Effect of Flank Rotation on the Photovoltaic Properties of Dithieno[2,3-d:2′,3′-d′]benzo[1,2-b:4,5-b′]dithiophene-Based Narrow Band Gap Copolymers

Side chain engineering has been an effective approach to modulate the solution processability, optoelectronic properties and miscibility of conjugated polymers (CPs) for organic/polymeric photovoltaic cells (PVCs). As compared with the most commonly used method of introducing alkyl chains, the employment of alkyl-substituted aryl flanks would provide two-dimensional (2-D) CPs having solution processability alongside additional merits like deepened highest occupied molecular orbital (HOMO) energy levels, increased absorption coefficient and charger transporting, etc. In this paper, the triple C≡C bond was used as conjugated linker to decrease the steric hindrance between the flanks of 4,5-didecylthien-2-yl (T) and dithieno[2,3-d:2′,3′-d′]benzo[1,2-b:4,5-b′]dithiophene (DTBDT) core. In addition, an alternating CP derived from 4,5-didecylthien-2-yl-ethynyl (TE) flanked DTBDT, and 4,9-bis(4-octylthien-2-yl) naphtho[1,2-c:5,6-c′]bis[1,2,5]thiadiazole (DTNT), named as PDTBDT-TE-DTNT, was synthesized and characterized. As compared with the controlled PDTBDT-T-DTNT, which was derived from 4,5-didecylthien-2-yl flanked DTBDT and DTNT, the results for exciton dissociation probability, density functional theory (DFT), time-resolved photoluminescence (PL) measurements, etc., revealed that the lower steric hindrance between TE and DTBDT might lead to the easier rotation of the TE flanks, thus contributing to the decrease of the exciton lifetime and dissociation probability, finally suppressing the short-circuit current density (JSC), etc., of the photovoltaic devices from PDTBDT-TE-DTNT.


Introduction
Polymeric photovoltaic cells (PVCs) have attracted much attention due to advantages such as the printable fabrication process, and their ultrathin, lightweight and flexible properties.

Synthesis of the Monomers and Copolymer
2.1.1. 2-iodo-4,5-didecylthiophene (T-I) 4,5-Didecylthiophene (2.00 g, 5.49 mmol) was dissolved in chloroform (35 mL) and HOAc (7 mL) under argon at 0 • C, followed by slow addition of N-iodosuccinimide (NIS) (1.73 g, 7.68 mmol) under dark. The mixture was stirred for 20 min and allowed to warm to room temperature (r.t.). After stirring for 4 h, dilute aqueous Na 2 S 2 O 3 (20 mL) was added to the solution then the solution was extracted with chloroform. The combined organic layer was dried over anhydrous Na 2 SO 4 and concentrated. Purification by silica gel chromatography using n-hexane gave white oil. The product was obtained to 2.30 g (yield: 85.2%). 1  T-I (1.00 g, 2.04 mmol), palladium chloride (28.95 mg, 0.16 mmol), cuprous iodide (24.83 mg, 0.13 mmol), triphenylphosphine (0.21 g, 0.80 mmol) were diluted with triethylamine (5 mL) and THF (5 mL) under nitrogen, 2-trimethylsilyl-acetylene (0.28 g, 2.86 mmol) was slowly added to the mixture. After all of the solution had been added, the mixture was heated at 65 • C with stirring for 1 h. It was then poured into water (300 mL) and extracted with ethyl acetate, the organic phase was washed with brine (200 mL), dried with anhydrous Na 2 SO 4 and concentrated. The crude product was purified by column chromatography (hexane eluent) to obtain 0.73 g of clear oil (yield: 75.0%). 1 [1,2-b:4,5-b ]dithiophene-5,10-diketone (0.50 g, 1.50 mmol) was quickly added and the mixture was stirred at 50 • C for 1 h. Then a solution of SnCl 2 ·2H 2 O (2.70 g, 12.00 mmol) in 10% HCl (30 mL) was added and stirred for additional 1.5 h and then poured into the ice water. The mixture was extracted twice with petroleum ether (PE). The organic phase was dried over Na 2 SO 4 and concentrated to afford crude product, which was purified by silica gel chromatography using PE as the eluent to yield a yellow solid (yield: 1.46 g, 91.3%). 1  Under nitrogen at −35 • C, n-BuLi (1.6 M in hexane, 0.86 mL, 1.38 mmol) was added dropwise to the 60 mL of anhydrous THF solution of DTBDT-TE (0.50 g, 0.46 mmol), the reaction mixture was stirred for 2 h. Then, at −35 • C, hexane solution of Me 3 SnCl (0.29 g, 1.47 mmol) was added in one portion. The reaction mixture was stirred at −35 • C for 0.5 h and then warmed to r.t. for 6 h. Subsequently, the reaction mixture was poured into water and extracted twice with PE. Then, the organic layer was dried over Na 2 SO 4 and concentrated to obtain the yellow crude product, which was recrystallized from iso-propanol (i-PrOH), finally obtaining pure DTBDT-TE as a yellow sheet-shaped crystal (yield: 0.55 g, 85.0%). 1 Figure S1). 13 Figure S2). DTBDT-TESn (182 mg, 0.13 mmol), DTNTC 8 Br (103 mg, 0.13 mmol), 10 mL of toluene and 2 mL of N,N-dimethylformide (DMF) were added to a 25 mL two-neck bottle. After being flushed with argon for 20 min, the catalyst Pd 2 (dba) 3 (1.5 mg) and P(o-tol) 3 (3 mg) were added, and the mixture was then purged with argon for 10 min. The solution was stirred and heated to reflux for 48 h under argon atmosphere. At the end of polymerization, the polymer was end-capped with 2-tributylstannylthiophene and 2-bromothiophene to remove bromo and trimethylstannyl end groups. The reaction was cooled to r.t., and the mixture was precipitated in methanol and filtered. Further purification was carried out by Soxhlet extraction using the sequence ethanol, acetone, hexane and toluene as the eluents to remove the residue catalyst and oligomers. Following that, the concentrated solutions of the copolymers in toluene were poured into methanol again (300 mL). The precipitation was collected and dried under vacuum overnight. Yield: 81%. M n = 21,628 g/mol with PDI of 3.14.

