Vertical Distribution to Optimize Active Layer Morphology for Efficient

8 hours ago - Find my institution .... Vertical Distribution to Optimize Active Layer Morphology for Efficient All-Polymer Solar Cells by J71 as a Com...
2 downloads 0 Views 3MB Size
Article Cite This: Macromolecules XXXX, XXX, XXX−XXX

pubs.acs.org/Macromolecules

Vertical Distribution to Optimize Active Layer Morphology for Efficient All-Polymer Solar Cells by J71 as a Compatibilizer Siqi Liu,†,‡,⊥ Dong Chen,†,‡,⊥ Weihua Zhou,*,†,§ Zoukangning Yu,†,§ Lie Chen,†,‡ Feng Liu,*,∥ and Yiwang Chen*,†,‡ †

College of Chemistry, ‡Institute of Polymers and Energy Chemistry (IPEC), and §School of Material Science and Engineering, Nanchang University, 999 Xuefu Avenue, Nanchang 330031, China ∥ School of Chemistry and Chemical Engineering, Shanghai Jiaotong University, 800 Dongchuan Road, Shanghai 200240, China

Downloaded by BOSTON UNIV at 18:42:04:128 on May 31, 2019 from https://pubs.acs.org/doi/10.1021/acs.macromol.9b00411.

S Supporting Information *

ABSTRACT: From the miscibility viewpoint, the mechanism of how the third polymer influences the morphology evolution of an active layer composed of a polymer donor and a polymer acceptor remains to be unknown. In this article, J71 was incorporated into the PBDB-T:PNDI-2T-TR(5) system to afford a ternary all-polymer solar cell showing the maximal power conversion efficiency of 9.12%. It is observed that J71 could provide complementary absorption in addition to the formation of cascade energy-level alignment with other polymers, demonstrating a Förster resonance energy transfer from J71 to PBDB-T. By the combination of dynamic mechanical analysis and differential scanning calorimetry techniques, the miscibility in the J71:PNDI-2T-TR(5) blend is better than that in the PBDB-T:PNDI-2T-TR(5) blend according to Fox equation analysis, showing Flory−Huggins interaction parameters of −1.003 and 0.006, respectively. J71 could serve as a compatibilizer to optimize the morphology of the active layer, leading to enrichment of PNDI-2T-TR(5) on the film surface due to the much lower surface energy of PNDI-2T-TR(5). The optimized vertical distribution of PNDI-2T-TR(5) with enhanced crystallization could be further demonstrated by grazing incidence wideangle X-ray scattering and atomic force microscopy, facilitating electron transfer and collection in the normal device structure for improved photovoltaic performance.

1. INTRODUCTION Organic solar cells have attracted considerable attention due to their unique advantages in fabricating lightweight and flexible devices suitable for roll-to-roll technology, exhibiting the highest power conversion efficiency (PCE) value of 17.3%.1−8 In addition, all-polymer solar cells (all-PSCs), comprising a polymer donor and an acceptor, exhibit outstanding thermal and mechanical stability, as well as facile modification for optimized optoelectronic properties suitable for long-term operation in flexible and large-scale devices.9−16 Through precise chemical structure design, a large number of novel polymer acceptors such as ITIC-like polymers and perylenediimide (PDI)- and naphthalenediimide (NDI)-based polymers have been synthesized, resulting in an inspiring breakthrough in the PCE value from 3 to 11%.17−21 Recently, the novel polymer acceptor poly[[N,N′-bis(2octyldodecyl)-naphthalene-1,4,5,8-bis(dicarboximide)-2,6diyl]-alt-5,5′-(2,2′-bithiophene)] named as N2200 has been developed, exhibiting excellent absorption behavior even in the near-infrared region, good solubility in organic solvents, and high electron mobility.22 Unfortunately, N2200 also suffers from its own disadvantages such as low absorption coefficient, limiting the increase in photocurrent and efficiency, and high crystallinity, leading to serious phase separation with unbalanced hole and electron transport after blending with donor polymers.23−25 To solve the inherent drawbacks of © XXXX American Chemical Society

