A comparative study of the effect of fluorine substitution on the photovoltaic performance of benzothiadiazole-based copolymers

Abstract

Two conjugated alternating copolymers to be used as the donor materials of the active layers in polymer solar cells have been designed and synthesized via a Stille coupling reaction. The alternating structure consisted of 4,8-bis((2-ethylhexyl)oxy)benzo[1,2-b:4,5-b′]dithiophene (BDT) as the donor unit and benzo[c][1,2,5]thiadiazole (BT) or fluorinated benzo[c][1,2,5]thiadiazole (FBT) as the acceptor unit, along with a thiophene group as the π-bridge between the donor and acceptor units. Since the donor units have attached alkoxy pendant chains, both polymers were soluble in common organic solvents. UV-vis spectra of both copolymers exhibited broad absorption bands in the range of 325–900 and 380–900 nm, respectively, and corresponding low band gaps of 1.82 and 1.80 eV. After fluorination of the BT unit, the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) energy levels of the polymer were lowered and estimated to be −5.51 and −3.71 eV, respectively. It was found that substitution of an F atom into the BT units facilitated the intramolecular charge transfer. In comparison with the nonfluorinated polymer, the photovoltaic performance of the fluorinated polymer was significantly improved due to the enhanced J_sc and V_oc. Based on the ITO/PEDOT:PSS/polymer:PC₆₁BM/LiF/Al device structure, the optimal device efficiency was obtained from a device with a blend of PBDTFBT and PC₆₁BM at a weight ratio of 1 : 1. For this blend ratio, the values of J_sc and power conversion efficiency (PCE) obtained at room temperature are 7.98 mA cm⁻² and 3.62%, respectively, under the illumination of AM 1.5 (100 mW cm⁻²).

---

1. Introduction

Over the past few decades, polymer solar cells (PSCs) have attracted considerable research interest due to their unique advantages and great potential for cheap and renewable energy applications, such as the low cost solution fabrication process, large area, light weight and practicability in flexible devices.^1,2 The most prominent approach to the design of low band gap polymers for PSCs is the donor–acceptor (D–A) type concept because of the vast possibilities in the unit combinations of electron donors and electron acceptors to tailor the energy levels of the conjugated polymers.^3–5 Although a wider band gap is detrimental to harvesting more light from the solar spectrum, the larger gap resulting from a lower HOMO in the polymer has the advantage of raising the open circuit voltage (V_oc).⁶ It is well-known that low band gap polymers with high V_oc are crucial to attaining the objective of high performance PSCs. In order to achieve the design goal of low band gap and high V_oc, we have to lower the energy levels of both the HOMO and the LUMO of the conjugated polymer concurrently. However, adding electron-withdrawing groups to the polymer can generally lower the energetic position of the LUMO level.⁷ Recently, fluorine has attracted much attention as an electron-withdrawing group in the synthesis of conjugated polymers for high-efficiency solar cells.^8–20 Due to the strong electron-withdrawing characteristic of the fluorine atom, the introduction of F as a substituent to the conjugated polymer would lower both energetic positions of the HOMO and LUMO levels of the synthesized polymer, as demonstrated by Brédas et al. in a theoretical study.²¹

The fused thiophene family is frequently used as the donor unit of D–A type polymers due to its stable quinoid form giving rise to a low band gap along with good electrochemical stability.^22–24 Among the numerous reported donor units, a benzo[1,2-b:4,5-b′]dithiophene (BDT) skeleton has been widely applied as a donor unit in D–A type polymers by virtue of its highly planar nature. In addition, the BDT unit may offer an easy way to link various substituents to improve the backbone planarity and to fine-tune the energy levels, and thus ameliorate the photovoltaic properties of the devices.²⁵ Recently, many BDT-based polymers with extended π-conjugated side group systems have garnered considerable interest because the extended π-conjugation side groups could enhance π-electron delocalization and thus exhibit promising photovoltaic properties.^26–31

Meanwhile, 2,1,3-benzothiadiazole (BT) is an electron-accepting (A) molecule showing high electron mobility. Therefore, it has been utilized to constitute some n-type semiconducting polymers.^32–34 Recently, BT has also been used as the acceptor unit in conjunction with a variety of electron-donating (D) units to synthesize low band gap donors for PSCs.^{26–28,35–38} Intense light absorption and good photochemical stabilities could be achieved for the D–A type polymers containing BT units.^39,40

On the other hand, on account of the facile tunability of the electronic structures, the conjugated polymers with alternating D–π–A (wherein π represents a conjugated bridge) architecture are particularly desirable for high-efficiency PSCs.⁴¹ It has been indicated that a simple D–π–A structure using thiophenes as shorter conjugated spacers may provide the conjugated copolymer with reduced steric hindrance, extended conjugation, enhanced absorption, and improved charge transport properties.⁴²

Furthermore, although the general literature dictates that the band gap and energy levels of conjugated polymers are primarily determined by the molecular structure of the polymer backbone, recent studies have indicated that many factors, such as solubility and processability, energy level and band gap, charge transporting ability and self-organization capability, are influenced by the substituted side chains of the polymer backbone.^43–47 You's group reported that the side chain plays a significant role in modulating the V_oc and J_sc of solar cells fabricated from polymers containing an identical conjugated backbone but possessing different alkyl side chains.⁴⁸

Consequently, although a few reports have studied the conjugated copolymers containing the BDT and BT units for applications in PSCs,^49–53 nevertheless, based on above-mentioned issues, attaching different alkyl side chains to the BDT unit will give conjugated polymers different optoelectronic properties such as energy level and band gap, V_oc and J_sc. Herein we report a fluorinated D–π–A copolymer incorporating 2-ethylhexyloxy-substituted benzodithiophene (BDT) as the donor unit and fluorinated benzothiadiazole (FBT) as the acceptor unit. To the best of our knowledge, no D–π–A copolymers incorporating 2-ethylhexyloxy-substituted BDT as the donor unit and fluorinated BT as the acceptor unit for PSCs have been studied. By contrast, researchers have synthesized a nonfluorinated analogue. The effect of fluorination on optoelectronic properties and device performance of the fabricated PSCs has been investigated in detail.

2. Experimental

2.1 Materials

Benzothiadiazole (Aldrich), hydrobromic acid (Aldrich), bromine (Aldrich), 2-(tributylstannyl)thiophene (Aldrich), tetrakis(triphenylphosphine)palladium(0) (Aldrich), butyllithium (Aldrich), benzo[1,2-b:4,5-b′]dithiophene-4,8-dione (Aldrich), 4,7-dibromo-5,6-difluorobenzo[c][1,2,5]thiadiazole (Lumtec), 2-ethylhexyl bromide (Aldrich), trimethyltin chloride (Aldrich), bis(triphenylphosphine)palladium(II) dichloride (Aldrich), poly(3,4-ethylenedioxythiophene)–poly(styrenesulfonate) (PEDOT:PSS, Aldrich) and phenyl-C₆₁-butyric acid methyl ester (PC₆₁BM, FEM Tech.) were used as received. All other reagents were used as received.

