang<sup>1</sup> , Fan Chen1,2, Bo Xiao1,2, Jing Yang1,2, Jianfeng Li 1,2, Xiaochen Wang<sup>1</sup> and Erjun Zhou<sup>1</sup> \*

*<sup>1</sup> CAS Key Laboratory of Nanosystem and Hierarchical Fabrication, CAS Center for Excellence in Nanoscience, National Center for Nanoscience and Technology, Beijing, China, <sup>2</sup> University of Chinese Academy of Sciences, Beijing, China*

Devolopment of organic solar cells with high open-circuit voltage (*V*OC) and power conversion efficiency (PCE) simutaniously plays a significant role, but there is no guideline how to choose the suitable photovoltaic material combinations. In our previous work, we developed "the Same-Acceptor-Strategy" (SAS), by utilizing the same electron-accepting segment to construct both polymeric donor and small molecular acceptor. In this study, we further expend SAS to use both the same electron-accepting and electron-donating units to design the material combination. The p-type polymer of PIDT-DTffBTA is designed by inserting conjugated bridge between indacenodithiophene (IDT) and fluorinated benzotriazole (BTA), while the n-type small molecules of BTA*x* (*x* = 1, 2, 3) are obtained by introducing different end-capped groups to BTA-IDT-BTA backbone. PIDT-DTffBTA: BTA*x* (*x* = 1–3) based photovolatic devices can realize high *V*OC of 1.21–1.37 V with the very small voltage loss (0.55–0.60 V), while only the PIDT-DTffBTA: BTA3 based device possesses the enough driving force for efficient hole and electron transfer and yields the optimal PCE of 5.67%, which is among the highest value for organic solar cells (OSCs) with a *V*OC beyond 1.20 V reported so far. Our results provide a simple and effective method to obtain fullerene-free OSCs with a high *V*OC and PCE.

Keywords: benzotriazole, indacenodithiophene, fullerene-free organic solar cells, high open-circuit voltage, non-fullerene acceptor

## INTRODUCTION

As one of the most promising technique in photoelectric conversion, bulk-heterojunction (BHJ) organic solar cells (OSCs) have been extensively studied. For a long time, fullerene derivatives have taken up the major part of the acceptor materials, benefited from their high electron affinity and electron mobility, as well as isotropic charge transport. (Guldi, 2000; von Hauff et al., 2005; Anthony et al., 2010; Eftaiha et al., 2014) It's not until the recent 2 years that non-fullerene small molecular acceptors (NFSMAs) with strong sunlight harvesting capability and tunable energy levels have drawn considerable attention and hundreds of novel NFSMAs have been developed (Hwang et al., 2015; Lin et al., 2015, 2016a,b; Zhong et al., 2015; Guo et al., 2016; Holliday et al., 2016; Li et al., 2016b, 2017; Wu et al., 2016; Duan et al., 2017a,b; Fan et al., 2017; Liu et al., 2017; Sun et al., 2017; Wang et al., 2017; Xiao B. et al., 2017a,b; Xu S. J. et al., 2017; Xu X. et al., 2017; Yang et al., 2017; Zhang G. et al., 2017; Zhang Z. G et al., 2017) To date, the power conversion efficiencies (PCEs) of the fullerene-free OSCs have reached up to 13%. (Xiao Z. et al., 2017a; Zhao et al., 2017). The short circuit current (JSC) and fill factor (FF) in these efficient fullerene-free OSCs could arrive as high as 18–25 mA cm−<sup>2</sup> and 60–70%, respectively. However, the open-circuit voltage (VOC) values remain relatively low (<1.0 V), because of the large energy loss (Eloss) (Xiao Z. et al., 2017b).