N2200, some research works focused on the development of novel p-type polymers such as PTzBI and PTzBI-Si, showing complementary absorption and good miscibility with N2200.26,27 An alternative route is the modification of N2200 via side chain and backbone engineering. By the introduction of a dye group 2-(1,1′-dicyanomethylene)-4-(3thienylmethylene)rhodanine (TR) via random copolymerization into the backbone of N2200, we successfully obtained the novel acceptor PNDI-2T-TR(5), exhibiting a much improved PCE of 8.13% after blending with PBDB-T due to higher absorption coefficient and reduced crystallization.28 To overcome the limitations of N2200-based all-PSCs, the solvent annealing and solvent additive dibenzyl ether techniques are utilized that could further enhance PCE values of solar cells.21 Another promising strategy is the ternary blending technique, and the third additive could be fullerenes or polymer donors. It should be noted that the maximal PCE value of 10% was achieved by the incorporation of PBTA-BO into PTzBI:N2200.29 Through a careful analysis of the chemical structure in the all-polymer ternary blends, it is found that both the third additive and the donor polymer have similar backbones and side chains. Although the authors Received: February 28, 2019 Revised: May 18, 2019

A

DOI: 10.1021/acs.macromol.9b00411 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

distribution with a higher PCE, which has not been reported previously.

believed that good miscibility in donor and acceptor blends is vital to improve the fill factor (FF) in all-PSCs, no direct confirmations about miscibility could be provided despite the grazing incidence wide-angle X-ray scattering (GIWAXS) result.30 In these years, the miscibility between the donor and acceptor in fullerene and nonfullerene systems has attracted much attention.31−35 The quantitative description of miscibility between two components can be described by Flory− Huggins interaction parameter χ. In general, the χ value below 0.5 illustrates that the two components should be miscible, whereas the χ value above 1.5 shows that the two components are immiscible.36−40 There are several techniques used to determine the Flory−Huggins interaction parameters, including the synchrotron radiation-based scanning transmission Xray microscopy, secondary-ion mass spectrometry, and so on.32,41 Unfortunately, these techniques require an advanced national laboratory platform, which is complex and difficult to perform. The melting-point depression method measured by differential scanning calorimetry (DSC) and ultraviolet−visible (UV−vis) absorption spectroscopy are demonstrated to be the simpler techniques suitable for common labs, although they have their own drawbacks from both theoretical and experimental viewpoints.42,43 The alternative dynamic mechanical analysis (DMA) has also been developed to determine the glass transition behavior of the active layer of PSCs, especially for the amorphous, conjugated polymer-based systems.44 Through analysis of the change of glass transition temperature (Tg) by the Fox equation, we are able to predict the miscibility between two components.45 Therefore, the incorporation of third polymer into binary systems to fabricate ternary all-PSCs is recognized to be effective to further improve PCE values. However, how to select the appropriate third polymer from the viewpoint of miscibility remains to be unknown. The role of miscibility between the third polymer and the other polymers affecting the morphological evolution of the active layer and eventual performance of PSCs needs deep investigation from the aspect of thermodynamics. In addition, how the miscibility influences the vertical distribution of donors and acceptors after incorporation of the third polymer remains to be unknown. In this article, the J71 polymer was chosen as the third additive for doping into the PBDB-T:PNDI-2T-TR(5) system, where the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) energy levels of J71 locate between the corresponding energy levels of PBDB-T and PNDI-2T-TR(5) to form a cascade energy-level alignment. The incorporation of J71 could afford a maximal efficiency of 9.12%, in addition to FF of 71%, Jsc of 14.6 mA/ cm2, and Voc of 0.88 V, which is one of the highest Voc value in all-PSCs. The enhanced PCE mainly contributes to optimization of phase separation in PBDB-T:PNDI-2T-TR(5) by J71. The χ values in PBDB-T:PNDI-2T-TR(5) and J71:PNDI2T-TR(5) blends are calculated to be 0.0061 and −1.0034, respectively, demonstrating that the miscibility between J71 and PNDI-2T-TR(5) is better than that between PBDB-T and PNDI-2T-TR(5). Moreover, the DMA technique was applied to investigate the miscibility between different components. Through Fox equation analysis, J71 and PBDB-T are suggested to be miscible. J71 could serve as a compatibilizer to optimize the morphology in PBDB-T:PNDI-2T-TR(5), leading to enrichment of PNDI-2T-TR(5) on the active layer surface due to much lower surface energy for optimized vertical