2.2 Synthesis of monomers

The three monomers, 4,7-bis(5-bromothiophen-2-yl)benzo[c][1,2,5]thiadiazole (A1), 4,7-bis(5-bromothiophen-2-yl)-5,6-difluorobenzo[c][1,2,5]thiadiazole (A2), and (4,8-bis((2-ethylhexyl)oxy)benzo[1,2-b:4,5-b′]dithiophene-2,6-diyl)bis(trimethylstannane) (D1), were synthesized according to the following steps and the synthetic route is shown in Scheme 1.

Scheme 1 Synthetic routes of the monomers. (a) BT monomer (A1), (b) FBT monomer (A2), (c) BDT monomer (D1).

2.2.1 Synthesis of 4,7-dibromobenzo[c][1,2,5]thiadiazole (1). A flask charged with a mixture of 2,1,3-benzothiadiazole (1.00 g, 7.3 mmol) and 100 mL hydrobromic acid was purged with argon bubbling for 10 min. Liquid bromine (1.00 mL, 19.5 mmol) was added dropwise to the above solution over 30 min. The reaction mixture was refluxed for 6 h. The solution was cooled to room temperature and precipitated with diethyl ether. The precipitate was filtered and washed twice with diethyl ether and then vacuum dried to give compound 1. Yield: 0.860 g, 40%. ¹H NMR (500 MHz, DMSO-d₆): δ (ppm) 7.91 (s, 2H). ¹³C NMR (125 MHz, CDCl₃): δ (ppm) 152.90, 132.30, 113.86. Anal. calcd for C₆H₂Br₂N₂S: C, 24.51; H, 0.69; N, 9.53. Found: C, 24.36; H, 0.71; N, 9.45.

2.2.2 Synthesis of 4,7-di(thiophen-2-yl)benzo[c][1,2,5]thiadiazole (2). In a 150 mL glass reactor, 4,7-dibromobenzo[c][1,2,5]thiadiazole (0.50 g, 1.70 mmol) was dissolved in 75 mL dry THF and stirred under an argon atmosphere. Pd(PPh₃)₂Cl₂ (0.024 g, 0.034 mmol) and 2-(tributylstannyl)thiophene (1.08 mL, 3.40 mmol) were added to the above solution and stirred for 3 h in an ice bath environment. The solution was then heated under reflux for one day. The reaction was cooled to room temperature. After filtration, the resulting product was recrystallized from THF and then dried in a vacuum. Yield: 0.229 g, 45%. ¹H NMR (500 MHz, DMSO-d₆): δ (ppm) 8.16 (d, 2H), 8.11 (s, 2H), 7.77 (d, 2H), 7.27 (t, 2H). ¹³C NMR (125 MHz, CDCl₃): δ (ppm) 153.26, 140.01, 128.66, 128.16, 127.45, 126.62, 126.39. Anal. calcd for C₁₄H₈N₂S₃: C, 55.97; H, 2.68; N, 9.33. Found: C, 55.78; H, 2.76; N, 9.37.

2.2.3 Synthesis of 4,7-bis(5-bromothiophen-2-yl)benzo[c][1,2,5]thiadiazole (A1). Compound 2 (0.300 g, 1.00 mmol) was suspended in a 250 mL glass reactor and DMF (75 mL) stirred in at RT. The solution was then cooled down to 0 °C and N-bromosuccinimide (NBS) (0.396 g, 2.20 mmol) was added dropwise over a period of 5 min. The stirred mixture was allowed to warm to room temperature and run for one day. To the reaction mixture was added 125 mL of deionized water and the reaction mixture was extracted with diethyl ether. Organic fractions were combined and washed with 5% (w/w) NaHCO₃ and then with distilled water three times. After drying over MgSO₄, the organic phase was recrystallized from THF to afford a dark red solid. Yield: 0.158 g, 35%. ¹H NMR (500 MHz, DMSO-d₆): δ (ppm) 8.16 (s, 2H), 7.97 (d, 2H), 7.41 (d, 2H). ¹³C NMR (125 MHz, THF-d₈): δ (ppm) 154.2, 141.8, 139.6, 131.1, 129.1, 128.0, 116.6. Anal. calcd for C₁₄H₆Br₂N₂S₃: C, 36.69; H, 1.32; N, 6.11. Found: C, 36.75; H, 1.39; N, 6.04.

2.2.4 Synthesis of 5,6-difluoro-4,7-di(thiophen-2-yl)benzo[c][1,2,5]thiadiazole (3). Compound 3 was prepared in a similar manner to compound 2. The final product was a dark yellow solid. Yield: 0.122 g, 40%. ¹H NMR (500 MHz, DMSO-d₆): δ (ppm) 8.28 (d, 2H), 7.96 (d, 2H), 7.37 (t, 2H). ¹³C NMR (125 MHz, THF-d₈): δ (ppm) 151.9, 149.6, 132.1, 131.6, 129.8, 127.9, 112.3. Anal. calcd for C₁₄H₆F₂N₂S₃: C, 49.98; H, 1.79; N, 8.33. Found: C, 49.87; H, 1.68; N, 8.29.

2.2.5 Synthesis of 4,7-bis(5-bromothiophen-2-yl)-5,6-difluorobenzo[c][1,2,5]thiadiazole (A2). Monomer A2 was prepared in a similar manner to monomer A1. The final product was a dark red solid. Yield: 0.087 g, 30%. ¹H NMR (500 MHz, DMSO-d₆): δ (ppm) 8.05 (d, 2H), 7.50 (d, 2H). ¹³C NMR (125 MHz, THF-d₈): δ (ppm) 151.4, 148.6, 134.8, 131.3, 130.8, 129.1, 111.6. Anal. calcd for C₁₄H₄Br₂F₂N₂S₃: C, 36.38; H, 0.87; N, 6.06. Found: C, 36.45; H, 0.92; N, 6.09.

