Medium Bandgap Polymers for Efficient Non-Fullerene Polymer Solar Cells—An In-Depth Study of Structural Diversity of Polymer Structure

A series of medium bandgap polymer donors, named poly(1-(5-(4,8-bis(5-(2-ethylhexyl)-4-fluorothiophen-2-yl)benzo [1,2-b:4,5-b′]dithiophen-2-yl)thiophen-2-yl)-5-((4,5-dihexylthiophen-2-yl)methylene)-3-(thiophen-2-yl)-4H-cyclopenta[c]thiophene-4,6(5H)-dione) (IND-T-BDTF), poly(1-(5-(4,8-bis(5-(2-ethylhexyl)-4-fluorothiophen-2-yl)benzo [1,2-b:4,5-b′]dithiophen-2-yl)-4-hexylthiophen-2-yl)-5-((4,5-dihexylthiophen-2-yl)methylene)-3-(4-hexylthiophen-2-yl)-4H-cyclopenta[c]thiophene-4,6(5H)-dione (IND-HT-BDTF), and poly(1-(5-(4,8-bis(5-(2-ethylhexyl)-4-fluorothiophen-2-yl)benzo [1,2-b:4,5-b′]dithiophen-2-yl)-6-octylthieno [3,2-b]thiophen-2-yl)-5-((4,5-dihexylthiophen-2-yl)methylene)-3-(6-octylthieno [3,2-b]thiophen-2-yl)-4H-cyclopenta[c]thiophene-4,6(5H)-dione (IND-OTT-BDTF), are developed for non-fullerene acceptors (NFAs) polymer solar cells (PSCs). Three polymers consist of donor-acceptor building block, where the electron-donating fluorinated benzodithiophene (BDTF) unit is linked to the electron-accepting 4H-cyclopenta[c]thiophene-4,6(5H)-dione (IND) derivative via thiophene (T) or thieno [3,2-b]thiopene (TT) bridges. The absorption range of the polymer donors based on IND in this study shows 400~800 nm, which complimenting the absorption of Y6BO (600~1000 nm). The PSC’s performances are also significantly impacted by the π-bridges. NFAs inverted type PSCs based on polymer donors and Y6BO acceptor are fabricated. The power conversion efficiency (PCE) of the device based on IND-OTT-BDTF reaches up to 11.69% among all polymers with a short circuit current of 26.37 mA/cm2, an open circuit voltage of 0.79 V, and a fill factor of 56.2%, respectively. This study provides fundamental information on the invention of new polymer donors for NFA-based PSCs.


Synthesis and Characterization
The synthesis procedure thorough description is provided in Scheme 1 which illustrates the synthesis of monomers and polymers. The structure of each compound was confirmed by 1 H and 13 C NMR analysis (Figures S1~S25).  Figure S26, the polymers had good thermal stability, IND-T-BDTF, IND-HT-BDTF, and IND-OTT-BDTF start to decompose at temperatures (T d , 5% weight loss) of 362, 361, and 385 • C, respectively. In differential scanning calorimetry, no observable melting behavior or glass transition was seen.
complemented with the absorption of Y6BO (600~1000 nm). In addition, the polymer donors may exhibit very good compatibility with NFA in the blend due to the structural similarity. Inverted PSCs type with a configuration of ITO/ZnO/polymer donor:Y6BO/MoO3/Ag were fabricated and tested. The PSC of the IND-OTT-BDTF device exhibited the highest PCE of 11.7% with a short circuit current ( ) of 26.4 mA/cm 2 , an open circuit voltage ( ) of 0.79 V, and a fill factor (FF) of 56.2%, respectively.