2. EXPERIMENTAL SECTION 2.1. Materials. Polymer donors of PBDB-T and J71 were supplied by Organtecsolar Materials, Inc. Chlorobenzene and 1,8-diiodooctane were provided by Sigma-Aldrich. Poly(3,4ethylenedioxythiophene):poly(styrene-sulfonate) (PEDOT:PSS) and perylenediimide functionalized with amino N-oxide (PDINO) were supplied by Solarmer Materials, Inc. The PNDI-2T-TR(5) polymer was synthesized as described previously.28 2.2. Device Fabrication. The fabrication procedure of solar cells with a normal device structure follows the sequence. The indium-tin oxide (ITO) glasses were cleaned by acetone, deionized water, and isopropyl alcohol using ultrasonic treatment for 30 min, respectively. Then, the ITO glasses were exposed to an ultraviolet lamp with a wavelength of 254 nm for 15 min. After that, the poly(3,4ethylenedioxythiophene):poly(styrene-sulfonate) (PEDOT:PSS) was spin-coated at 4000 rpm for 1 min. Then, ITO glasses covered by the PEDOT:PSS layer were heated at 150 °C for 15 min. The films were spun at 2700 rpm for 1 min from the PBDB-T:J71:PNDI-2T-TR(5) solution in chlorobenzene. The weight ratio of the donor and acceptor (D/A) was maintained at 1.5:1, and the concentration is kept at 13 mg/mL with 1,8-diiodooxane (DIO) volume fraction of 0.5 vol %. Then, the perylenediimide functionalized with amino N-oxide (PDINO) solution in methanol was spin-coated at 3000 rpm for 30 s. Al was finally deposited by thermal evaporation to afford an electrode with a thickness of about 100 nm. 2.3. Characterization. The current density−voltage (J−V) curves were determined by Keithley 2400, with currents measured by an Abet solar simulator Sun2000 under 100 mW/cm2 simulated AM 1.5G irradiation. These characterizations were carried out under nitrogen at 25 °C. The UV−vis spectra were characterized on a PerkinElmer Lambda 750 spectrophotometer. The polymer solutions in chlorobenzene were spin-coated on a quartz plate at 2700 rpm. The steady-state photoluminescence (PL) spectra were performed on a Shamrock sr-303i-B spectrograph from Andor Tech using a Xe flash lamp. The specimens used for PL measurements were the same as UV−vis characterization. DSC was characterized by TA DSC Q2000. The specimens used for DSC analysis were prepared by the solution casting method. The dried samples (2−5 mg) were heated to 350 °C and subsequently cooled to 20 °C at a rate of 10 °C/min under nitrogen protection. The corresponding parameters of melting temperature and enthalpy and crystallization temperature and enthalpy were determined. DMA was measured on a PerkinElmer Diamond DMA. The chlorobenzene solvent was used to dissolve the polymers to afford 30 mg/mL solutions at 70 °C for 12 h. Then, the solutions at the temperature of about 50 °C were drop-coated on a glass fiber mesh of 30 mm length and dried in a glove box under nitrogen for one day. The measurements were carried out in the tensile mode at a frequency of 1 Hz under nitrogen protection with a flow of 60 mL/min. The specimens were heated from −110 to 280 °C, and the heating rate was set to 3 °C/min. The variation of loss angle (tan δ), elastic modulus (E′), and loss modulus (E″) versus temperature was measured, and the glass transition temperatures were determined. Atomic force microscopy (AFM) measurements were characterized by a Digital Instrumental Nanoscope 31 operating in the tapping mode. The fabrication of the specimens used for AFM observation is similar to device preparation without the electron transporting layer and the Al electrode. The transmission electron microscopy (TEM) observation was performed on a JEM-2100F TEM. The films used for TEM characterization were the same as for AFM. Then, the specimens were soaked in H2O for 10 s, and the copper mesh grid was used to collect the film. The water contact angle of the films was measured on a Krüss DSA100s drop shape analyzer, and the surface energy was calculated according to the software. GIWAXS were carried out at beamline 7.3.3 at an Advanced Light Source, Berkeley, with an incident angle of 0.14°. The scattered X-rays were detected by a Dectris Pilatus 2 M photon counting detector. The B

DOI: 10.1021/acs.macromol.9b00411 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 1. (a) Chemical structures, (b) normalized UV−vis absorption spectra, and (c) energy-level alignments of PBDB-T, J71, and PNDI-2TTR(5).