Absorption Spectra of the Copolymer
The UV-Vis absorption spectra of the copolymer in solution and solid thin film were monitored on a UV-1800 spectrophotometer, and the absorption spectra and corresponding parameters are presented in Figure 1 and Table 1. The PDTBDT-TE-DTNT exhibited three absorption peaks at 423 nm, 545 nm and 643 nm, with shoulder absorption peaks at 699 nm. The absorption peaks at around 423 nm were attributed to the π−π * transition of the polymer backbone, and the absorption peaks at around 643 nm (0-1) and 699 nm (0-0) were respectively attributed to the intermolecular charge transfer (ICT) of the polymer [32,33]. Continuing on from solution to solid state, the absorption peak of PDTBDT-TE-DTNT at around 423 nm decreased, the absorption peak at round 643 nm clearly increased, and the on-set band gap wavelength varied from 789 nm to 792 nm. Additionally, the absorption spectra of the controlled polymer, e.g., PDTBDT-T-DTNT, were also monitored for comparison ( Figure S4). The PDTBDT-T-DTNT exhibited three main absorption peaks at 347 nm, 484 nm and 673 nm, with a shoulder absorption peak at around 726 nm in dilute solution. The absorption spectrum of PDTBDT-T-DTNT was almost unchanged, except that the relative absorption intensity at 347 nm was obviously decreased, and the on-set band gap wavelength of the PDTBDT-T-DTNT was red-shifted from 809 nm to 822 nm, when moving from solution to solid state. The corresponding optical band gap values ( ) of PDTBDT-TE-DTNT and PDTBDT-T-DTNT were 1.56 and 1.51 eV, respectively, according to the formula = 1240/λ .
It was found that PDTBDT-TE-DTNT presented a slightly broader optical band gap in contrast to that of PDTBDT-T-DTNT. Scheme 1. Synthesis routes of the monomers and copolymers.

Absorption Spectra of the Copolymer
The UV-Vis absorption spectra of the copolymer in solution and solid thin film were monitored on a UV-1800 spectrophotometer, and the absorption spectra and corresponding parameters are presented in Figure 1 and Table 1. The PDTBDT-TE-DTNT exhibited three absorption peaks at 423 nm, 545 nm and 643 nm, with shoulder absorption peaks at 699 nm. The absorption peaks at around 423 nm were attributed to the π−π * transition of the polymer backbone, and the absorption peaks at around 643 nm (0-1) and 699 nm (0-0) were respectively attributed to the intermolecular charge transfer (ICT) of the polymer [32,33]. Continuing on from solution to solid state, the absorption peak of PDTBDT-TE-DTNT at around 423 nm decreased, the absorption peak at round 643 nm clearly increased, and the on-set band gap wavelength varied from 789 nm to 792 nm. Additionally, the absorption spectra of the controlled polymer, e.g., PDTBDT-T-DTNT, were also monitored for comparison ( Figure  S4). The PDTBDT-T-DTNT exhibited three main absorption peaks at 347 nm, 484 nm and 673 nm, with a shoulder absorption peak at around 726 nm in dilute solution. The absorption spectrum of PDTBDT-T-DTNT was almost unchanged, except that the relative absorption intensity at 347 nm was obviously decreased, and the on-set band gap wavelength of the PDTBDT-T-DTNT was red-shifted from 809 nm to 822 nm, when moving from solution to solid state. The corresponding optical band gap values (E opt g ) of PDTBDT-TE-DTNT and PDTBDT-T-DTNT were 1.56 and 1.51 eV, respectively, according to the formula E opt g = 1240/λ film onset . It was found that PDTBDT-TE-DTNT presented a slightly broader optical band gap in contrast to that of PDTBDT-T-DTNT.

