All-Small-Molecule Organic Solar Cells Based on a Fluorinated Small Molecule Donor With High Open-Circuit Voltage of 1.07 V

A new small molecule donor with an acceptor-donor-acceptor (A-D-A) structure, namely DRTB-FT, has been designed and synthesized for all-small-molecule organic solar cells (ASM-OSCs). By introducing fluorine atoms on the thienyl substituent of the central benzodithiophene unit, DRTB-FT shows a low-lying highest occupied molecular orbital (HOMO) energy level of −5.64 eV. Blending with an A-D-A type acceptor F-2Cl, DRTB-FT based ASM-OSCs gave a power conversion efficiency (PCE) of 7.66% with a high open-circuit voltage (Voc) of 1.070 V and a low energy loss of 0.47 eV. The results indicate that high Voc of ASM-OSC devices can be obtained through careful donor molecular optimization.

From the perspective of donor materials, NFA based OSCs consist of polymer solar cells and small-molecule solar cells. Using polymer donor materials, both single junction and tandem OSCs have achieved the impressive power conversion efficiencies (PCEs) of 15-18% (Meng et al., 2018;An et al., 2019;Fan et al., 2019;Jiang et al., 2019;Lin et al., 2019;Wu Y. et al., 2019;Xiong et al., 2019;Yan et al., 2019). So far, all-small-molecule OSCs (ASM-OSCs), i.e., devices based on both small molecule donor and acceptor, have shown relatively lower performance compared with that of polymer donor based devices (Zhang Q. et al., 2014;Kan et al., 2015;Badgujar et al., 2016;Yang et al., 2017;Privado et al., 2018;Wang et al., 2018a;Duan et al., 2019). Nevertheless, with the advantages of versatile molecules design, less variation among different batches and thus easiness of property control for small molecules, ASM-OSCs are strongly believed to have the potential to achieve and even surpass the performance of polymer donor counterparts (Chen et al., 2013). To date, with the extensive study of small molecular design and device optimization for ASM-OSCs, the PCEs over 10% have been achieved (Bin et al., 2018;Wang et al., 2018a,b;Chen et al., 2019;Cheng et al., 2019;Gao et al., 2019;Ge et al., 2019;Yue et al., 2019;Zhou et al., 2019). Notably, a PCE over 14% has been obtained just recently by Wei et al. indicating the great potential of ASM-OSCs .
Presently, compared with the comprehensive study of small molecule NFA, less attention has been paid to small molecule donors, which in fact play a vital role same as acceptors in OSCs. Among many properties that should be considered for designing active layer materials, matched energy levels between donor and acceptor not only contribute to charge transfer and exciton separation, but also have great influence on the opencircuit voltage (V oc ) of OSC devices (Tang et al., 2017. Therefore, it is important to efficiently tune energy levels for designing active layer molecules. Among various strategies for tuning the energy level, introducing fluorine atoms into the designing molecule has been proved to be a simple way to downshift the frontier molecular orbital energy level, which has been used successfully in molecular design (Tang et al., 2018; Chen et al., 2019;Ge et al., 2019;Yue et al., 2019), e.g., PM6, a well-known polymer donor synthesized by Hou's group . Recently, we have reported an ASM-OSC based on a non-fluorinated donor DRTB-T and an acceptor F-2Cl, showing a PCE of 10.76%, but a relatively lower V oc of 0.969 V compared with other ASM-OSCs (Wang et al., 2018b). With these considerations, we have designed and synthesized a new small molecule donor with an A-D-A structure, namely DRTB-FT (as shown in Figure 1A), in which fluorine atoms are introduced on the thienyl substituent of central benzodithiophene (BDT) unit . As expected, DRTB-FT shows a lower highest occupied molecular orbital (HOMO) energy level of −5.64 eV in contrast to DRTB-T (−5.51 eV) . Notably, DRTB-FT has nearly the same solid film absorption ranging from 300 to 600 nm compared with that of DRTB-T, which is well complementary to that of an A-D-A type acceptor F-2Cl (500-800 nm). Thus, ASM-OSC based on DRTB-FT:F-2Cl gave a PCE of 7.66% with a high V oc of 1.070 V and a low energy loss (E loss ) of 0.47 eV. The results indicate that high V oc of ASM-OSC device can be obtained through careful donor molecular optimization. The moderate performance was mainly attributed to the low FF (0.532) and J sc (-13.46 mA cm −2 ) owing to the unmatched charge carrier mobilities and unfavorable active layer morphology.