Figure 2. (a) J−V characteristics and (b) EQE spectra of PSCs containing different contents of J71.

solar cell system.28 Unfortunately, the PSCs of PBDB-T:PNDI2T-TR(5) still exhibit a relatively low FF and PCE, which could be further enhanced by the incorporation of the third additive. Inspired by the similar chemical structure of J71 with PBDB-T containing a BDT core in the main chain of polymers, the J71 might be an appropriate third additive from the aspect of miscibility of “like dissolves like”. The chemical structures of PBDB-T, J71, and PNDI-2T-TR(5) are illustrated in Figure 1a. The corresponding proton NMR (1H NMR) spectra of three polymers are illustrated in Figure S1, and the number-average and weight-average molecular weight, in addition to polydispersity, are shown in Table S1 and Figure S2. It is well known that the molecular weights and dispersities showed the obvious effect on active layer morphology. For example, the domain size of the all-polymer-based PPDT2FBT:P(NDI2OD-T2) blend film dramatically reduced with the

specimens used for GIWAXS analysis was fabricated by spin-coating one layer of PEDOT:PSS onto a silicon wafer in the dimension of 1.5 cm × 1.5 cm, followed by spin-coating of PBDB-T:J71:PNDI-2TTR(5) solution in chlorobenzene with the concentration of 13 mg/ mL at 2700 rpm for 1 min. The resonant soft X-ray scattering (RSoXS) was carried out at beamline 11.0.1.2 at the Advanced Light Source, Berkeley. The specimens were fabricated similar to those using for TEM observation. The floating films in deionized water were transferred to a 1.5 mm × 1.5 mm, 100 nm-thick Si3N4 membrane supported by a 5 mm × 5 mm, 200 μm-thick Si frame (Norcada, Inc.). The poly(isoprene-b-styrene-b-2-vinyl pyridine) triblock copolymer with a spacing of 391 Å was used to calibrate the distance between the sample and the detector.

3. RESULTS AND DISCUSSION In our previous study, the acceptor PNDI-2T-TR(5) has been proven to be a better photovoltaic material than N2200 when it is blended with the PBDB-T donor to constitute a binary C

DOI: 10.1021/acs.macromol.9b00411 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

with the PCE result. It should be pointed out that J71:PNDI2T-TR(5) shows a much lower EQE value from 580 to 710 nm, as compared to that of PBDB-T:PNDI-2T-TR(5) and its ternary systems. Although the relative PBDB-T content decreases in ternary blends, the EQE at ∼625 nm was enhanced with less PBDB-T, which is probably contributing to the optimized active layer morphology induced by J71, showing an obvious increase of Jsc, FF, and PCE. The absorption and excitation behaviors of the films were investigated to explore the energy transfer between different components. Figure 3a shows normalized UV−vis absorption spectra of binary and ternary blend films. The binary PBDBT:PNDI-2T-TR(5) film shows double absorption peaks at 389 and 628 nm. After incorporation of J71, the absorption peak at about 628 nm reaches the maximum at 10 wt % J71. The appearance of a new peak at 589 nm demonstrates the improved absorption in the presence of J71. The PL spectra were further determined to study charge transfer in binary and ternary films, as illustrated in Figure 3b. The neat PBDB-T film exhibits a photoluminescence peak at 685 nm when excited at 570 nm, whereas the pristine J71 film shows a peak at 642 nm when excited at 540 nm. By mixing with PNDI-2T-TR(5), the corresponding photoluminescence peaks are effectively quenched, revealing efficient charge transfer between donors and acceptors. As shown in Figure 3c, the PBDB-T:J71:PNDI2T-TR(5) ternary film containing 10 wt % J71 with a DIO solvent additive exhibits the lowest PL intensity, illustrating the most efficient charge transfer in this sample, in association with the maximal J sc value in the corresponding device. Furthermore, the emission band of J71 overlaps with the absorption band of PBDB-T as shown in Figure 3d, demonstrating that a Förster resonance energy transfer from J71 to PBDB-T may occur. As is known, the energy transfer between two materials could be verified by testing the photoluminescence emission spectra of their pristines and mixed films. J71 exhibits photoluminescence at 649 nm when excited at 570 nm, which is seriously quenched by PBDB-T. Moreover, the PL intensity at 685 nm of PBDB-T gradually increases as the J71 content increases, illustrating efficient energy transfer from J71 to PBDB-T (Figure 3e), which is consistent with the EQE results. To confirm whether an additional charge transfer occurred between PBDB-T and J71, the device based on PBDB-T:J71 (1:1, wt/wt) was fabricated. As observed from the J−V characteristics of the PBDB-T:J71 system, the Jsc is 0.81 mA/cm2 in Figure 3f, revealing that almost Förster resonance energy transfer occurs between PBDB-T and J71. The energy transfer between PBDB-T and J71 implies that they should be miscible. Unfortunately, how miscible are J71:PNDI-2T-TR(5) and PBDB-T:PNDI-2T-TR(5) mixtures is not known. In binary polymer blends, the change of glass transition temperature upon blending ratio was always used to predict the miscibility between two components. If the blend exhibits double glass transitions with temperature close to the respective single component Tg value, the two components in the blends are thus defined to be immiscible. If the blends show only one glass transition with the temperature between two single polymers, which could be described by the Fox equation, the two components are supposed to be completely miscible. In this article, we performed the widely used thermal analysis techniques including DMA and DSC to investigate how miscible are different components.42,44 As shown in Figure 4, the tan δ versus temperature curves illustrates the