Aggregation of the Copolymers in Solution and Solid State
To investigate the aggregation characteristics of the CPs in solution, temperature-dependent absorption spectra (TD-Abs) of the copolymers in o-dichlorobenzene (o-DCB) solution were monitored ( Figure 2). Obvious changing of the absorption spectra of PDTBDT-TE-DTNT and PDTBDT-T-DTNT in o-DCB solution was found during the heating process. From 25 to 105 °C, the absorbance of the λ0-0 peaks at 699 nm and the λ0-1 peaks at 643 nm for PDTBDT-TE-DTNT gradually decreased and blue-shifted by 20 and 32 nm, respectively, indicating that the degree of torsion of the unit on the polymer conjugated backbone was aggravated, the coplanarity of the molecule was destroyed, and the effective conjugate length and the degree of conjugation were reduced, thus causing a blue shift in the absorption spectrum [33][34][35][36]. For the PDTBDT-T-DTNT, the absorbance of the λ0-0 peaks at 726 nm and the λ0-1 peaks at 673 nm was continuously reduced and blue-shifted to 689 nm and 580 nm. In addition, the relative intension rations of the λ0-1/λ0-0 peaks varied between 0.96:1 and 1.23:1 for PDTBDT-TE-DTNT and 1.05:1 and 1.97:1 for PDTBDT-T-DTNT. It was clear that the blue shift value of λ0-1 peak and the decrease of λ0-0 peak for PDTBDT-TE-DTNT were significantly smaller than those for PDTBDT-T-DTNT, and the results indicated that PDTBDT-TE-DTNT exhibited stronger aggregation than PDTBDT-T-DTNT in dilute solution [28,[33][34][35][36].

Aggregation of the Copolymers in Solution and Solid State
To investigate the aggregation characteristics of the CPs in solution, temperature-dependent absorption spectra (TD-Abs) of the copolymers in o-dichlorobenzene (o-DCB) solution were monitored ( Figure 2). Obvious changing of the absorption spectra of PDTBDT-TE-DTNT and PDTBDT-T-DTNT in o-DCB solution was found during the heating process. From 25 to 105 • C, the absorbance of the λ 0-0 peaks at 699 nm and the λ 0-1 peaks at 643 nm for PDTBDT-TE-DTNT gradually decreased and blue-shifted by 20 and 32 nm, respectively, indicating that the degree of torsion of the unit on the polymer conjugated backbone was aggravated, the coplanarity of the molecule was destroyed, and the effective conjugate length and the degree of conjugation were reduced, thus causing a blue shift in the absorption spectrum [33][34][35][36]. For the PDTBDT-T-DTNT, the absorbance of the λ 0-0 peaks at 726 nm and the λ 0-1 peaks at 673 nm was continuously reduced and blue-shifted to 689 nm and 580 nm. In addition, the relative intension rations of the λ 0-1 /λ 0-0 peaks varied between 0.96:1 and 1.23:1 for PDTBDT-TE-DTNT and 1.05:1 and 1.97:1 for PDTBDT-T-DTNT. It was clear that the blue shift value of λ 0-1 peak and the decrease of λ 0-0 peak for PDTBDT-TE-DTNT were significantly smaller than those for PDTBDT-T-DTNT, and the results indicated that PDTBDT-TE-DTNT exhibited stronger aggregation than PDTBDT-T-DTNT in dilute solution [28,[33][34][35][36].
To gain insight into the aggregation characteristics of CPs in solid state, X-ray diffraction (XRD) analyses of the polymer films cast from CB solution onto glass substrate were measured. As shown in Figure 3, The diffraction peaks of PDTBDT-TE-DTNT were located at 2θ = 22.31 • and 2θ = 3.31 • , corresponding to the π-π stacking d-spacing of 3.98 Å and interlayer d-spacing of 26.66 Å based on Bragg's law (i.e., λ = 2dsinθ) [28]. The controlled polymer PDTBDT-T-DTNT films provided two peaks at about 2θ = 3.27 • and 2θ = 23.01 • , corresponding to an interlayer d-spacing of 26.99 Å and a d-spacing To gain insight into the aggregation characteristics of CPs in solid state, X-ray diffraction (XRD) analyses of the polymer films cast from CB solution onto glass substrate were measured. As shown in Figure 3, The diffraction peaks of PDTBDT-TE-DTNT were located at 2θ = 22.31° and 2θ = 3.31°, corresponding to the π-π stacking d-spacing of 3.98 Å and interlayer d-spacing of 26.66 Å based on Bragg's law (i.e., λ = 2dsinθ) [28]. The controlled polymer PDTBDT-T-DTNT films provided two peaks at about 2θ = 3.27° and 2θ = 23.01°, corresponding to an interlayer d-spacing of 26.99 Å and a d-spacing of 3.86 Å, respectively. It was shown that PDTBDT-TE-DTNT possessed a shorter interlayer d-spacing distance and a longer π-π stacking distance than PDTBDT-T-DTNT.