2.2.6 Synthesis of 4,8-bis((2-ethylhexyl)oxy)benzo[1,2-b:4,5-b′]dithiophene (4). Benzo[1,2-b:4,5-b′]dithiophene-4,8-dione (0.500 g, 2.27 mmol) was suspended in a 250 mL glass reactor and a 1 M NaOH solution (100 mL) was stirred in at room temperature. Zinc powder (0.383 g, 5.96 mmol) was added to the stirred mixture and the reaction was heated under reflux for 1 h. Tetrabutylammonium bromide (0.007 g, 0.023 mmol) was added to the reaction mixture as a phase transfer catalyst and then 2-ethylhexyl bromide (1.02 mL, 5.96 mmol) was added. The reaction was refluxed for 12 h. After cooling to room temperature, 150 mL of distilled water was added to the reaction mixture and the reaction was extracted with dichloromethane. Organic fractions were combined and washed with 5% (w/w) NaHCO₃ and then with distilled water three times. After drying over MgSO₄, the solvent was stripped off by a rotary evaporator and the residue was purified by column chromatography on silica gel with CH₂Cl₂ to afford a light yellow liquid. Yield: 0.304 g, 30%. ¹H NMR (500 MHz, DMSO-d₆): δ (ppm) 7.72 (d, 2H), 7.50 (d, 2H), 4.15 (d, 4H), 1.85 (t, 2H), 0.80–1.80 (m, 28H). ¹³C NMR (125 MHz, THF-d₈): δ (ppm) 145.33, 132.16, 130.61, 126.57, 120.91, 76.66, 41.35, 31.15, 29.90, 24.55, 23.79, 14.82, 11.99. Anal. calcd for C₂₆H₃₈O₂S₂: C, 65.21; H, 7.99. Found: C, 65.34; H, 8.07.

2.2.7 Synthesis of (4,8-bis((2-ethylhexyl)oxy)benzo[1,2-b:4,5-b′]dithiophene-2,6-diyl)bis(trimethylstannane) (D1). To a solution of 4,8-bis((2-ethylhexyl)oxy)benzo [1,2-b:4,5-b′]dithiophene (0.200 g, 0.45 mmol) in THF (50 mL) was added dropwise n-BuLi (0.40 mL, 1.0 mmol, 2.5 M in hexane) at −78 °C under argon. The reaction was kept at −78 °C for 2 h. Then trimethylchlorostannane (1.13 mL, 1.13 mmol) was added. The reaction mixture was allowed to warm to room temperature and to react for 24 h, and it was poured into water (100 mL). The crude compound was extracted with diethyl ether three times. Organic fractions were combined and washed with 5% (w/w) NaHCO₃ and then with distilled water three times. The combined organic layers were dried over anhydrous MgSO₄ and evaporated to dryness. The light yellow solid was recrystallized from isopropanol to afford a white solid. Yield: 0.100 g, 30%. ¹H NMR (500 MHz, DMSO-d₆): δ (ppm): 7.50 (s, 2H), 4.13 (d, 4H), 1.75 (t, 2H), 0.40–1.60 (m, 34H). ¹³C NMR (125 MHz, CDCl₃): δ (ppm): 143.95, 141.07, 134.55, 133.59, 128.67, 76.33, 41.37, 31.24, 29.94, 24.61, 23.87, 14.87, 12.04, −8.2. Anal. calcd for C₃₂H₅₄O₂S₂Sn₂: C, 49.76; H, 7.04. Found: C, 49.68; H, 7.12.

2.3 Synthesis of polymers

Both polymers were synthesized through a Stille coupling reaction in DMF. The nonfluorinated copolymer PBDTBT was synthesized using a 1 : 1 molar ratio of D1 to A1, while the fluorinated copolymer PBDTFBT was synthesized from a 1 : 1 molar ratio of D1 to A2 in a similar manner. The detailed synthetic procedures for both copolymers are described below and shown in Scheme 2.

Scheme 2 Polymer synthesis. (a) PBDTBT copolymer, (b) PBDTFBT copolymer.

2.3.1 Synthesis of the copolymer PBDTBT. In a 125 mL flask, the two monomers (0.4 mmol of each), (4,8-bis((2-ethylhexyl)oxy)benzo[1,2-b:4,5-b′]dithiophene-2,6-diyl)bis(trimethylstannane) (D1) and 4,7-bis(5-bromothiophen-2-yl)benzo[c][1,2,5]thiadiazole (A1) were dissolved in 75 mL of dry DMF and then flushed with argon for 10 min. Following that, 0.012 mmol (0.014 g) of Pd(PPh₃)₄ was added, and the reactant was purged with argon for another 20 min. The reaction mixture was then heated at 120 °C for 48 h under an argon atmosphere. The solution was cooled and poured into 100 mL of methanol. The crude polymer was precipitated and collected as a deep red powder, which was then subjected to Soxhlet extraction with methanol, hexane, and THF. The polymer was recovered from the THF fraction by rotary evaporation to give PBDTBT as a dark red solid. Yield: 0.110 g, 35%. ¹H NMR (500 MHz, THF-d₈, δ ppm): 8.25 (d, 2H), 8.10 (d, 2H), 7.75 (s, 2H), 7.52 (d, 2H), 4.30 (d, 4H), 1.90 (t, 2H), 0.80–1.80 (m, 28H). Anal. calcd for (C₄₀H₄₂N₂O₂S₅)_n: C, 64.65; H, 5.69; N, 3.77. Found: C, 64.49; H, 5.81; N, 3.64. GPC (THF): , , PDI = 1.10.

2.3.2 Synthesis of copolymer PBDTFBT. The copolymer PBDTFBT was synthesized from D1 and A2 in a similar manner to the copolymer PBDTBT. The final product was a red-violet solid. Yield: 0.100 g, 30%. ¹H NMR (500 MHz, THF-d₈, δ ppm): 7.80 (d, 2H), 7.60 (s, 2H), 7.50 (d, 2H), 4.30 (d, 4H), 1.90 (t, 2H), 0.80–1.80 (m, 28H). Anal. calcd for (C₄₀H₄₀F₂N₂O₂S₅)_n: C, 61.66; H, 5.18; N, 3.59. Found: C, 61.43; H, 5.26; N, 3.67. GPC (THF): , , PDI = 1.21.

2.4 Device fabrication

The device structure of the PSCs for current density–voltage (J–V) measurements is ITO/PEDOT:PSS/polymer:PC₆₁BM/LiF/Al. PBDTBT and PBDTFBT were used as the p-type donor polymers and PC₆₁BM acted as the n-type acceptor in the active layer. Before device fabrication, the ITO-coated glass substrates were first cleaned by sequential ultrasonic treatment in acetone, detergent, de-ionized water, methanol and isopropyl alcohol. The ITO surface was spin-coated with ca. 80 nm layer of PEDOT:PSS in the nitrogen-filled glovebox. After the substrate was dried for 10 min at 150 °C, spin coating of the active layer on the surface was continued. The polymer:PC₆₁BM blend solutions were spin-cast onto the PEDOT:PSS layer at 800 rpm for 30 s with different weight ratios in 1,2-dichlorobenzene (DCB) as the active layer. The thicknesses obtained for the blend films were ca. 110 nm. The devices were completed by evaporation of metal electrodes Al with an area of 6 mm² defined by masks.

