Tuning Surface Energy of Conjugated Polymers via Fluorine Substitution of Side Alkyl Chains: Influence on Phase Separation of Thin Films and Performance of Polymer Solar Cells.

Different contents of fluorine in side alkyl chains were incorporated into three conjugated polymers (namely, PBDTTT-f13, PBDTTT-f9, and PBDTTT-f5) whose backbones consist of benzodithiophene donors and thienothiophene acceptors. These three fluorinated polymers, in comparison with the well-known analogue PTB7-Th, show comparable energy levels and optical band gaps. However, the fluorination of side alkyl chains significantly changed the surface energy of bulk materials, which leads to distinctly different self-assembly behaviors and phase separations as being mixed with PC71BM. The increased mismatch in surface energies between the polymer and PC71BM causes larger scale phase domains, which makes a sound explanation for the difference in their photovoltaic properties.

Introduction

A significant progress in bulk heterojunction (BHJ) polymer solar cells (PSCs) has been achieved through developments in both novel molecular design and optimized device architectures.[1−4] As a result, the performance of BHJ-PSCs has exceeded 11% in power conversion efficiency (PCE).[5,6] In most of the cases, the active layer of BHJ-PSCs is a blend of electron-rich polymers as a donor and electron-deficient fullerene derivatives as an acceptor to form an interpenetrating nanoscopic network. For typical conjugated polymers, the exciton diffusion length is rather short, usually on the order of tens of nanometers. If the phase domains are bigger than the diffusion length, the recombination possibility of excitons will dramatically increase.[7,8] Therefore, a well-defined phase separation with appropriate domain size is critical to efficient charge dissociation and also centrally important for providing the continuous percolation pathway for smooth charge transport with minimized charge losses.

To optimize the phase separation of a photoactive layer, many efforts have been focused on manipulating device-processing conditions, including blend ratio,[9,10] thermal/solvent annealing,[11−13] and introduction of additives.[14,15] Beyond the conditional control, the nature of the materials determined by their molecular structures is more decisive to the phase segregation of BHJ blend films.[16,17] In this regard, a compromise is often made between the polymer donor and [6,6]-phenyl-C61-butyric acid methyl ester (PCBM) acceptors. That is, a good phase separation demands a difference between their surface energies, that is, an immiscible nature, whereas a great difference is not favorable for the formation of uniform phase separation.[18,19] In this regard, surface energy was first introduced to explain the phase separation phenomenon by Kim et al. in 2010.[20] They found that the terminal groups of the P3HT polymer have great influence on the surface energy, similar to the surface morphology of P3HT/PCBM blend films. Following this strategy, Sun et al.[21] were able to tune the surface morphology of BHJ thin films by controlling the content of cyanohexane side chains in conjugated polymers. They all observed that highly miscible blends with nanoscale phase separation could be achieved when two components of an active layer had a similar surface energy, whereas polymers with large surface energy differences to PCBM tend to result in a large domain size and a significant phase separation.

In recent years, fluorinated conjugated materials have been widely used in organic photovoltaics.[22] Fluorine atoms, because of their strong electron-withdrawing effect, would lower both the highest occupied molecular orbital (HOMO) level and the lowest unoccupied molecular orbital (LUMO) level.[23−26] In addition, this modification is in favor of higher photoelectron conversion as being applied in BHJ-PSCs.[27−30] Compared with the nonfluorinated counterparts, fluorinated polymers exhibit a series of unique features such as high thermal stability, enhanced chemical resistance, and low surface energy.[31] Given the phase separation of the blend films, whether the lowered surface energy is a positive or negative factor calls for an answer. According to our knowledge, there is no precedent work that systematically investigates the role of a fluorinated side alkyl chain acting in surface energy as well as the phase separation in BHJ blend films.

Herein, a series of conjugated polymers based on alternating benzodithiophene (BDT) units as an electron donor and thienothiophene (TT) units as an electron acceptor were synthesized, named PBDTTT-fx (x = 13, 9, 5) series polymers, where x stands for the number of fluorine atoms in the side chain of TT unit. The molecular structures of these polymers are presented in Scheme . The energy levels and photoabsorption abilities of fluorinated polymers are similar to those of their known counterpart PTB7-Th.[32] As being applied as donors in BHJ-PSCs, quite different device performances were obtained. The surface morphology analysis indicates that a well-defined phase separation of the BHJ blend film corresponds to higher device performance. The phase separation, we believe, should be correlated with the surface energy of individual polymers that are deduced from the results of a group of contact angle measurements. The surface energies of PBDTTT-f13, PBDTTT-f9, PBDTTT-f5, and PTB7-Th are 19.6, 22.3, 25.8, and 32.0 mJ/m2, respectively. As for the acceptor material, PC71BM owns a surface energy of 31.0 mJ/m2. A high-quality phase separation and a high device performance were achieved as PTB7-Th:PC71BM pair was applied as the photoactive materials. This work gives us a hint that the surface energy of the targeting materials should be taken into account as designing new polymer donors.