Electrochemical Characteristic of the Copolymers
The energy levels are important parameters to guide the selection of appropriate acceptors in PVCs. Therefore, cyclic voltammetry (CV) was employed to measure the redox behaviors of the copolymers and determine their energy levels [37,38]. The onset oxidation potential was determined by the CV curves and calibrated with the potential of ferrocene/ferrocenium (Fc/Fc + ), assuming the energy level of ferrocene/ferrocenium (Fc/Fc + ) to be −4.80 eV below the vacuum level [39]. The redox potential of Fc/Fc + under the above-mentioned conditions was +0.11 V. For the CV of ferrocene and the two polymer films on a glassy carbon electrode in 0.1 mol L −1 tetrabutylammonium hexafluorophosphate (Bu4NPF6) acetonitrile solution at a scan rate of 50 mV s −1 , the HOMO levels and LUMO levels were calculated by empirical formulas (EHOMO = -(Eox + 4.69) (eV) [39] and ELUMO = -(│EHOMO│-(eV). As shown in Figure 4a, the onset oxidation potential of PDTBDT-TE-DTNT was located at around 0.68 V, corresponding to the HOMO level of −5.23 eV, and the LUMO energy level was about -3.67 eV. Meanwhile, the HOMO and LUMO levels of PDTBDT-T-DTNT were about −5.32 eV and −3.81 eV, respectively. The elevated HOMO energy  To gain insight into the aggregation characteristics of CPs in solid state, X-ray diffraction (XRD) analyses of the polymer films cast from CB solution onto glass substrate were measured. As shown in Figure 3, The diffraction peaks of PDTBDT-TE-DTNT were located at 2θ = 22.31° and 2θ = 3.31°, corresponding to the π-π stacking d-spacing of 3.98 Å and interlayer d-spacing of 26.66 Å based on Bragg's law (i.e., λ = 2dsinθ) [28]. The controlled polymer PDTBDT-T-DTNT films provided two peaks at about 2θ = 3.27° and 2θ = 23.01°, corresponding to an interlayer d-spacing of 26.99 Å and a d-spacing of 3.86 Å, respectively. It was shown that PDTBDT-TE-DTNT possessed a shorter interlayer d-spacing distance and a longer π-π stacking distance than PDTBDT-T-DTNT.

Electrochemical Characteristic of the Copolymers
The energy levels are important parameters to guide the selection of appropriate acceptors in PVCs. Therefore, cyclic voltammetry (CV) was employed to measure the redox behaviors of the copolymers and determine their energy levels [37,38]. The onset oxidation potential was determined by the CV curves and calibrated with the potential of ferrocene/ferrocenium (Fc/Fc + ), assuming the energy level of ferrocene/ferrocenium (Fc/Fc + ) to be −4.80 eV below the vacuum level [39]. The redox potential of Fc/Fc + under the above-mentioned conditions was +0.11 V. For the CV of ferrocene and the two polymer films on a glassy carbon electrode in 0.1 mol L −1 tetrabutylammonium hexafluorophosphate (Bu4NPF6) acetonitrile solution at a scan rate of 50 mV s −1 , the HOMO levels and LUMO levels were calculated by empirical formulas (EHOMO = -(Eox + 4.69) (eV) [39] and ELUMO = -(│EHOMO│-(eV). As shown in Figure 4a, the onset oxidation potential of PDTBDT-TE-DTNT was located at around 0.68 V, corresponding to the HOMO level of −5.23 eV, and the LUMO energy level was about -3.67 eV. Meanwhile, the HOMO and LUMO levels of PDTBDT-T-DTNT were about −5.32 eV and −3.81 eV, respectively. The elevated HOMO energy

Electrochemical Characteristic of the Copolymers
The energy levels are important parameters to guide the selection of appropriate acceptors in PVCs. Therefore, cyclic voltammetry (CV) was employed to measure the redox behaviors of the copolymers and determine their energy levels [37,38]. The onset oxidation potential was determined by the CV curves and calibrated with the potential of ferrocene/ferrocenium (Fc/Fc + ), assuming the energy level of ferrocene/ferrocenium (Fc/Fc + ) to be −4.80 eV below the vacuum level [39]. The redox potential of Fc/Fc + under the above-mentioned conditions was +0.11 V. For the CV of ferrocene and the two polymer films on a glassy carbon electrode in 0.1 mol L −1 tetrabutylammonium hexafluorophosphate (Bu 4 NPF 6 ) acetonitrile solution at a scan rate of 50 mV s −1 , the HOMO levels and LUMO levels were calculated by empirical formulas (E HOMO = −(Eox + 4.69) (eV) [39] and E LUMO = −(|E HOMO |− E g ) (eV). As shown in Figure 4a, the onset oxidation potential of PDTBDT-TE-DTNT was located at around 0.68 V, corresponding to the HOMO level of −5.23 eV, and the LUMO energy level was about −3.67 eV. Meanwhile, the HOMO and LUMO levels of PDTBDT-T-DTNT were about −5.32 eV and −3.81 eV, respectively. The elevated HOMO energy level of PDTBDT-TE-DTNT would result in a decrease of the V OC of the PDTBDT-TE-DTNT-based PVCs in comparison to that for the PDTBDT-T-DTNT-based PVCs [40]. In spite that, the LUMO gaps between the copolymers and PC 71 BM were 0.26-0.39 eV, which would provide sufficient driving force to promote efficient exciton dissociation at the D-A interface, thereby ensuring energetically favorable electron transfer between the CPs and the PC 71 BM [41]. level of PDTBDT-TE-DTNT would result in a decrease of the VOC of the PDTBDT-TE-DTNT-based PVCs in comparison to that for the PDTBDT-T-DTNT-based PVCs [40]. In spite that, the LUMO gaps between the copolymers and PC71BM were 0.26-0.39 eV, which would provide sufficient driving force to promote efficient exciton dissociation at the D-A interface, thereby ensuring energetically favorable electron transfer between the CPs and the PC71BM [41].