Materials and Synthesis
DRTB-FT was synthesized following the procedure in Scheme 1. F-2Cl was prepared according to the reported procedure (Wang et al., 2018b). The materials were commercially available and used without further purification. All manipulations and reactions were carried out following the standard Schlenk technique under argon protection.

Measurements and Instruments
The 1 H and 13 C nuclear magnetic resonance (NMR) spectra were taken on by the Bruker AV400 Spectrometer. Matrix assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass were measured using a Bruker Autoflex III instrument. Xray diffraction (XRD) experiments were carried out with a Bruker Model D8 FOCUS X-ray diffractometer with Cu Kα radiation (λ = 1.5406 Å) at a current of 40 mA and a generator voltage of 40 kV. The UV-vis spectra were performed on a JASCO V-570 spectrophotometer. Cyclic voltammetry (CV) experiments were carried out using a LK98B II Microcomputerbased Electrochemical Analyzer and a conventional threeelectrode configuration with a saturated calomel electrode (SCE) as the reference electrode, a Pt wire as the counter electrode, and a glassy carbon electrode as the working electrode at room temperature. The measurement was carried out employing acetonitrile solution of tetrabutylammonium phosphorus hexafluoride (n-Bu 4 NPF 6 , 0.1 M) as the supporting electrolyte with the scan rate of 100 mV s −1 . The LUMO (lowest unoccupied molecular orbital) and HOMO values were obtained following the formula E LUMO = -(4.80 + E re onset ) and E HOMO = -(4.80 + E ox onset ), where E re onset and E ox onset can be estimated from the CV curves, respectively. The geometry structures of DRTB-FT was optimized through Density functional theory (DFT) calculations (B3LYP/6-31G * ) according to Gaussian 16 (Lee et al., 1988;Frisch et al., 2016).
The current density-voltage (J-V) characteristics of ASM-OSC devices were measured with a Keithley 2400 source-measure unit. Using a SAN-EI XES-70S1 AAA class solar simulator, photocurrent was performed under illumination with simulated 100 mW cm −2 AM 1.5G irradiation. The external quantum efficiency (EQE) curve was obtained by a QE-R Solar Cell Spectral Response Measurement System. SCHEME 1 | Synthetic route of DRTB-FT.
Frontiers in Chemistry | www.frontiersin.org Atomic force microscope (AFM) and Transmission electron microscopy (TEM) images were obtained on a Bruker MultiMode 8 in tapping mode and Philips Technical G2 F20 at 200 kV, respectively. Space charge limited current (SCLC) mobility was performed by a diode configuration of glass/ZnO/active layer/Al for electron mobility and ITO/PEDOT:PSS/active layer/Au for hole mobility, respectively. The results were fitted via a space charge limited form, being described as the equation: where J is the current density, ε 0 is the permittivity of free space (8.85 × 10 −12 F m −1 ), ε r is the relative dielectric constant of the transport medium, µ 0 is the hole or electron mobility, V (= V appl -V bi ) is the internal voltage in the device, where V appl is the applied voltage to the device and V bi is the built-in voltage due to the relative work function difference of the two electrodes, L is the film thickness of the active layer.

Fabrication of OSC Devices
ASM-OSC devices were fabricated using a conventional structure of glass/ITO/PEDOT:PSS/Donor:Acceptor/PDINO/Al, in which the electronic transport material PDINO was designed by . As a common procedure: the cleaned ITOcoated glasses were treated by ultraviolet-ozone for 15 min and spin-coated with PEDOT:PSS solution at 4,800 rpm for 20 s. Subsequently, the glasses were baked at 150 • C for 20 min and transferred into the glove box with argon gas. Chloroform solution of both donor and acceptor was spin-coated on the PEDOT:PSS at 1,700 rpm for 20 s. Afterward, PDINO was spin-coated on the active layer at 3,000 rpm for 20 s with the concentration of 1.0 mg/ml using EtOH as solvent. Finally, under the high vacuum, a cathode material Al was deposited for 50 nm onto PDINO layer. The work area of each device is about 0.04 cm 2 .