Scheme 1

Synthetic Route and Molecular Structures of PBDTTT-f13, PBDTTT-f9, and PBDTTT-f5

Results and Discussion

Ultraviolet–visible (UV–vis) spectra of PBDTTT-fx-based polymers in chloroform and in the film state are shown in Figure . The absorption peaks of PBDTTT-f13, PBDTTT-f9, and PTB7-Th are similar, whereas the absorption peak of PBDTTT-f5 shows a blue shift compared with those of others. The former three polymers exhibited an intense absorption from 500 to 800 nm with two peaks at around 650 and 700 nm, which can be ascribed to the intramolecular charge transfer and interchain interaction, respectively.[33,34] To understand the blue shift of PBDTTT-f5, polymer solutions were heated from 30 to 100 °C. As shown in Figure , the absorption spectra indicate a temperature-dependent feature. For all polymers, the peak at around 700 nm decreased with the increase in temperature and shifted to short wavelength. An extreme case is that the corresponding peak of PBDTTT-f5 almost disappeared as the temperature was elevated to 100 °C. Knowing the fact that the peak at around 700 nm corresponds to the interchain interactions of the polymers, it can be concluded that the intermolecular interaction between PBDTTT-f5 should be less stronger than the rest of the polymers.

Figure 1

UV–vis spectra of PBDTTT-f13, PBDTTT-f9, PBDTTT-f5, and PTB7-Th in solution (a) and in the film state (b).

Figure 2

UV–vis spectra of PBDTTT-f13 (a), PBDTTT-f9 (b), PBDTTT-f5 (c), and PTB7-Th (d) in chlorobenzene solution as being heated from 30 to 100 °C with a temperature interval of 10 °C.

The LUMO energy levels (ELUMO) of PBDTTT-f13, PBDTTT-f9, PBDTTT-f5, and PTB7-Th were determined by cyclic voltammetry (CV), as shown in Figure . The onset reduction potentials (Ered) of PBDTTT-f13, PBDTTT-f9, PBDTTT-f5, and PTB7-Th were 1.31, 1.32, 1.34, and 1.39 V, respectively. ELUMO of these polymers can be calculated using the equation ELUMO = −(Ered + 4.80) (eV). Therefore, the ELUMO values of PBDTTT-f13, PBDTTT-f9, PBDTTT-f5, and PTB7-Th are −3.49, −3.48, −3.46, and −3.41 eV, respectively. The HOMO energy levels (EHOMO) can be calculated according to the equation EHOMO = ELUMO + Eg,opt. All band gaps and energy levels are summarized in Table . The PBDTTT-fx series polymers show comparable energy levels with those of PTB7-Th, indicating that fluorination of the side chain has little effect on the energy levels of the polymers.

Figure 3

CV curves of PBDTTT-f13, PBDTTT-f9, PBDTTT-f5, and PTB7-Th in n-Bu4NPF6/acetonitrile solution.

Table 1

Absorption Peaks (λ), Optical Band Gaps (Eg,opt), and Energy Levels of the Polymers

	λ (nm)
polymer	solution	film	Eg,opta (eV)	LUMOb (eV)	HOMOc (eV)
PBDTTT-f13	635, 696	650, 710	1.59	–3.49	–5.08
PBDTTT-f9	635, 696	649, 709	1.59	–3.48	–5.08
PBDTTT-f5	620, 660	645, 689	1.63	–3.46	–5.09
PTB7-Th	638, 701	646, 703	1.61	–3.41	–5.02

Calculated from the absorption band edge of the polymers in the film state.

Calculated by Ered from the CV curves.

Calculated from the difference between LUMO levels and corresponding Eg,opt.

The surface morphologies of the as spin-coated films were observed using atomic force microscopy (AFM), as shown in Figure . The PBDTTT-fx series polymers, except for PBDTTT-f5, showed surface topologies similar to those of PTB7-Th. Interdigitated fibrous nanostructures with a width of around 10 nm were observed in all polymer films. Because the number of fluorine atoms is small, for example, PBDTTT-f5, the relatively ordered self-assembled nanostructures are formed in comparison with PTB7-Th. These results are consistent with the conclusion drawn by Cho et al.[35] that the interactions of fluorine can strengthen the interactions between the side alkyl chains. Further increasing the fluorine atoms should enhance this effect. Therefore, for PBDTTT-f13 and PBDTTT-f9, interdigitated fibrous nanostructures were observed in the films as well.