Hole Mobilities of the Blend Films from the Copolymers
The carrier charge mobility is an important parameter, and could affect charge carrier transport and recombination in photovoltaic devices. Higher mobility is beneficial to increasing the current density and decreasing unfavorable exciton recombination [42,43]. The space-charge-limited current (SCLC) method was used to determine the hole mobilities of the blend films from the polymer in the devices with configuration of ITO/PEDOT:PSS/polymer: PC71BM/MoO3/Ag. The mobilities of the blend films were calculated by the SCLC model, which is described by Equation (1) where J is current density, ɛ0 stands for the permittivity of free space, ɛr is the relative dielectric constant of the transport medium, which is assumed to be around 3 for the CPs, µ is the hole mobility, V is the internal potential in the devices and L is the thickness of the active layers. The J-V characteristics of the devices from the PDTBDT-TE-DTNT/PC71BM were presented in Figure 5, and the J-V curve of the devices from PDTBDT-T-DTNT/PC71BM is also provided for comparison. The hole mobilities for PDTBDT-TE-DTNT/PC71BM was about 1.85 × 10 −4 cm 2 ·V −1 ·s −1 , which is comparable to the value of 1.55 × 10 −4 cm 2 ·V −1 ·s −1 obtained for the PDTBDT-T-DTNT/PC71BM blend films. In addition, the results indicated that the replacement of T with TE did not influenced the hole mobilities of the copolymers.

Hole Mobilities of the Blend Films from the Copolymers
The carrier charge mobility is an important parameter, and could affect charge carrier transport and recombination in photovoltaic devices. Higher mobility is beneficial to increasing the current density and decreasing unfavorable exciton recombination [42,43]. The space-charge-limited current (SCLC) method was used to determine the hole mobilities of the blend films from the polymer in the devices with configuration of ITO/PEDOT:PSS/polymer: PC 71 BM/MoO 3 /Ag. The mobilities of the blend films were calculated by the SCLC model, which is described by Equation (1) where J is current density, ε 0 stands for the permittivity of free space, ε r is the relative dielectric constant of the transport medium, which is assumed to be around 3 for the CPs, µ is the hole mobility, V is the internal potential in the devices and L is the thickness of the active layers. The J-V characteristics of the devices from the PDTBDT-TE-DTNT/PC 71 BM were presented in Figure 5, and the J-V curve of the devices from PDTBDT-T-DTNT/PC 71 BM is also provided for comparison. The hole mobilities for PDTBDT-TE-DTNT/PC 71 BM was about 1.85 × 10 −4 cm 2 ·V −1 ·s −1 , which is comparable to the value of 1.55 × 10 −4 cm 2 ·V −1 ·s −1 obtained for the PDTBDT-T-DTNT/PC 71 BM blend films. In addition, the results indicated that the replacement of T with TE did not influenced the hole mobilities of the copolymers.
Polymers 2019, 11, 239 9 of 17 level of PDTBDT-TE-DTNT would result in a decrease of the VOC of the PDTBDT-TE-DTNT-based PVCs in comparison to that for the PDTBDT-T-DTNT-based PVCs [40]. In spite that, the LUMO gaps between the copolymers and PC71BM were 0.26-0.39 eV, which would provide sufficient driving force to promote efficient exciton dissociation at the D-A interface, thereby ensuring energetically favorable electron transfer between the CPs and the PC71BM [41].

Hole Mobilities of the Blend Films from the Copolymers
The carrier charge mobility is an important parameter, and could affect charge carrier transport and recombination in photovoltaic devices. Higher mobility is beneficial to increasing the current density and decreasing unfavorable exciton recombination [42,43]. The space-charge-limited current (SCLC) method was used to determine the hole mobilities of the blend films from the polymer in the devices with configuration of ITO/PEDOT:PSS/polymer: PC71BM/MoO3/Ag. The mobilities of the blend films were calculated by the SCLC model, which is described by Equation (1) [44] where J is current density, ɛ0 stands for the permittivity of free space, ɛr is the relative dielectric constant of the transport medium, which is assumed to be around 3 for the CPs, µ is the hole mobility, V is the internal potential in the devices and L is the thickness of the active layers. The J-V characteristics of the devices from the PDTBDT-TE-DTNT/PC71BM were presented in Figure 5, and the J-V curve of the devices from PDTBDT-T-DTNT/PC71BM is also provided for comparison. The hole mobilities for PDTBDT-TE-DTNT/PC71BM was about 1.85 × 10 −4 cm 2 ·V −1 ·s −1 , which is comparable to the value of 1.55 × 10 −4 cm 2 ·V −1 ·s −1 obtained for the PDTBDT-T-DTNT/PC71BM blend films. In addition, the results indicated that the replacement of T with TE did not influenced the hole mobilities of the copolymers.

