Comparing Benzodithiophene Unit with Alkylthionaphthyl and Alkylthiobiphenyl Side-Chains in Constructing High-Performance Nonfullerene Solar Cells

Using single-bonded and fused aromatic rings are two methods for extending the π-conjugation in the vertical direction of benzo [1,2-b:4,5-b′] dithiophene (BDT) unit. To investigate which method is more efficient in nonfullerene systems, two novel polymers based on alkylthionaphthyl and alkylthiobiphenyl substituted BDT named PBDTNS-FTAZ and PBDTBPS-FTAZ are designed and synthesized. Two polymers only exhibit small differences in structure, but huge differences in photovoltaic properties. They are studied by blended with 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). The device based on PBDTNS-FTAZ:ITIC showed the best power conversion efficiency (PCE) of 9.63% with the Voc of 0.87 V, a Jsc of 18.06 mA/cm2 and a fill factor of 61.21%, while the PBDTBPS-FTAZ:ITIC only exhibit a maximum PCE of 7.79% with a Voc of 0.86 V, a Jsc of 16.24 mA/cm2 and a relatively low fill factor of 55.92%. Therefore, extending π-conjugation with alkylthionaphthyl is more effective against constructing nonfullerene solar cells.


Synthesis of Polymer PBDTNS-FTAZ
Compounds BDTNSSn (128.1 mg, 0.1 mmol), FTAZ-Br (64.5 mg, 0.1 mmol)，Pd2(dba)3 (1.8 mg, 0.002 mmol) and P(o-tol)3 (3.6 mg, 0.012 mmol) were added into a flask. The flask was subjected to more than three successive cycles of vacuum followed by refilling with argon. Then, 4 mL of toluene was added. The reaction mixture was heated to 110 °C under argon atmosphere. Two hours later, the mixture was cooled to room temperature and polymer PBDTNS-FTAZ was precipitated by the addition of methanol, filtered and purified by Soxhlet extraction with methanol, chloroform and odichlorobenzene (o-DCB), respectively. The o-DCB solution was concentrated by evaporation and then precipitated into methanol. The deep purple solid was filtered to yield the desired polymer PBDTPS-FTAZ (118.1 mg, 82% yield). Mn: 46 kDa, PDI: 2.4.

Synthesis of PBDTNS-FTAZ and PBDTBPS-FTAZ
The compounds BDTNSSn and BDTBPSSn were synthesized as previously reported.

Synthesis of Polymer PBDTNS-FTAZ
Compounds BDTNSSn (128.1 mg, 0.1 mmol), FTAZ-Br (64.5 mg, 0.1 mmol), Pd 2 (dba) 3 (1.8 mg, 0.002 mmol) and P(o-tol) 3 (3.6 mg, 0.012 mmol) were added into a flask. The flask was subjected to more than three successive cycles of vacuum followed by refilling with argon. Then, 4 mL of toluene was added. The reaction mixture was heated to 110 • C under argon atmosphere. Two hours later, the mixture was cooled to room temperature and polymer PBDTNS-FTAZ was precipitated by the addition of methanol, filtered and purified by Soxhlet extraction with methanol, chloroform and o-dichlorobenzene (o-DCB), respectively. The o-DCB solution was concentrated by evaporation and then precipitated into methanol. The deep purple solid was filtered to yield the desired polymer PBDTPS-FTAZ (118.1 mg, 82% yield). Mn: 46 kDa, PDI: 2.4.