Figure 4

AFM height images of the spin-coated films of PBDTTT-f13 (a), PBDTTT-f9 (b), PBDTTT-f5 (c), and PTB7-Th (d).

To investigate the miscibility of PBDTTT-fx series polymers and PC71BM in blend films, AFM investigation on PBDTTT-fx:PC71BM and PTB7-Th:PC71BM blend films was performed. In Figure , all polymers revealed similar surface profiles. However, drastic differences between the surface morphologies of the blends were observed, as shown in Figure . In Figure a, the PBDTTT-f13:PC71BM blend shows the formation of large, isolated aggregates, and the domain sizes are more than hundreds of nanometers. This morphology is probably due to the low entropy of mixing between the polymer and PC71BM because immiscible materials tend to phase-segregate into large domains during the spin-coating process. Considering the relatively short diffusion length of excitons, which is usually less than 10 nm, these overlarge phase domains are disadvantageous for efficient exciton dissociation in the BHJ films. With a lower content of fluorine atoms, as shown in Figure b, the spherical domains did not exist in the morphology of the PBDTTT-f9:PC71BM blend, whereas some distinct aggregates still exist. On the other hand, the blend films of PBDTTT-f5:PC71BM show a favorable interpenetrating network in Figure c, which are similar to the morphology of PTB7-Th:PC71BM blends shown in Figure d. Besides, the PTB7-Th:PC71BM blend films show smoother surface [with a root-mean-square (rms) roughness of 1.4 nm] than PBDTTT-fx:PC71BM blend films (the rms roughness values of PBDTTT-f5:PC71BM, PBDTTT-f9:PC71BM, and PBDTTT-f13:PC71BM blends were 1.7, 2.1, and 3.3 nm, respectively). This well-tuned morphology is beneficial to efficient exciton dissociation and charge transportation. Because immiscibility in a donor/acceptor blend arises from the large surface energy difference between the donor and acceptor materials,[36] the homogeneous network of the PBDTTT-f5:PC71BM and PTB7-Th:PC71BM blends can be ascribed to the well-matched surface energies between the donor and the acceptor. For Figure i–l, the transmission electron microscopy (TEM) images accord with the AFM results.

Figure 5

AFM height (top row), adhesion (middle row), and TEM (bottom row) images of blend films prepared by the spin coating of PBDTTT-f13:PC71BM (a,e,i), PBDTTT-f9:PC71BM (b,f,j), PBDTTT-f5:PC71BM (c,g,k), and PTB7-Th:PC71BM (d,h,l). The blend ratios of polymer:PC71BM were 1:1.5 for PBDTTT-fx and PTB7-Th.

The measurement of surface energy was recorded to clarify the reason for the formation of different morphologies of blend films. The contact angles of the PBDTTT-fx series polymers, PTB7-Th, and PC71BM were measured using liquid drops of water (θwater) and monoethylene glycol (MEG) (θMEG), as shown in Table . The surface energies (γSV) were calculated by the following equation.where θ is the contact angle of a liquid drop on the surface of the samples and γLV, γLVd, and γLVp correspond to the surface energy of water and MEG, dispersive component, and polar component of the surface energy of water and MEG, respectively.

Table 2

Pictures of Water Contact Angle (θwater) and Ethylene Glycol Contact Angle (θMEG) on the Films of PBDTTT-fx, PTB7-Th, and PC71BM

The surface energies of PBDTTT-fx, PTB7-Th, and PC71BM are summarized in Table . The increase of fluorine atoms in the BDT side chain led to a lowered surface energy value, 19.6, 22.3, 25.8, and 32.0 mJ/m2 for PBDTTT-f13, PBDTTT-f9, PBDTTT-f5, and PTB7-Th, respectively. The surface energy of PC71BM was also measured to be 31.0 mJ/m2, which is close to that of PTB7-Th. The surface energy differences between donor polymers and the PC71BM acceptor (ΔγSV) decreased consistently from PBDTTT-f13 to PTB7-Th; as a result, an uniform BHJ blend was obtained from the PTB7-Th:PC71BM mixture because of the well-matched surface energies between PTB7-Th and PC71BM. On the basis of the AFM and contact angle results, therefore, better performance is expected from the BHJ solar cells fabricated with a PTB7-Th:PC71BM film as the active layer.