Photovoltaic Characteristics and Optical Modeling of the Device from the Copolymers
To explore the influence of the flanks on the photovoltaic properties of the copolymers of the PDTBDT-TE-DTNT and PDTBDT-T-DTNT, inverted photovoltaic devices (i-PVCs) with a configuration of ITO/PFN/active layer/MoO 3 /Ag, in which PDTBDT-TE-DTNT and/or PDTBDT-T-DTNT was used as electron donor material, and PC 71 BM was used as electron acceptor material, were fabricated following the reported procedures [25]. The current density vs voltage (J-V) characteristics of i-PVCs are shown in Figure 6a, as measured under illumination of AM 1.5G at 100 mW/cm 2 conditions, and the corresponding photovoltaic performances are listed in Table 2. The optimal weight ratios of the PDTBDT-TE-DTNT and PC 71 BM were about 1:1, while the weight ratios of the PDTBDT-TE-DTNT and PC 61 BM were varied from 1:1, 1:2 and 1:3, 3% DIO was employed as solvent additives ( Figure S5 and Table S1 characteristics of i-PVCs are shown in Figure 6a, as measured under illumination of AM 1.5G at 100 mW/cm 2 conditions, and the corresponding photovoltaic performances are listed in Table 2. The optimal weight ratios of the PDTBDT-TE-DTNT and PC71BM were about 1:1, while the weight ratios of the PDTBDT-TE-DTNT and PC61BM were varied from 1:1, 1:2 and 1:3, 3% DIO was employed as solvent additives ( Figure S5 and Table S1). The PCEs of the optimal i-PVCs from PDTBDT-TE-DTNT/PC71BM (W:W, 1:1) with 3% DIO as solvent additives, were about 3.97%, alongside with a VOC of 0.60 V, a JSC of 10.15 mA cm -2 and a FF of 65.16%. Meanwhile, i-PVCs from controlled CPs of PDTBDT-T-DTNT/PC71BM provided the maximal PCE of 7.57% with a VOC of 0.70 V, a JSC of 16.09 mA/cm 2 and a FF of 67.19%. It could be found that the VOC and JSC of i-PVCs from PDTBDT-TE-DTNT/PC71BM blend films were lower than those for i-PVCs from the PDTBDT-T-DTNT/PC71BM. The lower VOC of PDTBDT-TE-DTNT-based devices might be attributed to the enhanced HOMO energy level, and the broader band gap of PDTBDT-TE-DTNT might lead to the decreasing JSC for PDTBDT-TE-DTNT-based devices in contrast to that for PDTBDT-T-DTNT. However, it is noted that the estimated JSC of the devices from the copolymers were calculated under the assumption that the devices exhibited the same internal quantum efficiencies, charge transporting and collection characteristics, etc., except for the difference in light harvesting between them ( Figure S6), which were about 23.    Recently, it has been well demonstrated that the topography and morphology of the active layer (such as domain size) are critical to exciton separation, and charge carrier recombination and transport, which also play an important role in J SC . Tapping-mode atomic force microscopy (AFM) and transmission electron microscopy (TEM) were implemented to gain insight into the nature of topography and morphology of the blend films from the copolymers. As demonstrated by the AFM height images in Figure S7, the root-mean-square surface roughness (RMS) values were about 1.64 nm for PDTBDT-T-DTNT/PC 71 BM and 2.66 nm for PDTBDT-TE-DTNT/PC 71 BM blend films. The TEM images of the corresponding films indicated that that the PDTBDT-T-DTNT/PC 71 BM exhibited similar domain sizes to the PC 71 BM (dark regions) and the polymers matrix (bight regions) with PDTBDT-TE-DTNT/PC 71 BM blend films ( Figure S8) [45][46][47][48]. It is worthwhile pointing out that the results of the study on the relationship between RMS of the blend films and J SC for the PVCs are, thus far, contradictory [49,50]. However, we speculated that the lower J SC of the PDTBDT-TE-DTNT-based i-PVCs cannot be ascribed to the slight difference in RMS between the PDTBDT-T-DTNT/PC 71 BM and PDTBDT-TE-DTNT, because the blend films with the PDTBDT-TE-DTNT and PDTBDT-T-DTNT exhibited similar morphologies [45][46][47][48][49][50].
To gain insight into the origin of the large decrease in J SC of i-PVCs from PDTBDT-TE-DTNT, we also monitored the incident photon-to-electron conversion efficiencies (IPCEs) of i-PVCs from the copolymers (Figure 6b). The PDTBDT-TE-DTNT/PC 71 BM-based devices presented light response ranging from 300 nm to 800 nm with IPCEs of 1.2% to 48.8%. As compared with the PDTBDT-TE-DTNT-based devices, the PDTBDT-T-DTNT/PC 71 BM-based devices presented similar light response except that they exhibited higher IPCEs in the whole photocurrent response of 300 nm to 820 nm. In addition, the lower J SC of i-PVCs from PDTBDT-TE-DTNT might have led to the lower IPCEs of the devices in contrast with those for the devices from PDTBDT-T-DTNT. Following that, the characteristics of the photocurrent density (J ph ) versus the effective applied voltage (V eff ) of the i-PVCs were also measured [51,52]. In addition, the J ph is defined as per Equation (2) where J L and J D are the photocurrent densities under illumination and dark, respectively. V eff can be defined as the difference between V a and V 0 (3) where V 0 is the voltage at which the photocurrent is zero and V a is the applied external voltage bias. As shown in Figure 7, the J ph of the both the devices from PDTBDT-TE-DTNT and PDTBDT-T-DTNT increased sharply when V eff was low, and then saturated when the V eff increased, suggesting that most of the excitons had been dissociated with the help of a large reverse bias. Nevertheless, it could be found that the J ph of the PDTBDT-T-DTNT-based devices were always higher than those for PDTBDT-TE-DTNT. In addition, the ratio of J ph to J sat under short-circuit condition was defined as P diss to present the exciton dissociation probability, and in which J sat was the saturated current density. P diss of about 91% and 85% were obtained for the optimal i-PVCs from PDTBDT-T-DTNT and/or PDTBDT-TE-DTNT, respectively. The results indicated that the PDTBDT-TE-DTNT-based i-PVCs exhibited low P diss , which would result in lower IPCEs for the PDTBDT-TE-DTNT-based i-PVCs.