Materials Synthesis and Characterization
The synthesis of DRTB-FT is illustrated in Scheme 1. By Stille coupling reaction between compound 1 and 2, two commercially available BDT derivatives, compound 3 was synthesized. Subsequent Knoevenagel condensation reaction between compound 3 and 3-ethylrhodanine gave the final product in high yield. As shown in the experimental section above, DRTB-FT was characterized through NMR spectroscopy and MALDI-TOF MS. DRTB-FT exhibits good solubility in common organic solvents such as chlorobenzene, dichloromethane and chloroform. As shown in Figures 1B,C, the optimized geometry of DRTB-FT based on DFT (B3LYP/6-31G * ) displays a good coplanar conformation. The structural ordering of pristine DRTB-FT and DRTB-T film was investigated by XRD analysis. According to the XRD results ( Figure S4), a weak diffraction peak (100) at 4.44 (2θ) was observed for the DRTB-T film. In contrast,  Yang et al. (2017). DRTB-FT film exhibits a strong diffraction peak (100) at 4.46 (2θ), indicating stronger stacking and crystallinity than DRTB-T.

Optical and Electrochemical Properties
The UV-vis absorption spectra of DRTB-FT obtained in chloroform solution and in solid state are shown in Figure 1D. In chloroform solution, DRTB-FT exhibits an absorption maximum at 547 nm, and it shows a bathochromic-shifted and broadened absorption in the solid state ranging from 300 to 630 nm, which is well complementary to that of F-2Cl film. The optical bandgap (E g opt ) obtained from DRTB-FT film absorption onset (622 nm) is 1.99 eV. Moreover, a stronger shoulder peak was observed in film absorption spectrum in comparison to the non-fluorinated counterpart DRTB-T, which reveals the stronger intermolecular stacking and π-π interactions in DRTB-FT solid film. The electrochemical behavior of DRTB-FT was investigated by CV experiments. As shown in Figure S5, the estimated reduction and oxidation potentials of DRTB-FT are −1.17V and 0.84V (vs Fc/Fc + ), respectively. Therefore, the first oxidation event of DRTB-FT occurs at a higher oxidation potential compared with DRTB-T, leading to a deeper HOMO of −5.64 eV, which is in favor of obtaining a higher V oc in the devices when combined with appropriate acceptor. The optical and electrochemical data of DRTB-FT and DRTB-T are listed in Table 1.