Table 3

Contact Angles of PBDTTT-f13, PBDTTT-f9, PBDTTT-f5, PTB7-Th, and PC71BM Films and their Corresponding Surface Energies Calculated by Young’s Equation

polymer	θwater (deg)	θMEG (deg)	γSVd (mJ/m2)	γSVp (mJ/m2)	γSV (mJ/m2)	ΔγSVa (mJ/m2)
PBDTTT-f13	106.4	83.0	18.9	0.7	19.6	11.4
PBDTTT-f9	103.5	78.6	21.5	0.8	22.3	8.7
PBDTTT-f5	102.4	75.1	25.3	0.5	25.8	5.2
PTB7-Th	99.9	68.9	31.7	0.3	32.0	1.0
PC71BM	97.7	67.6	30.3	0.7	31.0

Calculated from the difference between γSV (PBDTTT-fx series polymers) and γSV (PC71BM).

The photovoltaic properties of these polymers were investigated in PSCs with a conventional device configuration: ITO/PEDOT:PSS (∼40 nm)/polymer:PC71BM/Ca (∼20 nm)/Al (∼80 nm). The performance of the PSCs was investigated by recording the respective current density–voltage (J–V) curves and external quantum efficiency (EQE) curves, as shown in Figure . Detailed parameters including open-circuit voltage (Voc), short-circuit current density (Jsc), fill factor (FF), and PCE are summarized in Table . The Jsc values were all consistent with the integrated photocurrents [Jsc(EQE)] determined by the EQE spectra in Figure b. The photovoltaic device based on the PBDTTT-f13:PC71BM showed the lowest PCE of only 1.98%, with Voc of 0.77 V, Jsc of 5.00 mA/cm2, and FF of 51.2%. The poor device performance of PBDTTT-f13:PC71BM can be attributed to the large-scale heterogeneities where the domain sizes are too large to achieve efficient exciton dissociation. However, the PTB7-Th:PC71BM device yields the highest PCE of 7.56%, which is almost 4 times enhanced than that of the PBDTTT-f13:PC71BM solar cell, with an increased Jsc of 14.14 mA/cm2 and FF of 68.3%. The result of a significantly enhanced Jsc value showed that the favorable interpenetrating morphology facilitates more efficient charge dissociation and transport in the active layer, which leads to a higher device performance.

Figure 6

J–V curves (a) and EQE curves (b) of the PSCs using the PBDTTT-fx series polymers and PTB7-Th as donor materials.

Table 4

Maxima J–V Characteristics and Integrated Jsc from the EQE of PSCs Based on Polymer:PC71BM Blends

polymer	Voc (V)	Jsc (mA/cm2)	FF (%)	PCEmax (%)	Jsc (EQE) (mA/cm2)
PBDTTT-f13	0.77	5.00	51.2	1.98	4.41
PBDTTT-f9	0.75	6.03	54.7	2.47	5.24
PBDTTT-f5	0.81	11.98	61.3	5.93	11.44
PTB7-Th	0.77	14.41	68.3	7.56	14.02

The PBDTTT-f9:PC71BM device showed a slightly increased PCE of 2.47% as compared with PBDTTT-f13:PC71BM, and the BHJ device based on PBDTTT-f5 showed a moderately increased PCE of 5.93%. The upward trend in PCE from PBDTTT-f13:PC71BM to PTB7-Th:PC71BM devices can be correlated with the well-tuned morphology of the active layer and the well-matched surface energies between donor polymers and the fullerene acceptor.

To gain an insight understanding of the tendency of PCEs, the charge extraction properties were studied by plotting the photocurrent density (Jph) versus the effective voltage (Veff) of the solar cells. Jph can be determined by the equation Jph = JL – JD, where JL and JD are the photocurrent densities under illumination and in the dark, respectively. Veff can be determined by the equation Veff = V0 – V, where V0 is the voltage at Jph = 0 and V is the corresponding applied voltage. As shown in Figure , Jph of each device reaches saturation (where the corresponding current density is defined as Jsat) at high Veff value, suggesting that all charge carriers generated in the photoactive layers are extracted by the electrodes. Jph under the maximum power point (where the corresponding current density is defined as Jmp) of each device is calculated from the J–V curves. Jsat, Jmp, and the ratio of Jmp/Jsat are summarized in Table . The ratio of Jmp/Jsat for the PSC based on PTB7-Th:PC71BM was 66%, indicating that more than half of the photogenerated carriers were collected by the electrodes at the maximum power point. The efficient charge collection might be resulted from the well-tuned morphology of the blend films. By contrast, the ratio of Jmp/Jsat decreased to 63, 40, and 38% when PTB7-Th was replaced by PBDTTT-f13, PBDTTT-f9, and PBDTTT-f5, respectively. These results further confirm that a good phase separation should be responsible for the efficient extraction from the active layer to the electrodes.