Synthesis and Characterization
The synthesis procedure thorough description is provided in Scheme 1 which illustrates the synthesis of monomers and polymers. Compound 5 was prepared by the Knoevenagel condensation reaction between IND and 4. Compounds 6, 7, and 8 were prepared by the Stille coupling reaction between compound 5 and corresponding tributyl tin compounds with fairly high yields of 81.5, 82.9, and 59.4%, respectively. The Stille coupling reactions between BDTF and M1, M2, or M3 afforded the polymers IND-T-BDTF, IND-HT-BDTF, and IND-OTT-BDTF, respectively. In chlorinated organic solvents such as chloroform and chlorobenzene, the polymers were completely soluble. The number average molecular (Mn) weight/polydispersity index (PDI) of polymers IND-T-BDTF, IND-HT-BDTF, and IND-OTT-BDTF were 4.3 kDa/2.15, 7.5 kDa/1.62, and 22.5 kDa/2.42, respectively. Due to many insoluble parts after Soxhlet extraction by chloroform, IND-T-BDTF and IND-HT-BDTF possessed low molecular weight. According to Figure S26, the polymers had good thermal stability, IND-T-BDTF, IND-HT-BDTF, and IND-OTT-BDTF start to decompose at temperatures (Td, 5% weight loss) of 362, 361, and 385 °C, respectively. In differential scanning calorimetry, no observable melting behavior or glass transition was seen.

Optical and Electrochemical Behaviors
The absorption spectra of three polymer films are depicted in Figure 2 and their optical characteristics are summarized in Table 1. IND-T-BDTF film showed two broad absorption bands. The backbone's π-π* transition corresponds to a former absorption band at 400~500 nm. An absorption band in the longer wavelength region refers to intramolecular charge transfer (ICT) between the donor (BDTF) and acceptor (IND), a typical property of polymers with D-A arrangement, at a longer wavelength region (550~800 nm). The maximum absorption wavelengths of π-π* transition of polymer films were almost identical. The maximum absorption wavelength of ICT in IND-HT-BDTF film exhibited at 636 nm, which is shorter than that of IND-T-BDTF. This is due to the dihedral angle (see Figure S4) between the 3-hexylthiophene ring and IND of IND-HT-BDTF being larger than that of the IND-T-BDTF. The maximum absorption wavelength ICT in IND-OTT-BDTF film appeared at 668 nm, which is red-shifted than those of the IND-HT-BDTF and IND-T-BDTF. This suggests that thieno [3,2-b]thiopene (TT) exhibited higher aromaticity compared to thiophene, indicating increased electron delocalization in the polymer backbone [31,32]. The optical/electrochemical bandgaps (Table 1)

Optical and Electrochemical Behaviors
The absorption spectra of three polymer films are depicted in Figure 2 and their optical characteristics are summarized in Table 1. IND-T-BDTF film showed two broad absorption bands. The backbone's π-π* transition corresponds to a former absorption band at 400~500 nm. An absorption band in the longer wavelength region refers to intramolecular charge transfer (ICT) between the donor (BDTF) and acceptor (IND), a typical property of polymers with D-A arrangement, at a longer wavelength region (550~800 nm). The maximum absorption wavelengths of π-π* transition of polymer films were almost identical. The maximum absorption wavelength of ICT in IND-HT-BDTF film exhibited at 636 nm, which is shorter than that of IND-T-BDTF. This is due to the dihedral angle (see Figure S4) between the 3-hexylthiophene ring and IND of IND-HT-BDTF being larger than that of the IND-T-BDTF. The maximum absorption wavelength ICT in IND-OTT-BDTF film appeared at 668 nm, which is red-shifted than those of the IND-HT-BDTF and IND-T-BDTF. This suggests that thieno [3,2-b]thiopene (TT) exhibited higher aromaticity compared to thiophene, indicating increased electron delocalization in the polymer backbone [31,32]. The optical/electrochemical bandgaps (Table 1)   The energy levels of the polymers were measured by cyclic voltammetry. According to Figure  The polymers and the materials used in this research with those energy level diagrams are displayed in Figure 2b. Facile charge separation and transport processes are expected to happen in the devices. We also performed photoluminescence (PL) experiments to further investigate the exciton dissociation and charge transfer behavior from polymer donors to Y6BO. The PL spectra in Figure S28 showed broad emission at 700~850 nm. The fact that the PL emissions from polymer blend films containing Y6BO were almost quenched shows that the exciton dissociation and charge transfer in the blend films have successfully taken place.