2.5 Instrumentation

The molecular weights of the copolymers were analyzed by gel permeation chromatography (GPC) using a Young Lin Acme 9000 liquid chromatograph equipped with a 410 RI detector and a μ-Styragel column with THF as the carrier solvent. Thermogravimetric analyses (TGA) were conducted on a TA Instruments SDT-2960 analyzer operated at a heating rate of 10 °C min⁻¹ in air. Differential Scanning Calorimetry (DSC) thermograms were obtained with a TA Instruments modulated DSC 2920 analyzer operated at a heating rate of 10 °C min⁻¹ under a dry nitrogen purge. Ultraviolet-visible (UV-vis) spectroscopic analysis was performed in a PerkinElmer Lambda 35 UV-vis spectrophotometer. Cyclic voltammetric (CV) measurements were carried out in 0.1 M tetrabutylammonium perchlorate solution using a PGSTAT30 electrochemical analyzer (AUTOLAB Electrochemical Instrument, The Netherlands). The J–V curves were measured under illumination from a solar simulator, using a Keithley 2400 source meter. The intensity of the solar simulator was set with a primary reference cell and a spectral correction factor to give the performance under the AM 1.5 (100 mW cm⁻²) global reference spectrum (IEC 60904-9). EQE was measured with a QE-3000 (Titan Electro-Optics Co., Ltd.) lock-in amplifier under monochromatic illumination. Calibration of the incident light was performed with a monocrystalline silicon diode.

3. Results and discussion

3.1 Material design and physical properties

Recently, You's group proposed the “weak donor–strong acceptor” concept for the design of ideal donor polymers with small band gaps accompanied by deep-lying HOMO energy levels (<−5.40 eV).⁵⁴ Since lowering the HOMO energy level favors the achievement of higher V_oc, we adopted the “weak donor–strong acceptor” concept as our design motif in this study. However, in order to keep the band gap constant and increase the V_oc in the meantime, both the HOMO and LUMO energy levels of the conjugated polymer must be reduced concurrently. For fine-tuning both energy levels to increase the V_oc, substitution of F atoms onto the polymer backbone is one of the most efficient approaches. Because fluorine has small size and high electronegativity, it can be used to adjust the energy levels of organic molecules without giving rise to any of the significant steric hindrance usually generated by larger electro-withdrawing substituents.⁵⁵ Compared to the nonfluorinated polymer, the energy band gap of a fluorinated polymer is almost preserved. In addition, in order to improve the miscibility of the polymer without disturbing its crystallinity in the bulk heterojunction (BHJ) blend film, it is important to select side chains in the design strategy of the fluorinated polymer.

In this study, we have synthesized two new D–A type copolymers containing benzodithiophene (BDT) as the donor unit and nonfluorinated and fluorinated benzothiadiazole (BT) as the acceptor unit, respectively. Since the BDT unit is a weak donor fusing less electron-rich benzene with electron-rich thiophene and the BT moiety is a strong acceptor with strong electron-withdrawing conjugated aromatics, we apply this motif to design the donor polymers with small band gaps and high V_oc. In addition, benzothiadiazole (BT) is an electron-accepting heterocycle showing high electron mobility and BDT is a linearly symmetrical and heteroarene-fused coplanar unit. It is expected that a broad absorption band and high power conversion efficiency (PCE) could be attained for the D–A type copolymer using BDT as the donor and BT as the acceptor. On the other hand, two fluorine atoms were added to the BT unit in order to lower both the HOMO and LUMO energy levels of the conjugated polymer simultaneously and hence enhance the V_oc.

In the design of our donor polymers, conjugated polymers with an alternating D–π–A architecture are adopted due to the facile tunability of their electronic structures.⁴¹ As stated in the literature, thiophene or alkylthiophene units are generally used as π-bridges in D–π–A polymers to achieve efficient photovoltaic properties. It has been found that using thiophenes as short conjugated spacers between the electron donor and electron acceptor in a simple D–π–A structure may give the conjugated polymer good optoelectronic properties in the wavelength tuning, electronic coupling, and absorption capability between the donor and acceptor. Consequently, a polymer solar cell with red-shifted absorption and good power conversion efficiency may be attained.⁴⁰ Hence, in the present study the thiophene group was used as a spacer between BDT and BT units in our D–A type copolymer. Although the molecular structure of the polymer backbone has a great effect in determining the electronic properties of conjugated polymers, the solubilizing alkyl side chains also play an important role in the final photovoltaic performances of the resulting polymers. It has been found that many factors, such as solubility and processability, energy level and band gap, charge transporting ability and self-organization capability, are influenced by the substituted side chains of the polymer backbone. It has been suggested that long and branched side chains would weaken the intermolecular interactions and enhance V_oc while short and straight side chains would increase the intermolecular interactions and favor the short-circuit current (J_sc).⁵⁶ In addition, it is worth noting that introducing the electron-donating oxygen atoms to the donor units also can enhance the J_sc and FF of devices, although a moderate decrease in V_oc will occur.¹⁴ Therefore, we introduced two short and branched alkoxy side chains, i.e., 2-ethylhexyloxy groups, into the BDT unit to optimize the intermolecular interactions and hence lead to a balance between J_sc and V_oc. Thus, two D–π–A copolymers consisting of an alkoxy-substituted BDT donor unit, thiophene π-bridge, and the nonfluorinated and fluorinated BT acceptor units, respectively, were envisioned and synthesized. The synthetic routes toward the synthesis of monomers and conjugated copolymers are outlined in Schemes 1 and 2.

With the attachment of short and branched alkoxy side chains on the BDT unit, the band gap of the copolymers could be fine-tuned and the solubility in organic phases increases. Both copolymers dissolve well in common organic solvents such as THF, toluene, chloroform, and 1,2-dichlorobenzene due to the existence of two 2-ethylhexyloxy groups on the BDT unit. The quantitative solubility data for both copolymers in chloroform and 1,2-dichlorobenzene are larger than 80 mg mL⁻¹. Furthermore, due to the substitution of two fluorine atoms into the BT unit, the HOMO and LUMO energy levels of the fluorinated copolymer were expected to be lowered to enhance the V_oc. Both the nonfluorinated and fluorinated copolymers (PBDTBT and PBDTFBT) exhibited high glass transition temperatures (T_g) of 156 and 175 °C, respectively, as a result of the rigid donor and acceptor units. In particular, by virtue of the strong electronegativity of fluorine atoms, the fluorinated polymer exhibited a higher T_g which might arise from stronger intermolecular interactions resulting in a more compact architecture. In a similar trend, TGA results showed a higher T_d (5% degradation) for the fluorinated polymer compared to that of the nonfluorinated one. The relevant physical properties of both polymers are listed in Table 1.