Recently, by developing novel NFSMAs with high LUMO levels and choosing suitable p-type polymers, the resulted solar cells could realize high VOC values of beyond 1.0 V (Yang et al., 2014; Yu et al., 2014; Zhang et al., 2015, 2016; Baran et al., 2016; Fu et al., 2016; Li et al., 2016a; Liu et al., 2016; Ni et al., 2016; Chen et al., 2017; Ding et al., 2017; Xiao B. et al., 2017a; Zhan et al., 2017; Zhang Y. et al., 2017). In fact, there is always a problematic trade-off between JSC and VOC. Thus, very limited devices could simultaneously realize a high VOC of beyond 1.2 V and a high PCE (As shown in **Figure 1**) (Fu et al., 2016; Xiao B. et al., 2017a; Zhan et al., 2017; Zhang Y. et al., 2017). For example, Xie et al. reported that the OSCs containing poly(3-hexylthiophene) (P3HT) as electron donor and the oligomer F4TBT4 with four repeated fluorene and di-2-thienyl benzothiadiazole units as electron acceptor can output a high VOC above 1.2 V and a PCE of 4.12% (Fu et al., 2016). Our group synthesized a novel benzotriazole based acceptor, BTA2, which showed a high VOC of 1.22 V with an acceptable PCE of 4.5% with P3HT as donor (Xiao B. et al., 2017a). Recently, a pyrene-fused perylene diimide acceptor synthesized by Li and Sun et al. can achieve a high VOC of 1.21 V with a PCE of 5.10%, with the wide-bandgap polymer PBT1-EH as the donor (Zhan et al., 2017). Zhang et al. reported a combination of perylene monoimide (PMI)-based electron acceptor and a wide-bandgap polymer of PTZ1, which demonstrated a very high VOC of 1.3 V with a PCE of 6% (Zhang Y. et al., 2017). However, in the above-cases, they used a trial-anderror procedure and there is no guideline how to choose suitable p-type polymer and n-type NFSMA combination to realize such a high VOC and PCE simultaneously. Thus, finding a promising strategy to achieve a high VOC without sacrificing the other impact factors is essential for practical application. Furthermore, this kind of OSCs with a high VOC can also be used as a sub cell in the tandem devices to offer opportunities to realize a high VOC beyond 2.0 V (Xu et al., 2018), which will supply high enough voltage for solar-energy-driven water splitting (Walter et al., 2010; Luo et al., 2014).

Based on the molecular orbital theory, for donor-acceptor (D-A)-type conjugated materials, the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) energy level are mainly decided by the electrondonating and electron-withdrawing segments respectively. Thus, it may realize the close molecular energy levels by utilizing the same building blocks to build both donor and acceptor materials. By further slightly modulating the chemical structures, the enough LUMO-LUMO and HOMO-HOMO offsets could be realized to guarantee sufficient driving force for efficient hole and electron transfer and simultaneously result in an ultra-large Voc. In our previous work, we used "the Same-Acceptor-Strategy" (SAS), both p-type polymer of J61 and n-type small molecule of

BTA3 contain the same electron-accepting unit of BTA, which could realize a high Voc of 1.15 V (Tang et al., 2017). In this paper, we further adopted this feasible strategy and chose the classic electron-donating unit of indacenodithiophene (IDT) and electron-accepting segment of benzotriazole (BTA) to construct the polymer donor and small molecular acceptors. The donoracceptor (D-A) type copolymer of PIDT-DTffBTA, as shown in **Figure 2**, containing IDT as the donor unit and difluorosubstituted BTA as the acceptor unit and thiophene as π-bridge, was designed as the polymer donor. In addition, we utilized IDT and BTA to construct the conjugated backbone of the small molecular acceptors and fine-tuned the electron-withdrawing end-capped units to adjust their energy levels. In our previous work, we have proved that the energy levels of BTA-based small molecules could be fine-tuned and OSCs based P3HT: BTA1 (Xiao B. et al., 2017b), P3HT: BTA2 (Xiao B. et al., 2017a), and J61: BTA3 (Tang et al., 2017) could realize relatively higher VOC of 1.02, 1.22, and 1.15 V, respectively.