Device Characterization
The polymer solar cell was fabricated utilizing the conventional device structure of ITO/PEDOT:PSS/Polymer:ITIC/PFN/Al. The active area of solar cells was 0.1 cm 2 . The ITO coated glasses were cleaned in an ultrasonic machine with ITO detergent, deionized water, acetone and isopropanol sequentially for 20 min each step. The cleaned ITO substrates were treated with oxygen plasma for 6 min and then covered with PEDOT: PSS (Baytron PVP Al 4083) by spin coating. The thickness of PEDOT: PSS was about 35 nm. Then, the substrates were annealed at 160 • C for 20 min. Sequentially, the substrate was transferred to the glove box filled with N 2 atmosphere. The o-DCB solutions of polymers and ITIC were initially heated at 120 • C for 30 min and then stirred at room temperature for 6 h. The solutions were heated at 140 • C half an hour before spin coating. The solutions were hot spin-coated at 140 • C to make the active layer and the substrate was heated at 70 • C. After that, the PFN solutions were spin-coated above the active layer at 2500 rpm for 10 s as the buffer layer. Finally, the samples were transferred to a vacant chamber and Al (100 nm) was deposited in a high vacuum degree (5 × 10 −4 Pa) via a mask that constrains the active area of 0.1 cm 2 .
The current density-voltage (J-V) curves were measured by Keithley 2420 source meter under simulated AM 1.5 G irradiation (100 mW /cm 2 ) using a Newport solar simulator. The light intensity was calibrated by a standard silicon solar cell. The external quantum efficiency (EQE) of the solar cells was tested using a certified Newport incident photon conversion efficiency (IPCE) measurement.
Atomic force microscopy (AFM) measurement was performed by an Agilent 5400 with tapping mode. Transmission electron microscopy (TEM) images were obtained from a Hitachi H-7650 transmission electron microscope at an accelerating voltage of 100 kV. The absorption spectra were measured using a Hitachi U-400 UV-vis-NIR scanning spectrophotometer.

Synthesis and Characterization
The synthetic routes of the polymer PBDTNS-FTAZ and PBDTBPS-FTAZ are shown in Scheme 1. The polymers were obtained by a stille coupling reaction. The detailed synthetic procedure is displayed in the Experimental section. The thermal property of the polymers was measured by thermogravimetric analysis (TGA) in nitrogen atmosphere at the heating rate of 10 • C/min. As shown in Figure S1, PBDTNS-FTAZ and PBDTBPS-FTAZ exhibited high thermal stability with an onset decomposition temperature (Td) with a 5% weight loss located at 310 • C.

Optical Properties
The normalized ultraviolet-visible (UV-vis) absorption spectra of PBDTNS-FTAZ and PBDTBPS-FTAZ in dilute o-DCB and as a film are shown in Figure 1. The two polymers show very similar absorption features in both solutions and films. The PBDTNS-FTAZ and PBDTBPS-FTAZ two exhibit shoulder peaks in thin films, which were located at 588 and 597 nm. This could be attributed to the aggregations of the polymer chains. [54] The absorption peak of PBDTNS-FTAZ and PBDTBPS-FTAZ located at 544 and 552 nm, respectively, which can be assigned to the intramolecular charge-transfer between BDT and the FTAZ unit. In the film state, both two polymers exhibit a slight red-shift compared to the absorption spectra in dilute o-DCB solution, which could ascribe to the aggregation in the solid-state. The optical bandgaps (E g opt ) of PBDTNS-FTAZ and PBDTBPS-FTAZ are

Electrochemical Properties
The electrochemical properties of PBDTNS-FTAZ and PBDTBPS-FTAZ were characterized by utilizing the cyclic voltammetry (CV) method. The CV curves are shown in Figure 2

Photovoltaic Properties
To investigate the photovoltaic properties of these two polymers, BHJ polymer solar cells were fabricated using the conventional device structure (ITO/PEDOT:PSS/active layer/PFN/Al). The polymers and ITIC were dissolved in o-DCB. The solutions were initially stirred at 130 °C for one hour and then stirred at room temperature overnight. Devices with different donor/acceptor (D/A) ratios (1:1, 1:1.5, 1:2) were fabricated to explore the best D/A ratio of the blend. Current densityvoltage (J-V) characteristics of all devices were tested under AM 1.5G illumination. The J-V curves are shown in Figure 3 and the corresponding device parameters are summarized in Table 1. From Table 1, it can be concluded that both PBDTNS-FTAZ and PBDTBPS-FTAZ exhibited the best photovoltaic performance when the weight ratio between the polymer and ITIC were 1:1.5. The optimal PCEs (8.17%, 9.64%, 6.49%) of PBDTNS-FTAZ:ITIC were all higher than the PCEs of PBDTBPS-FTAZ:ITIC (7.08%, 7.79%, 7.07%). The PCEs of both polymers did not improve with post thermal annealing or processing additives. The device based on PBDTNS-FTAZ:ITIC showed the best

Electrochemical Properties
The electrochemical properties of PBDTNS-FTAZ and PBDTBPS-FTAZ were characterized by utilizing the cyclic voltammetry (CV) method. The CV curves are shown in Figure

Electrochemical Properties
The electrochemical properties of PBDTNS-FTAZ and PBDTBPS-FTAZ were characterized by utilizing the cyclic voltammetry (CV) method. The CV curves are shown in Figure 2.