Time-Resolved Photoluminescence of the Copolymers
Generally, polymer singlet excitons are the precursors for charger photogeneration taking place at the D-A interface, and the lifetime of the singlet excitons in CPs would determine their diffusion length in polymer films, which would exhibit a direct impact on the exciton dissociation and

Time-Resolved Photoluminescence of the Copolymers
Generally, polymer singlet excitons are the precursors for charger photogeneration taking place at the D-A interface, and the lifetime of the singlet excitons in CPs would determine their diffusion length in polymer films, which would exhibit a direct impact on the exciton dissociation and photocurrent generation characteristics of the photovoltaic devices [53,54]. The time-resolved photoluminescence (PL) of the PDTBDT-TE-DTNT and PDTBDT-T-DTNT was performed by time-correlated single photon counting (TCSPC) on the dilute solution of toluene (Figure 8). The τ 0 of 2.61 ns for PDTBDT-T-DTNT, and 1.92 ns for PDTBDT-TE-DTNT, was extracted from the time-resolved photoluminescence decay for the copolymer via single exponential fit. However, it could be found that the single exponential fit failed to accurately describe the PL decay of PDTBDT-TE-DTNT, and the deviations of PDTBDT-T-DTNT PL decay from the single exponential fit were very small. Double exponential fits were also implemented to understand the PL decay of the copolymers. The τ 1 and τ 2 of 0.347 and 2.741 ns were extracted from the PL decay of PDTBDT-T-DTNT, and the τ 1 and τ 2 of 0.486 ns and 2.17 ns were extracted from the PL decay of PDTBDT-TE-DTNT. As compared with the single exponential fit, the double exponential fits could more accurately describe the PL decay for both copolymers. This indicates that there were at least two types of excited state decay for the excitations of the copolymers, and the dominant excited states, which were mainly related with the charger photo-generation process of the photovoltaic devices, provided a lifetime of 2.741 ns for PDTBDT-T-DTNT and 2.171 ns for PDTBDT-TE-DTNT. In a word, the PDTBDT-TE-DTNT had a shorter PL lifetime, and the lower exciton dissociation probability of the devices might be attributed to the shorter PL lifetime when compared to that of PDTBDT-T-DTNT.

Time-Resolved Photoluminescence of the Copolymers
Generally, polymer singlet excitons are the precursors for charger photogeneration taking place at the D-A interface, and the lifetime of the singlet excitons in CPs would determine their diffusion length in polymer films, which would exhibit a direct impact on the exciton dissociation and photocurrent generation characteristics of the photovoltaic devices [53,54]. The time-resolved photoluminescence (PL) of the PDTBDT-TE-DTNT and PDTBDT-T-DTNT was performed by time-correlated single photon counting (TCSPC) on the dilute solution of toluene (Figure 8). The τ0 of 2.61 ns for PDTBDT-T-DTNT, and 1.92 ns for PDTBDT-TE-DTNT, was extracted from the time-resolved photoluminescence decay for the copolymer via single exponential fit. However, it could be found that the single exponential fit failed to accurately describe the PL decay of PDTBDT-TE-DTNT, and the deviations of PDTBDT-T-DTNT PL decay from the single exponential fit were very small. Double exponential fits were also implemented to understand the PL decay of the copolymers. The τ1 and τ2 of 0.347 and 2.741 ns were extracted from the PL decay of PDTBDT-T-DTNT, and the τ1 and τ2 of 0.486 ns and 2.17 ns were extracted from the PL decay of PDTBDT-TE-DTNT. As compared with the single exponential fit, the double exponential fits could more accurately describe the PL decay for both copolymers. This indicates that there were at least two types of excited state decay for the excitations of the copolymers, and the dominant excited states, which were mainly related with the charger photo-generation process of the photovoltaic devices, provided a lifetime of 2.741 ns for PDTBDT-T-DTNT and 2.171 ns for PDTBDT-TE-DTNT. In a word, the PDTBDT-TE-DTNT had a shorter PL lifetime, and the lower exciton dissociation probability of the devices might be attributed to the shorter PL lifetime when compared to that of PDTBDT-T-DTNT.