Here, as expected, these devices based on PIDT-DTffBTA: BTAx (x = 1–3) as acceptors exhibited the reduced energy loss below 0.60 eV and the according VOC in the range of 1.21–1.37 V were higher by nearly 0.3–0.5 V than that of [6,6]-phenyl-C71 butyric acid methyl ester (PC71BM) based device. Differently, the BTA2 and BTA1 based devices showed poor JSC below 0.4 mA cm−<sup>2</sup> while BTA3 gave a dramatically increased JSC of 8.68 mA cm−<sup>2</sup> . It is remarkable that the achieved PCE of 5.67% in BTA3 based device is among the highest values for OSCs with a VOC beyond 1.20 V reported so far. Our results provide a simple and feasible method to design both p-type and n-type photovoltaic materials for OSCs with high VOC and PCE.

#### RESULTS AND DISCUSSION

#### Theoretical Calculation

Calculations with density functional theory (DFT) at the B3LYP/6-31G(d) level are firstly performed to compare the energy levels of these photovoltaic molecules. The polymers were replaced with the dimers of the repeating units and the long alkyls were replaced with methyl groups to simplify the calculations. As shown in **Figure 3**, the calculted LUMO/HOMO levels of PIDT-DTffBTA and BTAx (x = 1, 2, 3) are −2.66/−4.61, −2.93/−5.06, −2.74/−4.99, and −3.15/−5.26 eV, respectively. The LUMO offsets between PITD-DTffBTA and three acceptors (BTA2, BTA1 and BTA3) (1ELUMOD−LUMO<sup>A</sup> ) are calculated to be 0.08, 0.27, and 0.49 eV, respectively, and the according HOMO offsets (1EHOMOD−HOMO<sup>A</sup> ) are 0.38, 0.45, 0.65 eV, respectively. These results reveal that utilizing the same building blocks to construct both p-type polymeric donor and n-type small molecular acceptors has the potential to realize similar LUMO levels and give rise to high voltage. Modulating the end-capped units can help optimize the energy offsets to achieve the enough driving force for charge transfer and produce high photocurrent.

#### Synthesis

The synthetic routes of the photovoltaic materials are depicted in Scheme S1. PIDT-DTffBTA was synthesized by

Stille-coupling reaction between 4,7-bis-(5-bromothiophen-2-yl)-5,6-difluoro-2-octyl-2H-benzotriazole and (4,4,9,9 tetrakis(4-hexylphenyl)-4,9-dihydro-s-indaceno [1,2-b:5,6 b']dithiophene-2,7-diyl)bis(trimethylstannane). BTAx were synthesized by Stille-coupling reaction and a Knoevenagel condensation with a yield of 70–80%. The number-average molecular weight (Mn) and polydispersity index (PDI) value of PIDT-DTffBTA are 62.8 kDa and 1.55, respectively, determined by gelpermeation chromatography (GPC) (see Figure S1). The temperature with the 5% weight loss (Td) of BTAx (x = 1–3) are 386, 405, and 396◦C, respectively, measured with thermogravimetric analysis (TGA) (Figure S2**)**. All the materials are soluble in common organic solvents, such as chloroform (CF), chlorobenzene (CB), and o-dichlorobenzene (o-DCB).

## Optical Properties

The UV–vis absorption spectra of the donor and acceptors in solution and films are shown in Figure S3 and **Figure 4A**, respectively, and their absorption characteristics are summarized in Table S1. The absorption of the BTA2 overlaps with that of PIDT-DTffBTA. The maximum absorption peaks red-shifts as the increase of the electron-withdrawing properties of the endcapped units. As a result, the absorption of BTA1 and BTA3 become more and more complementary to that of PITD-DTBTA, which would allow for an improved JSC compared to BTA2 based devices. The optical band gaps (E ◦pt <sup>g</sup> ) of PIDT-DTffBTA and BTAx (x = 2, 1, 3) calculated from the film absorption onsets are ca. 1.96, 2.00, 1.87, 1.76 eV, respectively.