Table 1 Molecular weights, thermal and optoelectronic properties of the synthesized copolymers

Polymer	Mn (kg mol^−1)	Mw (kg mol^−1)	PDI	Tg (°C)	Td^a (°C)	Eg^b (eV)	HOMO (eV)	LUMO^c (eV)
a Decomposition temperature at 5% weight loss.b Obtained by Tauc relation.c Obtained from LUMO = HOMO + Eg.
PBDTBT	11.2	12.3	1.10	156	326	1.82	−5.38	−3.56
PBDTFBT	12.6	15.3	1.21	175	331	1.80	−5.51	−3.71
P3HT (commercial)	13.2	17.0	1.29	—	—	1.90	−5.20	−3.30

---

3.2 Optical properties and band gaps

The normalized UV-vis absorption spectra for the films of the two synthesized polymers and P3HT are presented in Fig. 1. As seen from Fig. 1, the UV-vis absorption spectrum of the PBDTBT film exhibits a broad absorption band ranging from 325 to 900 nm with a weak absorption centered at around 346 nm and two strong peaks centered at about 433 and 550 nm. The two strong absorptions positioned at about 433 and 550 nm are well consistent with the dual-band optical feature arising from the “donor–acceptor” synthetic approach. From the dual-band spectrum, the first absorptions are attributed to π–π* transitions within the polymer backbone, whereas the last absorption band is ascribed to the intramolecular charge transfer (ICT) among BDT, π-bridge and BT units in the polymer backbone.⁵⁷ For the weak absorptions in the range of 800–900 nm which cannot be observed in the spectra of samples in some literature^49,51,52 because of the very small absorption coefficients in this region, it may be because the absorption coefficients of our samples are higher than those of the above-mentioned samples that we can still observe the absorptions in the range of 800–900 nm of the UV-vis spectra.

Fig. 1 (a) Normalized UV-vis absorption spectra of PBDTBT, PBDTFBT, and P3HT, (b) plot of (αhν)² vs. hν via Tauc relation for PBDTBT and PBDTFBT film.

In a similar manner, the spectrum of the PBDTFBT film also exhibits the dual-band of absorption ranging from 380 to 700 nm with tailing to around 900 nm. The first strong absorption centered at 435 nm is ascribed to the π–π* transitions within the polymer backbone, while the peak positioned at 579 nm arises from the ICT among the BDT, thiophene spacer and BT units in the polymer backbone. As compared to the two absorptions around 433 and 550 nm in the spectrum of PBDTBT film, it is noticeable that both absorptions are significantly enhanced and the peak arising from ICT is red-shifted. This may be attributed to the substitution of fluorine atoms into the BT units which has intensified the interchain interactions and ICT effect. In comparison with the spectra of both synthesized copolymers, it is noteworthy that the absorption maximum of P3HT just intervenes between the two strong absorption peaks of PBDTBT and PBDTFBT. This interesting observation suggests that both the PBDTBT and PBDTFBT copolymers may be employed in a tandem solar cell, stacking two cells with two active layers absorbing different parts of the solar spectrum.

The optical band gaps obtained from the Tauc relation⁵⁸ αhν = B(hν − E_opt)ⁿ for PBDTBT and PBDTFBT are 1.82 and 1.80 eV (Fig. 1(b)), respectively, and are shown in Table 1. Both optical band gaps are smaller than that (E^opt_g = 1.9 eV) of the widely used regioregular P3HT.

3.3 Electrochemical properties and energy levels

Both the PBDTBT and PBDTFBT copolymers display similar electrochemical oxidation characteristics, as seen from Fig. 2. The HOMO energy levels of both polymers can be calculated from the onset oxidation potential [E_ox,onset] based on the reference energy level of ferrocene (4.8 eV below the vacuum level, which is defined as zero) according to eqn (1). The LUMO level can be obtained from eqn (2) based on the E_g from Fig. 1(b). E_FC is the potential of the internal standard, the ferrocene/ferrocenium (Fc/Fc⁺) redox couple.

HOMO = −[E_ox,onset − E_FC + 4.8] eV (1)

LUMO = HOMO + E_g (2)

Fig. 2 Cyclic voltammograms of polymer films on an ITO substrate in CH₃CN/AcOH (v/v = 7/1) containing 0.1 M tetrabutylammonium perchlorate at a scan rate of 50 mV s⁻¹. (a) PBDTBT copolymer, (b) PBDTFBT copolymer.

As seen in Fig. 2, the E_ox,onset values for PBDTBT and PBDTFBT have been determined as 0.72 V and 0.85 V vs. Ag/Ag⁺, respectively. E_FC is 0.14 V vs. Ag/Ag⁺. Hence, the HOMO energies for PBDTBT and PBDTFBT have been evaluated to be −5.38 eV and −5.51 eV, respectively. Correspondingly, the LUMO levels determined from eqn (2) are −3.56 eV and −3.71 eV. Fig. 3 shows the energy level diagram of the HOMO and LUMO energy levels of PBDTBT and PBDTFBT relative to the work function of the electrodes. Recent studies dictate that the proposed “ideal” conjugated polymer should exhibit a HOMO energy level lower than −5.4 eV to acquire a high V_oc.⁵⁴ As evident from Fig. 3, it is obvious that both of the synthesized polymers may well match the requirement of “ideal” polymers for PSC applications. In comparison with the PBDTBT copolymer, the fluorinated material has even lower HOMO and LUMO levels than expected. Therefore, it is evident that fluorination of the conjugated copolymer is an effective way to lower both the HOMO and the LUMO energy levels and hence to enhance V_oc. Furthermore, it is noteworthy that both copolymers demonstrate superior deep-lying HOMO levels in comparison with most reported polymers with similar structures mentioned in ref. 47–51.

Fig. 3 Energy level diagram for the polymers in the polymer solar cell.

3.4 Density functional theory (DFT) calculation

In order to evaluate the influence of fluorine atoms on the electronic and optical properties of PBDTFBT, a model system based on the D–π–A repeating unit containing two fluorine atoms on the benzothiadiazole moiety was performed by using density functional theory (DFT) with the B3LYP/6-31G* method implemented in the Gaussian 09 program package. Alkyl chains are replaced by methyl groups to simplify the calculations. Computational study of the nonfluorinated analogue was also performed for comparison. In order to realize the UV-vis spectra of both copolymers in advance, the TDB3LYP/6-31G* method was used to estimate the excitation energies of two (PBDTBT-2 and PBDTFBT-2) and three (PBDTBT-3 and PBDTFBT-3) repeating units of the studied polymers (Fig. 4).