Photovoltaic Properties
To investigate the photovoltaic properties of these two polymers, BHJ polymer solar cells were fabricated using the conventional device structure (ITO/PEDOT:PSS/active layer/PFN/Al). The polymers and ITIC were dissolved in o-DCB. The solutions were initially stirred at 130 °C for one hour and then stirred at room temperature overnight. Devices with different donor/acceptor (D/A) ratios (1:1, 1:1.5, 1:2) were fabricated to explore the best D/A ratio of the blend. Current densityvoltage (J-V) characteristics of all devices were tested under AM 1.5G illumination. The J-V curves are shown in Figure 3 and the corresponding device parameters are summarized in Table 1. From Table 1, it can be concluded that both PBDTNS-FTAZ and PBDTBPS-FTAZ exhibited the best photovoltaic performance when the weight ratio between the polymer and ITIC were 1:1.5. The optimal PCEs (8.17%, 9.64%, 6.49%) of PBDTNS-FTAZ:ITIC were all higher than the PCEs of PBDTBPS-FTAZ:ITIC (7.08%, 7.79%, 7.07%). The PCEs of both polymers did not improve with post thermal annealing or processing additives. The device based on PBDTNS-FTAZ:ITIC showed the best

Photovoltaic Properties
To investigate the photovoltaic properties of these two polymers, BHJ polymer solar cells were fabricated using the conventional device structure (ITO/PEDOT:PSS/active layer/PFN/Al). The polymers and ITIC were dissolved in o-DCB. The solutions were initially stirred at 130 • C for one hour and then stirred at room temperature overnight. Devices with different donor/acceptor (D/A) ratios (1:1, 1:1.5, 1:2) were fabricated to explore the best D/A ratio of the blend. Current density-voltage (J-V) characteristics of all devices were tested under AM 1.5G illumination. The J-V curves are shown in Figure 3 and the corresponding device parameters are summarized in Table 1. From Table 1, it can be concluded that both PBDTNS-FTAZ and PBDTBPS-FTAZ exhibited the best photovoltaic performance when the weight ratio between the polymer and ITIC were 1:1.5. The optimal PCEs (8.17%, 9.64%, 6.49%) of PBDTNS-FTAZ:ITIC were all higher than the PCEs of PBDTBPS-FTAZ:ITIC (7.08%, 7.79%, 7.07%). The PCEs of both polymers did not improve with post thermal annealing or processing additives. The device based on PBDTNS-FTAZ:ITIC showed the best PCE of 9.63% with the V oc of 0.87V, a J sc of 18.06 mA/cm 2 and a fill factor of 61.21%, while the PBDTBPS-FTAZ:ITIC only exhibit a maximum PCE of 7.79% with a V oc of 0.86V, a J sc of 16.24 mA/cm 2 and a relatively low fill factor of 55.92%. The V oc , J sc and fill factor of PBDTNS-FTAZ:ITIC were promoted simultaneously compared to PBDTBPS-FTAZ:ITIC. In addition, device stability is another important parameter for solar cells. The storage tests were carried out to investigate the ambient stability of PBDTNS-FTAZ and PBDTBPS-FTAZ-based devices (Figure 4a,b). After storage for 192 h in air, the PBDTNS-FTAZ-based devices exhibited more excellent ambient stability than that of PBDTBPS-FTAZ-based devices.

Polymer
Ratios (w:w) The average values are obtained from 10 devices.    To further investigate the reason higher Jsc of PBDTNS-FTAZ were obtained compared to PBDTBPS-FTAZ. The external quantum efficiency (EQE) was performed and the curves are displayed in Figure 4c. Both the PBDTNS-FTAZ/ITIC and PBDTBPS-FTAZ/ITIC devices showed broad photoresponse from 470 to 720 nm. The EQE response of the PBDTNS-FTAZ-based device were higher than those of PBDTBPS-FTAZ in the whole spectral range, which could partly account for the higher Jsc of the PBDTNS-FTAZ-based device. The optimal devices possess a high and flat EQE value of over 65%. The calculated current density of 15.98 mA/cm 2 and 17.86 mA/cm 2 were obtained by integrating the EQE spectrum with the standard AM 1.5G solar spectrum, which is consistent with the measured Jsc value with little error (<5%).