The molecular energy levels of these four materials are determined by electrochemical cyclic voltammetry (CV, **Figure 4B**). The according results are listed in Table S1. Calculated with their onset oxidation potentials, the HOMO levels of PIDT-DTffBTA and BTAx (x = 2, 1, 3) are −5.34, −5.43, −5.46, and −5.49 eV, respectively. The LUMO levels are estimated to be −3.38, −3.43, −3.59, and −3.73 eV, respectively, by adding the optical bandgap to their HOMO levels. 1EHOMOD−HOMO<sup>A</sup> between the donor and the acceptors (BTA2, 1, 3) are 0.09, 0.12, 0.15 eV, respectively, and 1ELUMOD−LUMO<sup>A</sup> are respectively 0.05, 0.21, 0.35 eV. As expected, the very small 1ELUMOD−LUMO<sup>A</sup> produce the ultar-high offsets between the

HOMO level of the donor and the LUMO level of the acceptor, giving a chance to achieve the ultra-hight VOC (Armstrong et al., 2009).

### Photovoltaic Device Performance

To investigate the photovoltaic properties, photovoltaic devices are fabricated with a conventional device configuration of indium tin oxide (ITO)/ poly(3,4-ethylenedioxythiophene): poly(styrenesulfonate)(PEDOT:PSS)/PIDT-DTffBTA: BTAx (x = 1–3)/Ca/Al. The optimized photovoltaic characteristics are listed in **Table 1** and the optimal current density–voltage (J– V) curves and the corresponding external quantum efficiency (EQE) spectra are displayed in **Figure 5**. The detail optimization conditions are shown in Tables S2–S7. The device using the PIDT-DTffBTA: BTA1 and PIDT-DTffBTA: BTA2 blend (1:1 in wt %) with thermal annealing show nearly no performance with a PCE of 0.12 and 0.07%, respectively, which is likely due to their very high-lying LUMO, resulting in insufficient charge transfer from polymer to acceptor. Under the same conditions, the BTA3-based device (1:3 in wt %) exhibits improved solar cell performance with a PCE of 2.61%, which may be attributed to the decreased LUMO level. After solvent annealing, the devices using PIDT-DTffBTA: BTA3 show the highest PCE of 5.67% with the increased FF and JSC. It has been seen that after solvent annealing, the strong aggregation (Figure S4) can enhance domain purity and further improve charge transport in the active layer (see farther below), giving rise to the improved JSC and FF.

As expected, all of the three devices show high VOC (> 1.2 V), which are much higher than that of PIDT-DTffBTA: PC71BM ([6,6]-phenyl C<sup>71</sup> butyric acid methyl ester) based device (VOC = 0.88 V) and PIDT-DTffBTA: ITIC based device (VOC = 0.94 V). The VOC of the devices increase from 1.21, 1.27 to 1.37 V in the order of BTA3, BTA1, and BTA2. The energy loss values for the OSCs of PIDT-DTffBTA: BTAx (x = 2, 1, 3) are calculated to be 0.59, 0.60, and 0.55 eV, respectively, which is a result of the high VOC. Interestingly, the JSC for PIDT-DTffBTA: BTA2 and PIDT-DTffBTA: BTA1 based devices are very low (<0.5 mA cm<sup>2</sup> ), while the one for PIDT-DTffBTA:BTA3 based device dramatically reaches up to 8.68 mA cm−<sup>2</sup> , close to the value in PIDT-DTffBTA: PC71BM based device (9.06 mA cm−<sup>2</sup> ). Accordingly, the maximum EQE values of PIDT-DTffBTA: BTA2 and PIDT-DTffBTA: BTA1 based devices are below 5%, while the one of PIDT-DTffBTA: BTA3 based device reaches up to 50%. The according current density obtained by the integration of the EQE curves are 0.24, 0.54, 8.85 mA cm−<sup>2</sup> , respectively, which are consistent with the JSC values from the J–V curves within 5% error. Therefore, the OSCs based on PIDT-DTffBTA: BTA3 exhibits the best performance with a maximal PCE of 5.67%, which is among the highest values reported in the literature to date for NF OSCs with VOC >1.20V.