Figure 7

Jph–Veff curves of the PSCs based on PBDTTT-fx:PC71BM and PTB7-Th:PC71BM.

Table 5

Jsat and Jmp of the PSCs Based on PBDTTT-fx:PC71BM and PTB7-Th:PC71BM

polymer	Jsat (mA/cm2)	Jmp (mA/cm2)	Jmp/Jsat (%)
PBDTTT-f13	5.54	2.10	38
PBDTTT-f9	8.81	3.55	40
PBDTTT-f5	12.88	8.06	63
PTB7-Th	17.17	11.36	66

The charge transport property in the photoactive layers was also investigated using the space-charge-limited current equation to calculate the carrier mobilities. The corresponding J–V curves (Figures S1 and S2) and detailed data (Tables S1 and S2) are presented in the Supporting Information. For the pristine polymer films, the hole mobilities of the polymers increased with the number of F atoms in the side alkyl chains. These results indicate that increasing F contents in the side alkyl chains can improve the charge transport properties of PBDTTT-fx polymers. However, for the blend films, the hole mobilities showed a reverse trend; that is, they decreased with the number of F atoms in the side alkyl chains, and the electron mobilities also showed a downward trend when increasing the number of F atoms. These results suggest that the charge transportation inside of the active layers will be hindered with an increase in the number of F atoms because of the excessive phase separation in the blend films.

In addition to J–V and Jph–Veff characteristics, the alternating current impedance spectrometry (ACIS) of the devices was measured. Figure a–d shows the Nyquist plots of the solar cell based on PBDTTT-fx polymers and PTB7-Th blend with PC71BM at the corresponding Voc bias voltages in the dark. The complete device can be modeled as the three different regions for carrier relaxation. From high to low (i.e., left to right) frequency, the impedance responses correspond to the active layer (R1 and C1), the buffer layer region (R2 and C2), and the interface between the buffer layer and the active layer (R3 and C3). The equivalent circuit shown in Figure e was employed to fit the Nyquist plots. The parameters fitted from the ACIS are listed in Table . The equivalent circuit is represented as the combination of parallel resistance–capacitance circuits in series. Rs is regarded as the contribution of the series resistance between the two electrodes and the external circuit, which reflects the properties of the conductive electrode and is not addressed in the rest of the discussion. Generally, smaller resistances facilitate the transfer of charge carriers, thus leading to better performance.

Figure 8

(a–d) Nyquist plots of the PSCs based on different polymer:PC71BM blends as the active layers. (e) Equivalent circuit used to fit the Nyquist plots.

Table 6

Detailed Parameters of the Equivalent Circuits for the PSCs Operated in the Dark with Applied Bias Voltages Near Their Corresponding Voc

polymer	Rs (Ω)	R1 (Ω)	C1 (nF)	R2 (Ω)	C2 (nF)	R3 (Ω)	C3 (nF)
PBDTTT-f13	61.9	308.8	8.4	9079.0	1.9	11 233.0	2.4
PBDTTT-f9	66.6	92.0	9.0	440.3	17.9	1399.0	3.5
PBDTTT-f5	29.6	76.7	5.7	238.5	5.2	323.9	13.2
PTB7-Th	51.8	32.6	14.5	57.1	27.4	172.4	5.7

For the PSCs using PBDTTT-f13 as a donor, R1 shows the largest value among PBDTTT-fx and PTB7-Th polymers. Owing to the decrease in the amount of fluorine atoms in the alkyl side chain, the value of R1 shows a downward trend from PBDTTT-f13 to PTB7-Th, indicating more efficient exciton transport arising from the formation of an interpenetrating network in the BHJ layer, and this result is in accordance with the AFM image. The downward trend of R2 and R3 from PBDTTT-f13:PC71BM to PTB7-Th:PC71BM device also indicates the influence of surface morphology on the device performance.

Conclusions

In summary, a series of donor-/acceptor-conjugated polymers bearing different alkyl chains with varying fluorine contents were designed and synthesized. Their surface energies were systematically investigated and applied as a metric for the interpretation of phase separation of the blend films of polymer and PC71BM. Taking the surface energy of PC71BM (31.0 mJ/m2) as a reference, the differences between the polymer and PC71BM are increasing with the fluorine content in the alkyl chains, and consequently, the phase domains of the corresponding films also increased. These variation tendencies make a sound explanation to the device performances when these polymers are applied as donors in PSCs. For PBDTTT-f5 and PTB7-Th whose surface energies were close to that of PC71BM, their blend films showed clear phase boundaries and smaller phase domains of ∼40 nm, and the corresponding PSCs achieved higher PCEs in the form of better Jsc and FF. Phase separation of BHJ blend films is one of the most important issues that can greatly affect the device performances. The present research demonstrated that surface energy is one of the key factors that should be taken into account as designing new conjugated polymers for highly efficient PSCs.