PBDTNS-FTAZ
The relationship of the photocurrent (Jph) and the effective applied voltage (Veff) were studied to further analyze the charge recombination process in the devices based on PBDTNS-FTAZ and PBDTBPS-FTAZ. The Jph is defined by JL-JD, where JL and JD refer to the current density under AM 1.5G illumination and in the dark, respectively. V is the applied voltage; and V0 refers to the effective voltage where JL=JD [55][56][57][58][59]. As displayed in Figure 5, the Jph of PBDTBPS-FTAZ-based devices could not reach the saturation current (Jsat) even the applied voltage is over 3V, indicating the relatively severe charge recombination in the PBDTBPS-FTAZ:ITIC blend. On the contrary, the Jph of the PBDTNS-FTAZ-based devices reached the saturation value (Jsat) at a low applied voltage of 1 V, proving that the photogenerated excitons in PBDTNS-FTAZ-based devices were fully dissociated to free charges. The Jsat of the PBDTNS-FTAZ-based device was higher than the Jsat of PBDTBPS-FTAZbased device, demonstrating that photogenerated current is larger in PBDTNS-FTAZ devices. The charge recombination and exciton dissociation in the PBDTNS-FTAZ:ITIC and PBDTBPS-FTAZ:ITIC device can be further investigated by the Jph/ Jsat under short-circuit conditions. The Jsat of PBDTNS-FTAZ and PBDTBPS-FTAZ-based devices were 18.83 mA/cm 2 and 17.17 mA/cm 2 , respectively. The Jph/ Jsat of PBDTNS-FTAZ and PBDTBPS-FTAZ-based devices were 97.5% and 94.5%, respectively, proving that the PBDTNS-FTAZ-based device exhibited a better exciton dissociation efficiency at the D/A interfaces and then was collected at the electrodes with little recombination. As shown in Figure  6, the power-law dependence of photocurrent on light intensity (P), Jsc vs P α was plotted. The α value for PBDTNS-FTAZ:ITIC and PBDTBPS-FTAZ:ITIC was 0.985 and 0.938, respectively. The α value approaching one demonstrating there was little bimolecular recombination in PBDTNS-FTAZ:ITIC devices, which could account for the higher Jsc and FF of the PBDTNS-FTAZ-based devices [60][61][62][63]. To further investigate the reason higher J sc of PBDTNS-FTAZ were obtained compared to PBDTBPS-FTAZ. The external quantum efficiency (EQE) was performed and the curves are displayed in Figure 4c. Both the PBDTNS-FTAZ/ITIC and PBDTBPS-FTAZ/ITIC devices showed broad photoresponse from 470 to 720 nm. The EQE response of the PBDTNS-FTAZ-based device were higher than those of PBDTBPS-FTAZ in the whole spectral range, which could partly account for the higher J sc of the PBDTNS-FTAZ-based device. The optimal devices possess a high and flat EQE value of over 65%. The calculated current density of 15.98 mA/cm 2 and 17.86 mA/cm 2 were obtained by integrating the EQE spectrum with the standard AM 1.5G solar spectrum, which is consistent with the measured J sc value with little error (<5%).
The relationship of the photocurrent (J ph ) and the effective applied voltage (V eff ) were studied to further analyze the charge recombination process in the devices based on PBDTNS-FTAZ and PBDTBPS-FTAZ. The J ph is defined by J L -J D , where J L and J D refer to the current density under AM 1.5G illumination and in the dark, respectively. V is the applied voltage; and V 0 refers to the effective voltage where J L = J D [55][56][57][58][59]. As displayed in Figure 5, the J ph of PBDTBPS-FTAZ-based devices could not reach the saturation current (J sat ) even the applied voltage is over 3V, indicating the relatively severe charge recombination in the PBDTBPS-FTAZ:ITIC blend. On the contrary, the J ph of the PBDTNS-FTAZ-based devices reached the saturation value (J sat ) at a low applied voltage of 1 V, proving that the photogenerated excitons in PBDTNS-FTAZ-based devices were fully dissociated to free charges. The J sat of the PBDTNS-FTAZ-based device was higher than the J sat of PBDTBPS-FTAZ-based device, demonstrating that photogenerated current is larger in PBDTNS-FTAZ devices. The charge recombination and exciton dissociation in the PBDTNS-FTAZ:ITIC and PBDTBPS-FTAZ:ITIC device can be further investigated by the J ph /J sat under short-circuit conditions. The J sat of PBDTNS-FTAZ and PBDTBPS-FTAZ-based devices were 18.83 mA/cm 2 and 17.17 mA/cm 2 , respectively. The J ph /J sat of PBDTNS-FTAZ and PBDTBPS-FTAZ-based devices were 97.5% and 94.5%, respectively, proving that the PBDTNS-FTAZ-based device exhibited a better exciton dissociation efficiency at the D/A interfaces and then was collected at the electrodes with little recombination. As shown in Figure 6, the power-law dependence of photocurrent on light intensity (P), J sc vs. P α was plotted. The α value for PBDTNS-FTAZ:ITIC and PBDTBPS-FTAZ:ITIC was 0.985 and 0.938, respectively. The α value approaching one demonstrating there was little bimolecular recombination in PBDTNS-FTAZ:ITIC devices, which could account for the higher J sc and FF of the PBDTNS-FTAZ-based devices [60][61][62][63].