The frontier molecular orbitals of the polymers were determined from Density functional theory (DFT) at the B3LYP/6-31G** level of the Gaussian 09 (ver.9.5) software (Wallingford, CT, UK) [33]. For making computation easier, the polymer alkyl chains were represented by methyl groups, and two repeating units to represent the polymer itself. Wave functions in the HOMO state of the polymers were delocalized along the backbone and BDTF unit, as shown in Figure S29

Photovoltaic Property
Inverted-type PSCs with a configuration of ITO/ZnO/donor polymers:Y6BO/MoO3/Ag were evaluated to know the photovoltaic performances of polymers. First, several processing parameters, including the active layer thickness and D-A blend ratios were investigated for observing how they affect the photovoltaic performance. The optimum blend ratio between the polymers and Y6BO was performed at 3:4  The energy levels of the polymers were measured by cyclic voltammetry. According to Figure  The polymers and the materials used in this research with those energy level diagrams are displayed in Figure 2b. Facile charge separation and transport processes are expected to happen in the devices. We also performed photoluminescence (PL) experiments to further investigate the exciton dissociation and charge transfer behavior from polymer donors to Y6BO. The PL spectra in Figure S28 showed broad emission at 700~850 nm. The fact that the PL emissions from polymer blend films containing Y6BO were almost quenched shows that the exciton dissociation and charge transfer in the blend films have successfully taken place.
The frontier molecular orbitals of the polymers were determined from Density functional theory (DFT) at the B3LYP/6-31G** level of the Gaussian 09 (ver.9.5) software (Wallingford, CT, UK) [33]. For making computation easier, the polymer alkyl chains were represented by methyl groups, and two repeating units to represent the polymer itself. Wave functions in the HOMO state of the polymers were delocalized along the backbone and BDTF unit, as shown in Figure S29

Photovoltaic Property
Inverted-type PSCs with a configuration of ITO/ZnO/donor polymers:Y6BO/MoO 3 /Ag were evaluated to know the photovoltaic performances of polymers. First, several processing parameters, including the active layer thickness and D-A blend ratios were investigated for observing how they affect the photovoltaic performance. The optimum blend ratio between the polymers and Y6BO was performed at 3:4 and with 130 nm for IND-T-BDTF and (Table S1). Figure 3a,b display the current density (J) vs applied voltage (V) curves under illumination and dark with photovoltaic parameters of the devices, and their optimum processing conditions are summarized in Table 2. and with 130 nm for IND-T-BDTF and 120 nm for IND-HT-BDTF and IND-OTT-BDTF active layer optimum thickness achieved (Table S1). Figure 3a,b display the current density ( ) vs applied voltage ( ) curves under illumination and dark with photovoltaic parameters of the devices, and their optimum processing conditions are summarized in Table 2.         We measured the devices and vs. light intensity to better understand the charge recombination mechanisms. From the relationship between the vs. illuminated light intensity ( ℎ ) (expressed by = ( ℎ ) α ), the bimolecular recombination process can be observed. According to Figure 5a, the α values of IND-T-BDTF, IND-HT-BDTF, and IND-OTT-BDTF based devices were 0.76, 0.89, and 1.05, respectively. As for the device with IND-OTT-BDTF, the bimolecular recombination process was minimized in the device [37]. Thus, IND-T-BDTF shows poorer PCE than those based on IND-HT-BDTF, and IND-OTT-BDTF devices. Additionally, using the formula = ( kT/q) × ln ( ℎ ) (where , , and T are the Boltzmann constant, elementary charge, and the temperature in Kelvin), it is possible to determine the device trap-assisted