Fig. 4 (a) Absorption calculation by dimer model using TDB3LYP/6-31G*, (b) absorption calculation by trimer model using TDB3LYP/6-31G*, both calculations assuming gas phase with 20 states and 50–50 singlet and triplets with maxima at ∼470 nm and ∼770 nm.

As seen from Table 2 and Fig. 4, PBDTFBT was predicted to have both a similar band gap and UV-vis absorption spectrum to PBDTBT. The calculated band gaps for both copolymers are very close to the values obtained from their UV-vis spectra as shown in Fig. 1(b). Both the HOMO and LUMO energy levels were slightly lower in PBDTFBT than those in PBDTBT. The simulated spectra in Fig. 4 for both species show dual-band absorption peaks that greatly resemble the observed experimental spectra of PBDTBT and PBDTFBT. The simulated data from the DFT calculations were nearly in agreement with the experimental results estimated from the cyclic voltammograms and UV-vis spectra. Our results confirm that the previous discovery of the utility of F atoms in lowering the energetic position of both the HOMO and LUMO levels of related conjugated polymers. Due to a similar band gap but more strong UV-vis absorptions along with a lower HOMO level, the PBDTFBT-based BHJ device could offer an enhanced J_sc and a larger V_oc than those of the nonfluorinated analogue (PBDTBT). Furthermore, using the shorter analog of these two polymers, the distributions of the molecular orbital wave functions of the two oligomers are calculated and displayed in Fig. 5. For both oligomers, the HOMO wave functions are well delocalized along the polymer backbone composed of donor and acceptor units, whereas their LUMO wave functions are more localized at the acceptor moieties. These images further corroborate the formation of a well-defined D–π–A structure and the intramolecular charge transfer (ICT) behavior between the donor unit and the acceptor unit of the material.

Table 2 HOMO, LUMO, and E_g obtained from DFT calculation

Polymer	HOMO (eV)	LUMO (eV)	Eg (eV)
PBDTBT	−4.75	−2.95	1.80
PBDTFBT	−4.91	−3.07	1.84

---

Fig. 5 The calculated HOMO and LUMO orbitals of the polymers by B3LYP/6-31G*. (a) PBDTBT copolymer, (b) PBDTFBT copolymer.

3.5 Hole mobility and photovoltaic properties

The charge-carrier mobilities of both copolymers and P3HT were determined by using the space-charge-limited current (SCLC) method in the direction perpendicular to the electrodes. The hole mobilities are estimated to be 7.32 × 10⁻⁴, 1.26 × 10⁻³, and 2.94 × 10⁻⁴ cm² V⁻¹ s⁻¹ for PBDTBT, PBDTFBT, and P3HT, respectively. The hole mobility of PBDTFBT is remarkably higher than those of P3HT and the copolymer without fluorines (PBDTBT). To compare the PCE of PBDTFBT with those of the nonfluorinated analogue (PBDTBT) and P3HT, we prepared the blend solutions of polymer:PC₆₁BM as the active layer. Because the light absorbance is affected by the amount of PC₆₁BM, we investigated the effect of weight ratio of PC₆₁BM on the device performance with varying weight fractions of PC₆₁BM. The device structure employed was ITO/PEDOT:PSS/polymer:PC₆₁BM/LiF/Al. The active layers were spin-coated from 1,2-dichlorobenzene (DCB) solutions of the polymers and PC₆₁BM, and the optimized weight ratios of polymer to PC₆₁BM were 1 : 2, 1 : 1, and 1 : 1 for PBDTBT, PBDTFBT, and P3HT, respectively. The photovoltaic performances of the devices for the active layers cast at room temperature (RT) were carried out under illumination from a solar simulator at 100 mW cm⁻² light intensity. The corresponding short-circuit current (J_sc), open-circuit voltage (V_oc), fill factor (FF), and power conversion efficiency (PCE, η) of the three polymers are listed in Table 3. The current density–voltage (J–V) characteristics of the cells are shown in Fig. 6. We observed the best performance from the device with the PBDTFBT:PC₆₁BM blend film, which gave the highest short-circuit current of about 7.98 mA cm⁻² and power conversion efficiency of up to 3.62% without post-treatment or any additives.

Table 3 Photovoltaic characteristics of devices for the polymer:PC₆₁BM blends

Polymer	Polymer:PC61BM (w/w)	Voc (V)	Jsc (mA cm^−2)	FF (%)	η (%)
PBDTBT	1 : 2	0.75	6.71	48.2	2.44
PBDTFBT	1 : 1	0.87	7.98	52.2	3.62
P3HT	1 : 1	0.58	7.34	58.7	2.50

---

Fig. 6 J–V characteristics of devices under AM 1.5 simulated solar illumination at an intensity of 100 mW cm⁻².

As seen in Fig. 6 and Table 3, it is noticeable that the J_sc of PBDTFBT is higher than that of PBDTBT. In comparison with P3HT, the ability of the nonfluorinated and fluorinated polymers to acquire higher J_scs is likely to be due to the higher hole mobilities of both polymers. In particular, with the substitution of fluorine atoms, the PBDTFBT has even higher hole mobility than that of PBDTBT and hence affords a higher J_sc. In addition, it is noteworthy that the V_oc value of PBDTFBT is the highest among the three polymers because the introduction of F into the conjugated backbone has lowered the HOMO and LUMO energy levels of the copolymer. Because the values of J_sc and V_oc are simultaneously higher for PBDTFBT, the power conversion efficiency of the solar cell for PBDTFBT is the highest among the three polymers. However, both synthesized copolymers display lower FF values compared to P3HT. This may imply that if the FF value is high enough, the PBDTBT copolymer may also have a higher PCE than that of P3HT. Nevertheless, both polymers show superior V_ocs and similar or better J_scs, PCEs, and mobilities under the processing conditions without post-treatment or any additives compared with most of the reported polymers mentioned in ref. 49–53.

With regard to the lower FF value, it normally arises from shunt resistance, series resistance, and film-forming properties, etc. We suspect that the low FF values of both polymers are mainly owing to the non-optimized morphology and thickness of the copolymer/PC₆₁BM blend film. Therefore, an attempt to improve the PCEs of both materials by using different solvents or adding additives to the polymer/PC₆₁BM blend and modifying the thickness of the blend film is still ongoing and higher PCEs are anticipated.