Charge Transport Characteristics
The hole mobility of PBDTNS-FTAZ:ITIC and PBDTBPS-FTAZ:ITIC were characterized using the space-charge-limited-current (SCLC) method to study the charge transport ability. The J-V curves were shown in Figure S2. The hole only diodes were fabricated with the structure of ITO/PEDOT:PSS/active layer/Au. The hole mobility of PBDTNS-FTAZ and PBDTBPS-FTAZ are 6.64 × 10 −5 cm 2 /Vs and 2.28 × 10 −5 cm 2 /Vs, respectively. The PBDTNS-FTAZ exhibited higher hole mobility than PBDTBPS-FTAZ, which may result in a higher Jsc and FF in the PBDTNS-FTAZ-based devices.

Morphology Characterization
Transmission electron microscopy (TEM) and atomic force microscopy (AFM) measurements were performed to look into the surface morphology and the inside bulk structure of the active layer. As shown in Figure 7, the AFM and TEM images could tell that the PBDTNS-FTAZ:ITIC blend film possessed a smooth surface with a relatively low RMS of 1.17 and small aggregated microdomains. Indicating the PBDTNS-FTAZ polymer has excellent miscibility with ITIC. On the contrary, PBDTBPS-FTAZ:ITIC blend film exhibits a rough surface with a large RMS of 1.98 and a large scale of phase separation, indicating the aggregation of polymers and acceptors are severe in the blend film thus limiting the excitons separation and transportation. Thus, we can conclude that the compatibility of PBDTNS-FTAZ:ITIC is much better than that of PBDTBPS-FTAZ:ITIC, resulting in a much better morphology. A favorable morphology is vital for high-performance PSC, which could explain the higher Jsc and FF obtained in PBDTNS-FTAZ:ITIC device [64].

Charge Transport Characteristics
The hole mobility of PBDTNS-FTAZ:ITIC and PBDTBPS-FTAZ:ITIC were characterized using the space-charge-limited-current (SCLC) method to study the charge transport ability. The J-V curves were shown in Figure S2. The hole only diodes were fabricated with the structure of ITO/PEDOT:PSS/active layer/Au. The hole mobility of PBDTNS-FTAZ and PBDTBPS-FTAZ are 6.64 × 10 −5 cm 2 /Vs and 2.28 × 10 −5 cm 2 /Vs, respectively. The PBDTNS-FTAZ exhibited higher hole mobility than PBDTBPS-FTAZ, which may result in a higher Jsc and FF in the PBDTNS-FTAZ-based devices.