recombination. If n becomes 1, the band-to-band recombination process is dominated in the device. Trap-assisted recombination mechanism predominates in the devices when n is close to 2 [21]. In contrast to the device with IND-T-BDTF (n = 1.47) which demonstrated the most undesired trap- We measured the devices J sc and V OC vs. light intensity to better understand the charge recombination mechanisms. From the relationship between the J sc vs. illuminated light intensity (P light ) (expressed by J sc = (P light ) α ), the bimolecular recombination process can be observed. According to Figure 5a, the α values of IND-T-BDTF, IND-HT-BDTF, and IND-OTT-BDTF based devices were 0.76, 0.89, and 1.05, respectively. As for the device with IND-OTT-BDTF, the bimolecular recombination process was minimized in the device [37]. Thus, IND-T-BDTF shows poorer PCE than those based on IND-HT-BDTF, and IND-OTT-BDTF devices. Additionally, using the formula V OC = (nkT/q) × ln (P light ) (where k, q, and T are the Boltzmann constant, elementary charge, and the temperature in Kelvin), it is possible to determine the device trap-assisted recombination. If n becomes 1, the band-to-band recombination process is dominated in the device. Trap-assisted recombination mechanism predominates in the devices when n is close to 2 [21].

Morphology Study
The molecular ordering information is very important for understanding the overall photovoltaic properties of PSCs. To understand the ordering features of the active layers, we measured grazing incidence wide-angle X-ray scattering (GIWAXS). Figure 6 showed GIWAXS images (Figure 6a,c) and direction line cuts in-plane (IP) and out-of-plane (OOP) (Figure 6b,d) of neat polymers and polymer:Y6BO blending film. GIWAXS film was prepared in the same way as the preparation of devices on the silicon wafer. As shown in Figure 6b (OOP direction), IND-T-BDTF and IND-OTT-BDTF films exhibited a broad (010) peak at 1.71 and 1.65 Å −1 , respectively, corresponding to π-π (intermolecular) stacking distances of 3.67 and 3.81 Å, respectively. As for IND-HT-BDTF film, very weak (010) peak at 1.68 Å −1 (3.74 Å). The intermolecular stacking distances were increased in the order of IND-T-BDTF < IND-HT-BDTF < IND-OTT-BDTF due to the alkyl substituents on the πbridges on T and TT. By a peak (010) in the OOP direction cut, it is preferable to have the face-on orientation to the surface. This means that the vertical charge transport is favorable in the device [38]. The lamellar domain is represented by a broad (100) peak in the IND-T-BDTF, IND-HT-BDTF, and IND-OTT-BDTF at 0.274 (22.9), 0.266 (23.6), and 0.267 (23.5) Å −1 (Å) through the IP direction, respectively. Alkyl substituents on the π-bridges such as thiophene (T) and TT also affect the lamellar domain spacing distances. As shown in Figure 6d, a broad peak at 1.70 (3.70) Å −1 (Å) in OOP directions appeared, which is almost the same as the peak in Y6BO film (1.75 Å −1 ) (See Figure S32). Interestingly, two unknown scattering patterns in IP directions at 1.36 and 1.70 Å −1 for the blend films were found. The strength of the (010) peak along the OOP direction in blend films is more pronounced than the peak in neat polymer films. This indicates that the Y6BO acceptor is the key factor for the face-on orientation in the blend films. Although blend films primarily refer to the face-on molecular packing orientation of the Y6BO acceptor, the blend films may also form in a favored face-on orientation owing to D-A strong intermolecular interactions. Moreover, to know the active layer morphology, we also used transmission electron microscopy (TEM) and atomic force microscopy (AFM). The TEM images of the active layer ( Figure S33) showed that the IND-HT-BDTF and IND-OTT-BDTF blend films may create bicontinuous interpenetrating networks and superior nanoscale phase separation than the IND-T-BDTF blend films, respectively. Therefore, by effective charge-separation or transport, phase separation is preferred in the active layer based on IND-HT-BDTF and IND-OTT-BDTF can promote higher PCE of the related PSCs. The AFM height and phase images ( Figure S34