increasing polymer donor number-average molecular weight.16,46 As illustrated in Figure 1b, PBDB-T exhibits an absorption peak and a shoulder peak at 624 and 589 nm, respectively. PNDI-2T-TR(5) shows matched absorption with PBDB-T, having two strong absorption peaks at 394 and 704 nm, respectively. The pristine J71 exhibits two absorption peaks at 542 and 583 nm, which are complementary to the absorption gap of PBDB-T and PNDI-2T-TR(5). In addition, J71 could form cascade energy-level alignment with PBDB-T and PNDI2T-TR(5), facilitating charge separation and transport as illustrated in Figure 1c. Furthermore, the HOMO of J71 is lower than that of PBDB-T, which could enhance the Voc of PBDB-T- and PNDI-2T-TR(5)-based solar cells. The HOMO−HOMO offset and LUMO−LUMO offset are 0.58 and 1.01 eV for the PBDB-T and PNDI-2T-TR(5) pair and 0.51 and 0.69 eV for the J71 and PNDI-2T-TR(5) pair, respectively. The offset is large enough to drive efficient exciton dissociation due to the requirement of a minimum offset larger than 0.3 eV. Then, the influence of J71 on PBDB-T:PNDI-2T-TR(5) performance was investigated using a normal structure of ITO/ PEDOT:PSS/active layer/PDINO/Al.47 The J−V curves of all-PSCs are shown in Figures 2a and S3; meanwhile Tables 1 Table 1. Photovoltaic Parameters of the Optimized PBDBT:PNDI-2T-TR(5) PSCs with Different Contents of J71 under Simulated Solar Illumination (AM 1.5G, 100 mW/ cm2) J71 content (wt %)

Voc (V)

Jsc (mA/cm2)

0 5 10 20 100

0.85 0.87 0.88 0.89 0.92

13.58 14.38 14.63 14.10 11.03

(12.87)a (13.82) (14.30) (13.42) (10.58)

FF (%)

PCEave (%)

PCEmax (%)

65.45 68.00 71.02 67.70 67.20

7.51 8.43 9.05 8.33 6.67

7.59 8.50 9.12 8.47 6.80

a

The photocurrent is calculated by integrating the EQE spectra with the AM 1.5G solar spectrum.

and S2 illustrate the corresponding parameters. PBDBT:PNDI-2T-TR(5) exhibits a PCE of 7.59% with a Jsc of 13.58 mA/cm2, FF of 65.45%, and Voc of 0.85 V. In contrast, the J71:PNDI-2T-TR(5)-based binary PSCs show a lower PCE of 6.80% due to a smaller Jsc of 11.03 mA/cm2, despite a higher FF of 67.20% and a Voc of 0.92 V. After incorporation of J71 as the third additive into the PBDB-T:PNDI-2T-TR(5) system, the PCE values of specimens show an obvious enhancement. It is worth noting that the specimen at the J71 content of 10 wt % attains the highest PCE value of 9.12%, with a Jsc of 14.63 mA/cm2, FF of 71.02%, and Voc of 0.88 V. The total D/A weight ratios are maintained at 1.5:1, and the solutions were spun at 2700 rpm for 1 min to afford the active layers. The device with the highest PCE value is based on PBDB-T:J71:PNDI-2T-TR(5) with a weight ratio of 135:15:100. J71 is suggested to be an effective additive to enhance the performance of the binary PBDB-T:PNDI-2TTR(5) system. The external quantum efficiency (EQE) spectra of the binary and ternary PSCs are illustrated in Figure 2b. The integrating current density provided by EQE curves is illustrated in Table 1. The EQE value ranging from 460 to 689 nm increases after the incorporation of J71, reaching the highest value at the J71 content of 10 wt %, which is consistent D