Morphology Characterization
Transmission electron microscopy (TEM) and atomic force microscopy (AFM) measurements were performed to look into the surface morphology and the inside bulk structure of the active layer. As shown in Figure 7, the AFM and TEM images could tell that the PBDTNS-FTAZ:ITIC blend film possessed a smooth surface with a relatively low RMS of 1.17 and small aggregated microdomains. Indicating the PBDTNS-FTAZ polymer has excellent miscibility with ITIC. On the contrary, PBDTBPS-FTAZ:ITIC blend film exhibits a rough surface with a large RMS of 1.98 and a large scale of phase separation, indicating the aggregation of polymers and acceptors are severe in the blend film thus limiting the excitons separation and transportation. Thus, we can conclude that the compatibility of PBDTNS-FTAZ:ITIC is much better than that of PBDTBPS-FTAZ:ITIC, resulting in a much better morphology. A favorable morphology is vital for high-performance PSC, which could explain the higher Jsc and FF obtained in PBDTNS-FTAZ:ITIC device [64].

Charge Transport Characteristics
The hole mobility of PBDTNS-FTAZ:ITIC and PBDTBPS-FTAZ:ITIC were characterized using the space-charge-limited-current (SCLC) method to study the charge transport ability. The J-V curves were shown in Figure S2. The hole only diodes were fabricated with the structure of ITO/PEDOT:PSS/active layer/Au. The hole mobility of PBDTNS-FTAZ and PBDTBPS-FTAZ are 6.64 × 10 −5 cm 2 /Vs and 2.28 × 10 −5 cm 2 /Vs, respectively. The PBDTNS-FTAZ exhibited higher hole mobility than PBDTBPS-FTAZ, which may result in a higher J sc and FF in the PBDTNS-FTAZ-based devices.

Morphology Characterization
Transmission electron microscopy (TEM) and atomic force microscopy (AFM) measurements were performed to look into the surface morphology and the inside bulk structure of the active layer. As shown in Figure 7, the AFM and TEM images could tell that the PBDTNS-FTAZ:ITIC blend film possessed a smooth surface with a relatively low RMS of 1.17 and small aggregated microdomains. Indicating the PBDTNS-FTAZ polymer has excellent miscibility with ITIC. On the contrary, PBDTBPS-FTAZ:ITIC blend film exhibits a rough surface with a large RMS of 1.98 and a large scale of phase separation, indicating the aggregation of polymers and acceptors are severe in the blend film thus limiting the excitons separation and transportation. Thus, we can conclude that the compatibility of PBDTNS-FTAZ:ITIC is much better than that of PBDTBPS-FTAZ:ITIC, resulting in a much better morphology. A favorable morphology is vital for high-performance PSC, which could explain the higher J sc and FF obtained in PBDTNS-FTAZ:ITIC device [64].

Conclusions
In summary, two novel 2D-BDT-based polymers containing alkylthionaphthyl (PBDTNS-FTAZ) and alkylthiobiphenyl (PBDTBPS-FTAZ) as side-chains were synthesized and photovoltaic properties in nonfullerene systems were compared in detail. The PBDTNS-FTAZ exhibited almost the same absorption property as PBDTBPS-FTAZ. However, the PBDTNS-FTAZ shows the better miscibility with ITIC compared with PBDTBPS-FTAZ. Therefore, the PBDTNS-FTAZ exhibited better EQE response, fill factor and yielded a higher PCE of 9.64%。This work proved that extending the side-chain on the BDT unit with fused aromatic rings providing better planarity could facilitate the stacking of polymer, avoiding excessive phase separation and improve PCE when blended with ITIC.

Conclusions
In summary, two novel 2D-BDT-based polymers containing alkylthionaphthyl (PBDTNS-FTAZ) and alkylthiobiphenyl (PBDTBPS-FTAZ) as side-chains were synthesized and photovoltaic properties in nonfullerene systems were compared in detail. The PBDTNS-FTAZ exhibited almost the same absorption property as PBDTBPS-FTAZ. However, the PBDTNS-FTAZ shows the better miscibility with ITIC compared with PBDTBPS-FTAZ. Therefore, the PBDTNS-FTAZ exhibited better EQE response, fill factor and yielded a higher PCE of 9.64%. This work proved that extending the side-chain on the BDT unit with fused aromatic rings providing better planarity could facilitate the stacking of polymer, avoiding excessive phase separation and improve PCE when blended with ITIC.

Conflicts of Interest:
The authors declare no conflict of interest.