Morphology Study
The molecular ordering information is very important for understanding the overall photovoltaic properties of PSCs. To understand the ordering features of the active layers, we measured grazing incidence wide-angle X-ray scattering (GIWAXS). Figure 6 showed GIWAXS images (Figure 6a,c) and direction line cuts in-plane (IP) and out-of-plane (OOP) (Figure 6b,d) of neat polymers and polymer:Y6BO blending film. GIWAXS film was prepared in the same way as the preparation of devices on the silicon wafer. As shown in Figure 6b (OOP direction), IND-T-BDTF and IND-OTT-BDTF films exhibited a broad (010) peak at 1.71 and 1.65 Å −1 , respectively, corresponding to π-π (intermolecular) stacking distances of 3.67 and 3.81 Å, respectively. As for IND-HT-BDTF film, very weak (010) peak at 1.68 Å −1 (3.74 Å). The intermolecular stacking distances were increased in the order of IND-T-BDTF < IND-HT-BDTF < IND-OTT-BDTF due to the alkyl substituents on the π-bridges on T and TT. By a peak (010) in the OOP direction cut, it is preferable to have the face-on orientation to the surface. This means that the vertical charge transport is favorable in the device [38]. The lamellar domain is represented by a broad (100) peak in the IND-T-BDTF, IND-HT-BDTF, and IND-OTT-BDTF at 0.274 (22.9), 0.266 (23.6), and 0.267 (23.5) Å −1 (Å) through the IP direction, respectively. Alkyl substituents on the π-bridges such as thiophene (T) and TT also affect the lamellar domain spacing distances. As shown in Figure 6d, a broad peak at 1.70 (3.70) Å −1 (Å) in OOP directions appeared, which is almost the same as the peak in Y6BO film (1.75 Å −1 ) (See Figure S32). Interestingly, two unknown scattering patterns in IP directions at 1.36 and 1.70 Å −1 for the blend films were found. The strength of the (010) peak along the OOP direction in blend films is more pronounced than the peak in neat polymer films. This indicates that the Y6BO acceptor is the key factor for the face-on orientation in the blend films. Although blend films primarily refer to the face-on molecular packing orientation of the Y6BO acceptor, the blend films may also form in a favored face-on orientation owing to D-A strong intermolecular interactions. Moreover, to know the active layer morphology, we also used transmission electron microscopy (TEM) and atomic force microscopy (AFM). The TEM images of the active layer ( Figure S33 [3,2-b]indole-2,10-diyl)bis (methanylylidene))bis(5,6-difluoro-3-oxo-2,3-dihydro-1H-indene-2,1-diylidene)) dimalononitrile (Y6BO) [9], and (4,8-bis(5-(2-ethylhexyl)-4-fluorothiophen-2-yl)benzo-[1,2-b:4,5-b′]dithiophene-2,6-diyl)bis(trimethylstannane)(BDTF) [23] were synthesized according to previous reports. 4,6-Dibromo-1H,3H-thieno [3,4-c] furan-1,2-dione was purchased form Sunatech. All other chemicals used in this work were purchased from Sigma Aldrich Co and Alfa Aesar (A Johnson Matthey Company), and used without any further purification unless otherwise described. The 1 H and 13 C NMR spectra were measured with a JEOL JNM ECP-400 spectrometer. UV visible spectra were recorded on a JASCO V730 UV/Vis spectrophotometer. Matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) spectroscopy was conducted by using a Bruker Ultraflex spectrometer. Gel permeation chromatography (GPC) was measured on an Agilent 1200 series instrument with THF as the eluent. The thermogravimetric analysis (TGA) was carried out under the N2 atmosphere at a heating rate of 10 °C/min with TA Instrument Q600 (PH407 PUSAN KBSI). Cyclic voltammetry (CV) measurements were carried out by using a VersaSTAT3 potentiostat (Princeton Applied