Experimental Section

Materials and Instruments

3,3,4,4,5,5,6,6,7,7,8,8,8-Tridecafluoro-1-octanol, 3,3,4,4,5,5,6,6,6-nonafluoro-1-hexanol, and 3,3,4,4,4-pentafluoro-1-butanol were purchased from Energy Chemical. 1,6-Dibromohexane, 1,8-dibromooctane, 1,10-dibromodecane, tetrabutylammonium bromide (TBAB), tetrakis(triphenylphosphine)palladium, tetrabutylammonium hexafluorophosphate (n-Bu4NPF6), trichlorobenzene, o-dichlorobenzene (o-DCB), and pyridine were purchased from J&K Scientific. Oxalyl chloride was purchased from Tokyo Chemical Industry Co., Ltd. Dichloromethane (DCM) and petroleum ether were purchased from Titan Scientific Co., Ltd. Sodium chloride, sodium hydroxide, anhydrous magnesium sulfate, MEG, N,N-dimethylformamide (DMF), N-methyl pyrrolidone (NMP), methanol, hexane, ether, acetone, and isopropanol were purchased from Sinopharm Chemical. All of these chemicals were used without further purification, unless otherwise noted. Toluene, DMF, and DCM were purified by a solvent purification system manufactured by Innovative Technology Ltd. PTB7-Th (Mn = 57.6 kg·mol–1, Mw = 106.7 kg·mol–1, and PDI = 1.85) and PC71BM were purchased from Solarmer Materials Inc. ITO glass was obtained from CSG Holding Co., Ltd. Ca and Al were acquired from Zhongnuo Advanced Material (Beijing) Technology Co., Ltd.

Contact angle measurements were recorded on a KRÜSS DSA100 surface analysis system. The gel permeation chromatography (GPC) measurement was recorded on a Agilent PL-GPC220 chromatograph at 160 °C with trichlorobenzene as an eluent. 1H NMR spectra were recorded on an Avance III 400 MHz spectrometer produced by Bruker Co. Mass spectra were taken on a Thermo ISQ mass spectrometer using a direct exposure probe. UV–vis spectra were recorded on a Cary 2000 spectrometer produced by Agilent Technologies. CVs of polymers were recorded on a CHI600 voltammetric analyzer at room temperature in a 0.1 M n-Bu4NPF6 solution under nitrogen gas protection. The J–V characteristics were measured using a Keithley 236 Source Measure Unit, under AM 1.5G illumination from a xenon-lamp-based solar simulator (SAN-Electric Co., Ltd.) with an irradiation intensity of 100 mW/cm2. The EQE was measured using QE-R3011 from Enli Technology Co., Ltd. The AFM images were obtained using a Multimode 8 microscope (Bruker, Santa Barbara, USA). The ACIS measurements were recorded with an ac signal with an rms amplitude of 10 mV over the frequency range of 1 Hz–1 MHz in the dark on an IM6 electrochemical workstation (ZAHNER ZENNIUM, Germany). The TEM images were obtained using a Hitachi HT7700 microscope (Japan) operated at 120 kV.

Synthesis of Monomers and PBDTTT-fx Polymers

Synthesis of Compound A1

To a solution of 3,3,4,4,5,5,6,6,7,7,8,8,8-tridecafluoro-1-octanol (10.00 g, 27.50 mmol), TBAB (0.44 g, 1.38 mmol), and 1,6-dibromohexane (10.10 g, 41.30 mmol) in hexane (60 mL), an aqueous solution of NaOH (20.00 g, 0.50 mol) in water (40 mL) was added slowly. The reactant was heated to 70 °C, stirred for 12 h, and then cooled down to room temperature. After the removal of hexane, the product was extracted with diethyl ether. The organic phase was combined and dried over MgSO4. The organic solvent was evaporated under reduced pressure to yield a yellow oil (compound A1) (8.18 g, 56%). δH (400 MHz; CDCl3; TMS): 3.70 (t, J = 6.9 Hz, 2H), 3.44 (t, J = 6.5 Hz, 2H), 3.41 (t, J = 6.8 Hz, 2H), 2.39 (m, 2H), 1.86 (m, 2H), 1.58 (m, 2H), 1.43 (m, 4H). δF (400 MHz; CDCl3): −80.96 (m, 3F), −113.54 (m, 2F), −122.00 (m, 2F), −122.98 (m, 2F), −123.78 (m, 2F), −126.26 (m, 2F). GC–MS m/z: calcd 526.02; found, 526.24. Compounds B1 and C1 were synthesized with a protocol similar to that used to synthesize compound A1. For compound B1: yield: 7.85 g, 45%. δH (400 MHz; CDCl3; TMS): 3.70 (t, J = 6.9 Hz, 2H), 3.42 (m, 4H), 2.44 (m, 2H), 1.85 (m, 2H), 1.56 (m, 2H), 1.37 (m, 8H). δF (400 MHz; CDCl3): −81.12 (m, 3F), −113.73 (m, 2F), −124.67 (m, 2F), −126.09 (m, 2F). GC–MS m/z: calcd 454.06; found, 454.15.