To study the cause of the different JSC, we first investigate the exciton generation and separation by measuring the photoluminescence (PL) in these blends. As shown in **Figure 6A**, at the excitation wavelength of 480 nm, emission from the blend with BTA2 is close to the emission of the pristine PIDT-DTffBTA film and the PL quenching efficiency is only 16% (Li Z. et al., 2016). The inefficient quenching of PL indicates the excitons quick recombination rather than efficient separation. The shape of PL from the blend with BTA1 is close to the emission of the pristine BTA1 film but the polymer PL quenching efficiency significantly raises up to 88%, indicating that excitons initially generated on the PIDT-DTffBTA can transfer to BTA1 in high yield but they are inefficiently quenched by the heterojunction (Hoke et al., 2013). Considering the fine film morphology (as shown in **Figure 7**), the inefficient quenching of PL in PIDT-DTffBTA:BTA2 and PIDT-DTffBTA:BTA1 is owing to the too small 1ELUMOD−LUMO<sup>A</sup> (0.05 eV) and 1EHOMOD−HOMO<sup>A</sup> (0.12 eV), respectively. The too small energy offsets could reduce the overall exciton dissociation efficiency and create exergonic pathways for charge recombination of holes in PIDT-DTffBTA or electrons in BTA1, obviously increasing the voltage loss. As the increase of 1ELUMOD−LUMO<sup>A</sup> and 1EHOMOD−HOMO<sup>A</sup> , the driving force for the hole and electron tranfer are obviously improved. Hence, BTA1 shows similar results with BTA2 but BTA3 can completely quench the PL of PIDT-DTffBTA with a quenching efficiency of ∼100% and the luminous efficiency of BTA3 in the blend is as low as 25%, as shown in **Figures 6B,C**. The high PL quenching efficiency of both PIDT-DTffBTA and BTA3 suggests the improved electron and hole transfer in BTA3 based devices, which can partially explain its high photocurrent. Though JSC and FF can be improved, they still below 9 mA/cm<sup>2</sup> and 0.60, respectively, which may be attributed to



its incomplete fluorescence quenching of BTA3, resulting in the modest hole transfer from BTA3 to polymer. Besides, time resolved photoluminescence measurements were also performed on neat and blended films with the same excited wavelength (Figure S5). All the time-resolved PL (TRPL) lifetime data fitted with bi-exponential decay model are summarized in **Table 2**. The fast decay is related to the exciton dissociation due to charge transfer at the donor/acceptor interfaces. The little changes in τ and f for the PIDT-DTffBTA:BTA2 and PIDT-DTffBTA:BTA1 blend films can be attributed to the poor exciton separation arising from the limitation of the small driving force. Moreover, the exciton lifetimes of PIDT-DTffBTA:BTA3 based blend film

FIGURE 7 | (A–C) The height images for the PIDT-DTffBTA:BTA2, PIDT-DTffBTA:BTA1, and PIDT-DTffBTA:BTA3 blend films without solvent annealing, respectively; (E–G) The phase images for the PIDT-DTffBTA:BTA2, PIDT-DTffBTA:BTA1, and PIDT-DTffBTA:BTA3 blend films without solvent annealing, respectively; (D,H) Is the height and phase images for the PIDT-DTffBTA:BTA3 blend films with solvent annealing, respectively.

TABLE 2 | Fitting parameters of PL decay dynamics of the various films with excitation at 450 nm.


is clearly quenched as compared to neat films, suggesting both efficient hole and electron transfer at the donor/acceptor interface. Compared to the other two control device, the well improved exciton dissociation efficiency may explain part of the remarkably increased JSC in the PIDT-DTffBTA:BTA3 based device. The PL quenching behavior of the TRPL measurements agrees well with the steady-state PL measurements discussed above.