3.6 External quantum efficiency and solar cell performance

External quantum efficiency (EQE) measurements with monochromatic waves from 300 to 900 nm were also performed on the solar cell devices with the optimized weight ratios of polymer to PC₆₁BM (Fig. 7). As depicted in the figure, both EQE spectra of the PBDTBT and PBDTFBT blend films exhibit a photoresponsive region ranging from 300 to 900 nm, somewhat similar to the optical data. The spectrum of PBDTBT shows two remarkable peaks positioned at 360 and 553 nm, whereas that of PBDTFBT displays two peaks centered at around 339 and 580 nm. Both spectra are well consistent with their optical data. The first peak of both spectra arises mainly from the absorption of PC₆₁BM, and the higher EQE value for the PBDTBT device is due to the higher weight ratio of PC₆₁BM in the blend film. Conversely, the latter peak of both spectra obviously resulted from the UV-vis absorption of both copolymers. Comparing both spectra, it is apparent that most of the EQE values of the PBDTFBT blend film are higher than those of the PBDTBT one by virtue of the stronger and red-shifted UV-vis absorption of PBDTFBT. The higher EQE values and enhanced J_sc for the PBDTFBT blend film are attributed to the rapid charge transfer of the dissociated excitons in the polymer backbone as a result of the higher hole mobility of PBDTFBT. It may suggest that substitution of fluorine atoms onto the BT units of the polymer backbone may facilitate the intramolecular charge transfer between the BDT and BT moieties.

Fig. 7 EQE spectra of PBDTBT/PC₆₁BM and PBDTFBT/PC₆₁BM blends.

4. Conclusions

In this study, both nonfluorinated and fluorinated copolymers based on BDT and BT units have been synthesized and utilized as the donor materials in the active layer of BHJ-type polymer solar cells. The branched alkoxy group was introduced as a substituent on the BDT unit and a thiophene π-bridge was inserted into the BDT and BT units to optimize the performance of the conjugated polymers. The nonfluorinated polymer PBDTBT displayed a wide UV-vis absorption ranging from 325 to 900 nm and an optical band gap of around 1.82 eV. In comparison to PBDTBT, the PBDTFBT copolymer exhibited a similar and red-shifted absorption spectrum as well as a band gap of ca. 1.80 eV. Both copolymers exhibited high hole mobilities compared to that of P3HT, which might arise from the coplanarity of the polymer main chains. Both copolymers displayed superior deep-lying HOMO levels in comparison with most of reported polymers with similar structures. The fluorinated copolymer PBDTFBT depicted a low-lying HOMO level of 5.51 eV along with a high V_oc of 0.87 V. It can be concluded that substitution of fluorine atoms onto the polymer backbone is an effective approach to control the HOMO and LUMO energy levels without a large change in the energy bandgap. As evident from the UV-vis spectra, the ICT among the D–π–A units of the backbone was much increased after fluorine substitution. Through the modification by fluorination on the BT units, the photovoltaic performance of PBDTFBT at room temperature was significantly improved due to the enhanced J_sc and V_oc, and the power conversion efficiency of the device reached 3.62% under white light illumination (100 mW cm⁻²).

Acknowledgements

We gratefully acknowledge the support of the Ministry of Science and Technology in Taiwan through Grant NSC 99-2221-E-390-001-MY3.