Synthesis of Compound A2

Compound A1 (5.00 g, 9.48 mmol) was dissolved in a mixture of water (15 mL) and NMP (85 mL). The reaction mixture was heated to 100 °C for 5 days before cooling down to room temperature. After the addition of the saturated aqueous solution of NaCl (200 mL), the aqueous phase was extracted with diethyl ether. The organic phase was washed with water and dried over MgSO4. After the removal of the solvent, the residue was purified by column chromatography (SiO2, DCM) to afford a pale yellow oil (compound A2) (2.22 g, 50%). δH (400 MHz; CDCl3; TMS): 3.69 (t, J = 6.9 Hz, 2H), 3.63 (t, J = 6.6 Hz, 2H), 3.44 (t, J = 6.5 Hz, 2H), 2.39 (m, 2H), 1.56 (m, 4H), 1.38 (m, 4H). δF (400 MHz; CDCl3): −81.00 (m, 3F), −113.58 (m, 2F), −122.04 (m, 2F), −123.02 (m, 2F), −123.82 (m, 2F), −126.30 (m, 2F). GC–MS m/z: calcd 464.10; found, 464.18. Compounds B2 and C2 were synthesized with a protocol similar to that used to synthesize compound A2. For compound B2: yield: 3.88 g, 64%. δH (400 MHz; CDCl3; TMS): 3.69 (t, J = 6.9 Hz, 2H), 3.62 (t, J = 6.6 Hz, 2H), 3.43 (t, J = 6.6 Hz, 2H), 2.38 (m, 2H), 1.55 (m, 4H), 1.32 (s, 8H). δF (400 MHz; CDCl3): −81.18 (m, 3F), −113.78 (m, 2F), −124.72 (m, 2F), −126.14 (m, 2F). GC–MS m/z: calcd 392.14; found, 392.29. For compound C2: yield: 6.46 g, 55%. δH (400 MHz; CDCl3; TMS): 3.66 (t, J = 6.9 Hz, 2H), 3.61 (t, J = 6.6 Hz, 2H), 3.42 (t, J = 6.6 Hz, 2H), 2.33 (m, 2H), 1.55 (m, 4H), 1.28 (s, 12H). δF (400 MHz; CDCl3): −85.83 (t, 3F), −117.30 (s, 2F). GC–MS m/z: calcd 320.18; found, 320.32.

Synthesis of Compound M1

To a solution of 4,6-dibromo-3-fluoro-thieno[3,4-b]thiophene-2-carboxylic acid (0.30 g, 0.83 mmol) in dry toluene (10 mL), oxalyl dichloride (0.21 g, 1.67 mmol) was added slowly. After the addition of two drops of DMF, the reaction mixture was stirred at room temperature for 2 h. Toluene and the remaining oxalyl dichloride were removed under reduced pressure before adding compound A2 (3.87 g, 8.33 mmol) in dry pyridine (10 mL). Then, the reaction mixture was stirred for further 12 h. After the removal of pyridine, the product was extracted with DCM. The organic phase was washed with water and dried over MgSO4. The crude product was purified by column chromatography (SiO2, DCM/PE = 2:1) to afford a yellow oil (compound M1) (0.23 g, 34%). δH (400 MHz; CDCl3; TMS): 4.32 (t, J = 6.6 Hz, 2H), 3.71 (t, J = 6.9 Hz, 2H), 3.46 (t, J = 6.5 Hz, 2H), 2.40 (m, 2H), 176 (m, 2H), 1.61 (m, 2H), 1.44 (m, 4H). δF (400 MHz; CDCl3): −80.81 (m, 3F), −113.43 (m, 2F), −118.00 (s, 1F), −121.90 (m, 2F), −122.89 (m, 2F), −123.68 (m, 2F), −126.14 (m, 2F). GC–MS m/z: calcd 805.87; found, 805.99. Compounds M2 and M3 were synthesized similarly to compound M1. For compound M2: yield: 0.22 g, 27%. δH (400 MHz; CDCl3; TMS): 4.32 (q, J = 6.5 Hz, 2H), 3.70 (t, J = 6.9 Hz, 2H), 3.44 (t, J = 6.6 Hz, 2H), 2.39 (m, 2H), 1.75 (m, 2H), 1.58 (m, 2H), 1.37 (m, 8H). δF (400 MHz; CDCl3): −81.05 (m, 3F), −113.67 (m, 2F), −118.03 (s, 1F), −124.62 (m, 2F), −126.03 (m, 2F). GC–MS m/z: calcd 733.90; found, 733.86. For compound M3: yield: 0.252 g, 34%. δH (400 MHz; CDCl3; TMS): 4.32 (q, J = 6.5 Hz, 2H), 3.67 (t, J = 6.9 Hz, 2H), 3.43 (t, J = 6.6 Hz, 2H), 2.34 (m, 2H), 1.75 (m, 2H), 1.56 (m, 2H), 1.33 (m, 12H). δF (400 MHz; CDCl3): −85.73 (s, 3F), −117.20 (s, 2F), −118.03 (s, 1F). GC–MS m/z: calcd 661.94; found, 662.07.