Computational Consideration of the Copolymers
Assuming that the only difference between the two polymers is that the donor units have different flanks, it makes sense that discrepancy between the optoelectronic and aggregation, etc., of the copolymers would mainly result from their different flanks. In addition, computational calculations were implemented by using density functional theory (DFT) calculations with the B3LYP/6-31G (d,p) basis set in Gaussian 09, striking a balance between prediction of the conformation and a completion of the calculations within a reasonable time, and the alkyl side groups were replaced with methyl [55]. The trans-(or cis-) coplanar conformations of the DTBDT-T and DTBDT-TE were respectively defined by the dihedral angles (torsion angels) between the flanks of dialkylthiophene in T or TE and DTBDT planar core (Figure 9). The optimized ground-state geometries of the DTBDT-T-DTNT and DTBDT-TE-DTNT systems were presented in Figure 9a, in which the dihedral angles were about 75.5 • and 1.5 • , respectively. Meanwhile, the relaxed potential-energy scans of the T in DTBDT-T-DTNT and/or TE in DTBDT-TE-DTNT were implemented, and the results signified that the TE flanks in DTBDT-TE-DTNT exhibited narrower potential energy wells than that for the T flanks in DTBDT-T-DTNT, and indicated that the TE flanks in the DTBDT-TE-DTNT presented lower rotating energy barrier near the minimum energy conformation in contrast to T in the DTBDT-T-DTNT, which might lead to easier rotation of the TE flanks, and the enhancement of the energy relaxation channels of the excited states [56], thus contributing to the shorter PL lifetime and the lower exciton dissociation probability of the PDTBDT-TE-DTNT and/or PDTBDT-TE-DTNT-based PVCs compared with those for PDTBDT-T-DTNT and/or PDTBDT-T-DTNT-based PVCs (Figures 7 and 8).
groups were replaced with methyl [55]. The trans-(or cis-) coplanar conformations of the DTBDT-T and DTBDT-TE were respectively defined by the dihedral angles (torsion angels) between the flanks of dialkylthiophene in T or TE and DTBDT planar core (Figure 9). The optimized ground-state geometries of the DTBDT-T-DTNT and DTBDT-TE-DTNT systems were presented in Figure 9a, in which the dihedral angles were about 75.5° and 1.5°, respectively. Meanwhile, the relaxed potential-energy scans of the T in DTBDT-T-DTNT and/or TE in DTBDT-TE-DTNT were implemented, and the results signified that the TE flanks in DTBDT-TE-DTNT exhibited narrower potential energy wells than that for the T flanks in DTBDT-T-DTNT, and indicated that the TE flanks in the DTBDT-TE-DTNT presented lower rotating energy barrier near the minimum energy conformation in contrast to T in the DTBDT-T-DTNT, which might lead to easier rotation of the TE flanks, and the enhancement of the energy relaxation channels of the excited states [56], thus contributing to the shorter PL lifetime and the lower exciton dissociation probability of the PDTBDT-TE-DTNT and/or PDTBDT-TE-DTNT-based PVCs compared with those for PDTBDT-T-DTNT and/or PDTBDT-T-DTNT-based PVCs (Figures 7 and 8).

Conclusions
In conclusion, we synthesized an alternating conjugated polymer, named as PDTBDT-TE-DTNT, from DTBDT-TE and DTNT, and the controlled polymer PDTBDT-T-DTNT

Conclusions
In conclusion, we synthesized an alternating conjugated polymer, named as PDTBDT-TE-DTNT, from DTBDT-TE and DTNT, and the controlled polymer PDTBDT-T-DTNT was also prepared for comparison. Their optimal i-PVCs provided PCEs of 3.97% and 7.57%, with a V OC of 0.60 V and 0.70 V, J SC of 10.15 mA/cm 2 and 16.09 mA/cm 2 and FF of 65.16% and 67.19% under 100 mW/cm 2 illumination (AM 1.5G), respectively. Subsequently, comparative investigation of the absorption, aggregation, charge transporting, photovoltaic characteristics, etc., of PDTBDT-TE-DTNT and PDTBDT-T-DTNT were implemented. It was found that that the replacement of T with TE flanks would lead to a slight broadening of the band gap and elevation of the HOMO energy levels of the PDTBDT-TE-DTNT relative to PDTBDT-T-DTNT. Moreover, the optical calculations, exciton dissociation probability, time-resolved photoluminescence measurements and computational considerations results revealed that the lower torsional barrier of the TE flanks might contribute to the enhancement of the relaxation patterns of the excitons in the PDTBDT-TE-DTNT, thus leading to a decrease in the photo-induced excited state lifetime and exciton dissociation probability, which could be mainly attributed to the lower J SC of the PVCs from PDTBDT-TE-DTNT, in contrast to those for PDTBDT-T-DTNT and PDTBDT-T-DTNT-based PVCs.