References

1. F. C. Krebs, S. A. Gevorgyan and J. Alstrup, J. Mater. Chem., 2009, 19, 5442–5451.
2. T. D. Nielsen, C. Cruickshank, S. Foged, J. Thorsen and F. C. Krebs, Sol. Energy Mater. Sol. Cells, 2010, 94, 1553–1571.
3. J. Peet, J. Y. Kim, N. E. Coates, W. L. Ma, D. Moses, A. J. Heeger and G. C. Bazan, Nat. Mater., 2007, 6, 497–500.
4. X. Zhan, Z. Tan, B. Domercq, Z. An, X. Zhang, S. Barlow, Y. Li, D. Zhu, B. Kippelen and S. R. Marder, J. Am. Chem. Soc., 2007, 129, 7246–7247.
5. E. Zhou, S. Yamakawa, K. Tajima, C. Yang and K. Hashimoto, Chem. Mater., 2009, 21, 4055–4061.
6. W. Shockley and H. J. Queisser, J. Appl. Phys., 1961, 32, 510–519.
7. Y. J. Cheng, S. H. Yang and C. S. Hsu, Chem. Rev., 2009, 109, 5868–5923.
8. H. Y. Chen, J. H. Hou, S. Q. Zhang, Y. Y. Liang, G. W. Yang, Y. Yang, L. P. Yu, Y. Wu and G. Li, Nat. Photonics, 2009, 3, 649–653.
9. H. X. Zhou, L. Q. Yang, A. C. Stuart, S. C. Price, S. B. Liu and W. You, Angew. Chem., Int. Ed., 2011, 50, 2995–2998.
10. S. C. Price, A. C. Stuart, L. Q. Yang, H. X. Zhou and W. You, J. Am. Chem. Soc., 2011, 133, 4625–4631.
11. C. Chang, L. Zuo, H. Yip, Y. Li, C. Li, C. Hsu, Y. Cheng, H. Chen and A. K.-Y. Jen, Adv. Funct. Mater., 2013, 23, 5084–5090.
12. J. You, L. Dou, K. Yoshimura, T. Kato, K. Ohya, T. Moriarty, K. Emery, C.-C. Chen, J. Gao, G. Li and Y. Yang, Nat. Commun., 2013, 4, 1446.
13. N. Wang, Z. Chen, W. Wei and Z. Jiang, J. Am. Chem. Soc., 2013, 135, 17060–17068.
14. T. L. Nguyen, H. Choi, S.-J. Ko, M. A. Uddin, B. Walker, S. Yum, J.-E. Jeong, M. H. Yun, T. J. Shin, S. Hwang, J. Y. Kim and H. Y. Woo, Energy Environ. Sci., 2014, 7, 3040–3051.
15. M. Zhang, X. Guo, S. Zhang and J. Hou, Adv. Mater., 2014, 26, 1118–1123.
16. Y. Liu, J. Zhao, Z. Li, C. Mu, W. Ma, H. Hu, K. Jiang, H. Lin, H. Ade and H. Yan, Nat. Commun., 2014, 5, 5293.
17. J. W. Jo, S. Bae, F. Liu, T. P. Russell and W. H. Jo, Adv. Funct. Mater., 2015, 25, 120–125.
18. J. Wolf, F. Cruciani, A. E. Labban and P. M. Beaujuge, Chem. Mater., 2015, 27, 4184–4187.
19. J. Yuan, Y. Zou, R. Cui, Y. H. Chao, Z. Wang, M. Ma, Y. He, Y. Li, A. Rindgen, W. Ma, D. Xiao, Z. Bo, X. Xu, L. Li and C. S. Hsu, Macromolecules, 2015, 48, 4347–4356.
20. J. Yuan, L. Qiu, Z. Zhang, Y. Li, Y. He, L. Jiang and Y. Zou, Chem. Commun., 2016 10.1039/c6cc01771a.
21. J. L. Brédas and A. J. Heeger, Chem. Phys. Lett., 1994, 217, 507–512.
22. K. Lee and G. A. Sotzing, Macromolecules, 2001, 34, 5746–5747.
23. G. A. Sotzing and K. Lee, Macromolecules, 2002, 35, 7281–7286.
24. B. Lee, M. S. Yavuz and G. A. Sotzing, Macromolecules, 2006, 39, 3118–3124.
25. L. Ye, S. Q. Zhang, L. J. Huo, M. J. Zhang and J. H. Hou, Acc. Chem. Res., 2014, 47, 1595–1603.
26. M. J. Zhang, Y. Gu, X. Guo, F. Liu, S. Q. Zhang, L. J. Huo, T. P. Russell and J. H. Hou, Adv. Mater., 2013, 25, 4944–4949.
27. Q. Liu, X. C. Bao, S. G. Wen, Z. K. Du, L. L. Han, D. Q. Zhu, Y. H. Chen, M. L. Sun and R. Q. Yang, Polym. Chem., 2014, 5, 2076–2082.
28. C. Y. Kuo, W. Y. Nie, H. H. Tsai, H. J. Yen, A. D. Mohite, G. Gupta, A. M. Dattelbaum, D. J. William, K. C. Cha, Y. Yang, L. Wang and H. L. Wang, Macromolecules, 2014, 47, 1008–1020.
29. J. H. Kim, C. E. Song, B. S. Kim, I. N. Kang, W. S. Shin and D. H. Hwang, Chem. Mater., 2014, 26, 1234–1242.
30. K. Kranthiraja, K. Gunasekar, W. Cho, M. Song, Y. G. Park, J. Y. Lee, Y. Shin, I. N. Kang, A. Kim, H. J. Kim, B. S. Kim and S. H. Jin, Macromolecules, 2014, 47, 7060–7069.
31. J. X. Wang, M. J. Xiao, W. C. Chen, M. Qiu, Z. K. Du, W. G. Zhu, S. G. Wen, N. Wang and R. Q. Yang, Macromolecules, 2014, 47, 7823–7830.
32. J. Chen and Y. Cao, Acc. Chem. Res., 2009, 42, 1709–1718.
33. A. C. Arias, J. D. MacKenzie, R. Stevenson, J. J. Halls, M. Inbasekaran, E. P. Woo, D. Richards and R. H. Friend, Macromolecules, 2001, 34, 6005–6013.
34. H. J. Snaith, N. C. Greenham and R. H. Friend, Adv. Mater., 2004, 16, 1640–1645.
35. M. Svensson, F. L. Zhang, S. C. Veenstra, W. J. H. Verhees, J. C. Hummelen, J. M. Kroon, O. Inganas and M. R. Andersson, Adv. Mater., 2003, 15, 988–991.
36. N. Blouin, A. Michaud and M. Leclerc, Adv. Mater., 2007, 19, 2295–2300.
37. A. J. Moule, A. Tsami, T. W. Bunnagel, M. Forster, N. M. Kronenberg, M. Scharber, M. Koppe, M. Morana, C. J. Brabec, K. Meerholz and U. Scherf, Chem. Mater., 2008, 20, 4045–4050.
38. S. Zhang, Y. Guo, H. Fan, Y. Liu, H. Y. Chen, G. Yang, X. Zhan, Y. Liu, Y. Li and Y. Yang, J. Polym. Sci., Part A: Polym. Chem., 2009, 47, 5498–5508.
39. R. S. Kularatne, P. Sista, H. Q. Nguyen, M. P. Bhatt, M. C. Biewer and M. C. Stefan, Macromolecules, 2012, 45, 7855–7862.
40. R. S. Kularatne, F. J. Taenzler, H. D. Magurudeniya, J. Du, J. W. Murphy, E. E. Sheina, B. E. Gnade, M. C. Biewer and M. C. Stefan, J. Mater. Chem., 2013, A1, 15535–15543.
41. X. Guo, M. Zhang, L. J. Huo, F. Xu, Y. Wu and J. H. Hou, J. Mater. Chem., 2012, 22, 21024–21031.
42. A. V. Patil, W. H. Lee, K. Kim, Y. S. Lee, I. N. Kang and S. H. Lee, Synth. Met., 2011, 161, 1838–1844.
43. M.-H. Chen, J. Hou, Z. Hong, G. Yang, S. Sista, L.-M. Chen and Y. Yang, Adv. Mater., 2009, 21, 4238–4242.
44. N. Blouin, A. Michaud and M. Leclerc, Adv. Mater., 2007, 19, 2295–2300.
45. W. Yue, Y. Zhao, S. Y. Shao, H. K. Tian, Z. Y. Xie, Y. H. Geng and F. S. Wang, J. Mater. Chem., 2009, 19, 2199–2206.
46. E. G. Wang, M. Wang, L. Wang, C. H. Duan, J. Zhang, W. Z. Cai, C. He, H. B. Wu and Y. Cao, Macromolecules, 2009, 42, 4410–4415.
47. W. Nie, C. M. MacNeill, Y. Li, R. E. Noftle, D. L. Carroll and R. C. Coffin, Macromol. Rapid Commun., 2011, 32, 1163–1168.
48. L. Q. Yang, H. X. Zhou and W. You, J. Phys. Chem. C, 2010, 114, 16793–16800.
49. Y. Huang, F. Liu, X. Guo, W. Zhang, Y. Gu, J. P. Zhang, C. C. Han, T. P. Russell and J. H. Hou, Adv. Energy Mater., 2013, 3, 930–937.
50. N. Wang, Z. Chen, W. Wei and Z. H. Jiang, J. Am. Chem. Soc., 2013, 135, 17060–17068.
51. H. Zhou, L. Yang, A. C. Stuart, S. C. Price, S. Liu and W. You, Angew. Chem., Int. Ed., 2011, 50, 2995–2998.
52. A. C. Stuart, J. R. Tumbleston, H. Zhou, W. Li, S. Liu, H. Ade and W. You, J. Am. Chem. Soc., 2013, 135, 1806–1815.
53. F. Wu, D. Zha, L. Chen and Y. Chen, J. Polym. Sci., Part A: Polym. Chem., 2013, 51, 1506–1511.
54. H. Zhou, L. Yang, S. Stoneking and W. You, ACS Appl. Mater. Interfaces, 2010, 2, 1377–1383.
55. H. J. Son, B. Carsten, I. H. Jung and L. Yu, Energy Environ. Sci., 2012, 5, 8158–8170.
56. L. Q. Yang, H. X. Zhou and W. You, J. Phys. Chem. C, 2010, 114, 16793–16800.
57. P. M. Beaujuge, C. M. Amb and J. R. Reynolds, Acc. Chem. Res., 2010, 43, 1396–1407.
58. J. Tauc, Amorphous and Liquid Semiconductors, Plenum, New York, 1974, pp. 159–220.

---