Synthesis of PBDTTT-fx Series Polymers

For polymer PBDTTT-f13, BDT-containing monomer (0.22 g, 0.25 mmol), compound M1 (0.20 g, 0.25 mmol), tetrakis(triphenylphosphine)palladium (Pd(PPh3)4) (14.30 mg, 0.01 mmol), DMF (1 mL), and toluene (5 mL) were added into a microwave reaction tube under N2. The reaction was carried out under microwave irradiation at 130 °C for 3 h before cooling down to room temperature. A large amount of methanol was poured into the reaction mixture, and the precipitate was collected by filtration. The resultant polymer was extracted successively with methanol, hexane, and DCM using Soxhlet extraction apparatus. The remaining solid in the thimble was dissolved in hot chlorobenzene, and the undissolved substance was removed by filtration. A large amount of methanol was added into the filtrate, and the polymer PBDTTT-f13 was collected by filtration and dried under vacuum (0.25 g, 82%). Anal. Calcd for C55H56F14O3S6: C, 54.00; H, 4.61. Found: C, 53.64; H, 4.84%. High-temperature GPC (1,2,4-trichlorobenzene, 160 °C): Mn = 22.5 kg·mol–1, Mw = 31.8 kg·mol–1, PDI = 1.41. Polymers PBDTTT-f9 and PBDTTT-f5 were synthesized similarly to PBDTTT-f13. For PBDTTT-f9: yield: 0.24 g, 88%. Anal. Calcd for C55H60F10O3S6: C, 57.37; H, 5.25. Found: C, 57.30; H, 5.32%. High-temperature GPC (1,2,4-trichlorobenzene, 160 °C): Mn = 58.7 kg·mol–1, Mw = 156.5 kg·mol–1, PDI = 2.66. For PBDTTT-f5: yield: 0.35 g, 98%. Anal. Calcd for C55H64F6O3S6: C, 61.20; H, 5.98. Found: C, 60.44; H, 6.11%. High-temperature GPC (1,2,4-trichlorobenzene, 160 °C): Mn = 68.3 kg·mol–1, Mw = 175.9 kg·mol–1, PDI = 2.58.

Fabrication of Polymer Solar Cells

The ITO substrates were cleaned by a sequential ultrasonic procedure in an aqueous solution of detergent, distilled water, acetone, and isopropanol. The cleaned ITO substrates were dried at 150 °C to remove the trace amount of solvent.

The configuration of PSCs was ITO/PEDOT:PSS (∼40 nm)/polymer:PC71BM/Ca (∼20 nm)/Al (∼80 nm). The cleaned ITO substrates were treated with UV–ozone. PEDOT:PSS was filtered through a 0.45 μm filter and spin-coated onto the ITO surface. Subsequently, the ITO substrates with PEDOT:PSS films were moved into an oven at 150 °C for 15 min. The blend solution with 10 mg of polymer and 15 mg of PC71BM in 1 mL of o-DCB was spin-coated on the PEDOT:PSS film as a photoactive layer. On top of the photoactive layer, 20 nm of Ca and 80 nm of Al were thermally evaporated under vacuum of 1.0 × 10–5 Pa with a shadow mask covered on the surface to define the active area to be 0.04 cm2. All of these operations except for the preparation of PEDOT:PSS films were conducted in a nitrogen-filled glovebox.