Research) with tetrabutylammonium hexafluorophosphate (0.1 M, Bu4NPF6) as the electrolyte in acetonitrile. The films thickness was measured with an Alpha-Step IQ surface profiler (KLA-Tencor Co., Milpitas, CA, USA) Grazingincidence wide-angle X-ray scattering (GIWAXS) spectra were obtained on the 3C beamline with 13 keV (λ = 0.123 nm) X-ray irradiation source and the beam size of 300 μm (height) × 23 μm (width) in the Pohang Accelerator Laboratory (PAL). A two-dimensional charge-coupled device detector (Mar165 CCD) was used, and the distance from the sample to the detector was 0.2 m. The X-ray beam angle of the incidence was chosen such that the beam would penetrate the entire active layer while minimizing scattering from the substrate: ~0.12°. The samples were partially completed devices so that the entire exposed surface is composed of an active layer on the Si wafer and were examined under ambient. Preparation of film for GIWAXS was followed the same as the preparation of the active layer. The ZnO layer was deposited on the Si wafer by sol-gel process giving a film of 25nm-thick. The polymer or blended polymer film (polymeric donor and Y6BO acceptor) was fabricated by spin-coating in chloroform with 0.5% of 1-chloronaphthalene (CN) as a processing additive. Then the film was annealed at 100 °C for 10 min in the glove box. The  , and used without any further purification unless otherwise described. The 1 H and 13 C NMR spectra were measured with a JEOL JNM ECP-400 spectrometer. UV visible spectra were recorded on a JASCO V730 UV/Vis spectrophotometer. Matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) spectroscopy was conducted by using a Bruker Ultraflex spectrometer. Gel permeation chromatography (GPC) was measured on an Agilent 1200 series instrument with THF as the eluent. The thermogravimetric analysis (TGA) was carried out under the N 2 atmosphere at a heating rate of 10 • C/min with TA Instrument Q600 (PH407 PUSAN KBSI). Cyclic voltammetry (CV) measurements were carried out by using a VersaSTAT3 potentiostat (Princeton Applied Research) with tetrabutylammonium hexafluorophosphate (0.1 M, Bu 4 NPF 6 ) as the electrolyte in acetonitrile. The films thickness was measured with an Alpha-Step IQ surface profiler (KLA-Tencor Co., Milpitas, CA, USA) Grazing-incidence wide-angle X-ray scattering (GIWAXS) spectra were obtained on the 3C beamline with 13 keV (λ = 0.123 nm) X-ray irradiation source and the beam size of 300 µm (height) × 23 µm (width) in the Pohang Accelerator Laboratory (PAL). A two-dimensional charge-coupled device detector (Mar165 CCD) was used, and the distance from the sample to the detector was 0.2 m. The X-ray beam angle of the incidence was chosen such that the beam would penetrate the entire active layer while minimizing scattering from the substrate:~0.12 • . The samples were partially completed devices so that the entire exposed surface is composed of an active layer on the Si wafer and were examined under ambient. Preparation of film for GIWAXS was followed the same as the preparation of the active layer. The ZnO layer was deposited on the Si wafer by sol-gel process giving a film of 25-nm-thick. The polymer or blended polymer film (polymeric donor and Y6BO acceptor) was fabricated by spin-coating in chloroform with 0.5% of 1-chloronaphthalene (CN) as a processing additive. Then the film was annealed at 100 • C for 10 min in the glove box. The scattering vector (q) and d spacing (d) were calculated from the equation: q = 4π sin (φ)/λ and q = 2π/d. Photoluminescence spectra of the polymers and blended films were obtained by a HOMOBA (fluolog-QM). The blend morphology was examined by the transmission electron microscope (HITACHI Hightech. HT-7800).

Synthesis of Monomers and Polymers
1 H and 13 C NMR spectra of synthesized compounds are displayed in supporting information ( Figure S1 to Figure S25).