Photovoltaic Properties
ASM-OSC devices with a conventional structure of ITO/PEDOT:PSS/active layer/PDINO/Al were fabricated. The devices performance were optimized by varying the donor to acceptor weight ratio (D/A, w/w) in the active layer, applying solvent vapor annealing (SVA) and thermal annealing (TA) treatment and etc. The detailed device parameters are provided in Tables S1-S3. The photovoltaic parameters of the as-cast and optimal devices are listed in Table 2 and the corresponding J-V characteristics are shown in Figure 2A. For the as-cast device,  a PCE of 5.72% with a high V oc of 1.10 V was obtained. An improved photovoltaic performance with a PCE of 7.66% and a high V oc of 1.07 V was achieved after TA treatment at 100 • C for 10 min, and the corresponding E loss is as low as 0.47 eV calculated via the equation E loss = E g opt -eV oc , where E g opt is the optical bandgap of F-2Cl film (1.54 eV). As expected, OSCs based on DRTB-FT:F-2Cl demonstrated a high V oc owing to the low-lying HOMO energy level of the DRTB-FT donor. In contrast, the DRTB-T:F-2Cl based device has a V oc of 0.969 V. However, limited by the low J sc (13.46 mA cm −2 ) and FF (0.532), the optimized device based on DRTB-FT:F-2Cl gave a moderate PCE (7.66%). The unsatisfactory device performance can be ascribed to the unfavorable active layer morphology and unmatched charge carrier mobilities, as will be discussed in the following section.
The EQE spectrum of the optimal device is presented in Figure 2B, which exhibits a broad photo-to-current response from 300 to 790 nm. However, the maximum efficiency is only 58% at 471 nm, indicating a lower photoelectron conversion process in comparison to that of DRTB-T based device (Wang et al., 2018b). The calculated J sc value obtained from the integration of the EQE curve is 13.01 mA cm −2 , which is in good agreement with the J sc value from the corresponding J-V curve with about 3% mismatch.
To explore the reason for lower J sc and FF of the ASM-OSC, the exciton dissociation and extraction process in the active layer of optimal device was investigated. Figure 3A shows the relation between photocurrent density J ph (J ph = J L -J D ) and effective voltage V eff (V eff = V 0 -V a ), where J L , J D , V 0, and V a are the light current density, the dark current density, the voltage at J ph = 0 and the applied voltage, respectively. It can be seen that the J ph curve of DRTB-FT:F-2Cl based device is hard to reach the saturation current density (J sat ) even at a high V eff . Lower J ph /J sat values of 85.9% and 61.0% were obtained under the short-circuit and maximal output power conditions, respectively, which demonstrated a poor exciton dissociation and charge extraction processes and limited J sc and FF. For further study, the bimolecular recombination degree in active layer was estimated on the basis of the formula of J sc ∝ P α , where α is the exponential factor. When α value was closer to 1, the less bimolecular recombination in blend film. The fitted slope (α) is calculated to be 0.964 for the devices based on DRTB-FT as shown in Figure 3B, which is relatively lower compared with that of DRTB-T based device (0.985) (Wang et al., 2018b), indicating more bimolecular recombination in DRTB-FT:F-2Cl active layer.
The influence of active layer morphology on device performance was investigated by AFM and TEM technology. Figure 4A shows that the surface topography of the as-cast DRTB-FT:F-2Cl blend film is relatively rough in terms of the large root-mean-square (RMS) roughness (9.2 nm), which reveals the existence of excessive aggregation in the as-cast film due to the strong stacking and crystallinity of donor molecules. For the optimal blend film ( Figure 4B), a relatively smaller RMS roughness of 4.4 nm was observed, which is beneficial to form better contact between the film and the electrode, and thus to promote charge extraction. TEM images are in consistent with the AFM, from Figure 4C, it can be seen that an oversized phase separation and more defects in the as-cast blend film, and the morphology becomes better obviously after TA treatment ( Figure 4D). However, there was still no clear fiber structure or suitable phase separation in the optimized blend film, which might account for the low J sc and FF.
To investigate the charge-transporting behavior of the studied devices, the electron and hole mobilities (µ e and µ h ) were measured by the SCLC method as described in experimental section. According to Figure S6, the electron mobility of µ e =2.11 × 10 −4 cm −2 V −1 s −1 and hole mobility of µ h = 4.51 × 10 −5 cm −2 V −1 s −1 were obtained for the as-cast device. After TA treatment, the charge mobilities were raised to µ e = 4.12 × 10 −4 cm −2 V −1 s −1 and µ h = 8.56× 10 −5 cm −2 V −1 s −1 , respectively. The relatively lower and unbalanced hole and electron mobilities could lead to more charge recombination, thus the lower FF and J sc .

CONCLUSION
In summary, we have designed and synthesized a new small molecule and A-D-A type donor named DRTB-FT, taking the fluorothienyl-substituted benzodithiophene as the central unit. The introduction of fluorine atoms enables DRTB-FT to get a low HOMO of −5.64 eV. When combined with an acceptor F-2Cl, the optimal ASM-OSC device gave a PCE of 7.66%, especially a high V oc of 1.070 V and a low E loss of 0.47 eV. The results indicate that high V oc of ASM-OSCs can be obtained through careful donor molecular optimization. On the other hand, morphology control also plays a critical role in ASM-OSCs. In this case, the low J sc and FF are caused by the unfavorable active layer morphology with large aggregation of small molecule donors. It is believed that high performance ASM-OSCs will be achieved through synergistic study of molecular design with complementary absorption, matched energy levels and morphology control.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/Supplementary Material.

AUTHOR CONTRIBUTIONS
CLiu synthesized the DRTB-FT and measured the opticalelectric properties of DRTB-T and DRTB-FT. NQ fabricated and optimized the devices, and wrote the manuscript. YS contributed to the morphology characterization. XK ran the calculations of DRTB-FT, HZ, CLi, XW, and YC conceived and directed the project. All authors contributed to the whole work.