Atom force microscope (AFM) is performed to investigate if the surface morphologies of the blend films play a role in the different JSC and FF. As shown in the height images (**Figures 7A–C**), the PIDT-DTffBTA:BTA2, PIDT-DTffBTA: BTA1, and PIDT-DTffBTA: BTA3 based films without solvent annealing exhibit smooth and uniform surface morphologies with small root-mean-square roughness (RMS) of 0.95, 0.79, and 0.87 nm, respectively. Interestingly, the RMS in PIDT-DTffBTA: BTA3 based film increases up to 5.29 nm after solvent annealing (**Figure 7D**), which may be a result of the enhanced intermolecular aggregation effect (Figure S4) (Zhong et al., 2017) and increased crystallinity of BTA3 in the blend films (Figure S6) (Li et al., 2013). As shown in the phase images (**Figures 7E–G**), it is clear that all of the blend films exhibit interpenetrating networks with different domain sizes. The three blends of PIDT-DTffBTA: BTAx (x = 1–3) without solvent annealing show thinner domain sizes below 10 nm. While the blend of PIDT-DTffBTA and BTA3 shows the sufficient phase separation behavior with the domain size of 40–50 nm after solvent annealing (**Figure 7H**), which is beneficial to the efficient charge separation and transport, giving rise to the reduced recombination loss and improved JSC and FF.

At last, the space charge limited current (SCLC) method is applied to measure the carriers mobilities (Figure S7). As listed in **Table 1**, the carriers mobilities observed in both PIDT-DTffBTA: BTA2 and PIDT-DTffBTA: BTA1 based devices are very low, on the order of only 10−<sup>5</sup> cm<sup>2</sup> V −1 s −1 for hole mobility (µh) and 10−<sup>7</sup> cm<sup>2</sup> V −1 s −1 for electron mobility (µe). These low and imbalance mobilities could result in the low FF observed in these devices (Earmme et al., 2013; Meng et al., 2015). The µ<sup>h</sup> of BTA3 based device is 3–6 times higher than that of two other devices while the µ<sup>e</sup> is higher about 2 order of magnitude, which is attributed to the dominant face-on orientation (Figure S6). As a result, the transport of holes and electrons in the PIDT-DTffBTA: BTA3 based device is faster and more balanced, which can effectively prevent the accumulation of charge and achieve higher FF and JSC.

## CONCLUSIONS

In this work, we applied a new design concept to construct the OSCs with ultrahigh VOC, utilizing the same building blocks (indacenodithiophene and benzotriazole) to design both p-type polymeric donor and n-type small molecular acceptors. The resulted non-fullerene acceptors showed high-lying LUMO levels, close to that of the donor polymer. With small voltage loss (0.55–0.60 V), all of the three devices show ultra-high VOC (1.21– 1.37 V). With the increase of the ability of attracting electron of acceptor, PIDT-DTffBTA: BTA3 device possesses the large driving force for efficient electron and hole transfer and yields a dramatically increased JSC. The achieved PCE of 5.67% is among the highest values for NF OSCs with a VOC >1.20 V reported so far (as shown in Table S8). The results here offer a new method to construct the promising OSCs with ultra-high VOC, which could contribute to further improve the performance of the OSCs.

## AUTHOR CONTRIBUTIONS

Device fabrication and photovoltaic performance studies were carried out by AT and JY. Materials synthesis was carried out by BX and JL. EZ, FC, and XW contributed to project planning and discussions. EZ had the idea, led the project, and prepared

### REFERENCES


the manuscript. All authors contributed to the manuscript preparation.

## ACKNOWLEDGMENTS

This work is supported by the National Natural Science Foundation of China (Nos. 21602040, 51473040, 51673048, 21504019, 51773046), the National Natural Science Foundation of Beijing (No. 2162045), the Chinese Academy of Sciences (QYZDB-SSW-SLH033) and the National Key Research and Development Program of China (2017YFA0206600).

## SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem. 2018.00147/full#supplementary-material


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2018 Tang, Chen, Xiao, Yang, Li, Wang and Zhou. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.