DOI: 10.1021/acs.macromol.9b00411 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 3. (a) Normalized absorption of the active layer with different contents of J71, (b) PL spectra of the films of PBDB-T and PBDB-T:PNDI2T-TR(5) excited at 570 nm and of J71 and J71:PNDI-2T-TR(5) excited at 540 nm, (c) PL spectra of PBDB-T and its binary and ternary blend films processed with or without DIO, (d) normalized absorption of PBDB-T and PL spectra of J71 excited at 570 nm, (e) PL spectra of PBDBT:J71 blend films excited at 570 nm, and (f) J−V characteristics of the device based on PBDB-T:J71 (1:1, wt:wt).

w w 1 = 1 + 2 Tg Tg1 Tg2

glass transition of pristine polymers and binary blends, and the elastic modulus versus temperature curves are shown in Figure S4. We define the major relaxation peak as glass transition, and the minor relaxation peaks are named as secondary relaxations or subglass transitions. The relating parameters are shown in Table 2. It is noted that pristine PBDB-T shows a major relaxation peak at 51.6 °C, and J71 exhibits a major relaxation peak at 61.6 °C with a secondary relaxation peak at 161.5 °C. Notably, the PNDI-2T-TR(5) shows the major glass transition at −19.9 °C and secondary relaxation peaks at 57.0 and 124.4 °C, respectively. The relatively low Tg at −19.9 °C might correspond to the relaxation of the NDI segment containing alkyl chains. After blending, the PBDB-T:J71 binary blends exhibit a Tg value between that of pristines PBDB-T and J71. We explored the miscibility between PBDB-T:J71 by analyzing the Tg values through the Fox equation45 as shown below

(1)

where Tg is the glass transition temperature of the mixing blend, Tg1 and Tg2 are the glass transition temperatures of PBDB-T and J71, respectively, w1 and w2 represent the weight fractions of PBDB-T and J71 in blends, respectively. It is observed from Figure 5a that the experimental data are very close to the calculated data, demonstrating that the two components PBDB-T and J71 are miscible. In contrast, the J71:PNDI-2T-TR(5) blend shows a relatively broad glass transition with temperature in the range between that of pristine J71 and PNDI-2T-TR(5), which could also be fitted with Fox equation analysis, illustrating that J71 and PNDI-2TTR(5) should be miscible with each other. In addition, the PBDB-T:PNDI-2T-TR(5) blend shows a Tg of 57.9 °C, which is close to the secondary relaxation temperature at 57.0 °C of E

DOI: 10.1021/acs.macromol.9b00411 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 4. DMA temperature scans of (a) pristines PBDB-T, J71, and PNDI-2T-TR(5); (b) PBDB-T:J71; (c) J71:PNDI-2T-TR(5) and PBDBT:PNDI-2T-TR(5); and (d) PBDB-T:J71:PNDI-2T-TR(5) blends.

0 °C should correspond to the NDI segment relaxation. In the PBDB-T:PNDI-2T-TR(5) blend, the relaxation of the NDI segment might be restricted in the presence of PBDB-T due to π−π interaction. By the incorporation of J71, the J71 molecular chains might serve as a compatibilizer to improve the miscibility between PBDB-T and PNDI-2T-TR(5), where J71 is supposed to locate at the interfaces between donor and acceptor phases. Due to good miscibility between J71 and PBDB-T, PNDI-2T-TR(5) might aggregate to form a new phase. Thus, the relaxation of the NDI segment becomes free, leading to the appearance of a tan δ peak at the lowtemperature region. As the J71 content further increases to 20 wt %, the excessive J71 molecules tend to mix well with PBDBT due to good miscibility between the two components. PNDI-2T-TR(5) is supposed to form a separated phase due to the appearance of relaxation at a relatively low-temperature region, which might be related to the vertical phase separation of PNDI-2T-TR(5) from the mixed phase. Based on DMA analysis, PBDB-T and J71 are miscible, whereas PBDB-T and PNDI-2T-TR(5) are less miscible than PNDI-2T-TR(5) and J71. As discussed, the DMA technique is unable to quantitatively describe the miscibility between different components. Then, the DSC technique was performed to investigate the Flory−Huggins interaction parameter (χ) by the melting-point depression method. Pristines PBDB-T and J71 show no crystallization or melting transition on cooling and heating curves. Therefore, the melting temperature and melting enthalpy of PBDB-T:PNDI2T-TR(5) and J71:PNDI-2T-TR(5) were determined, as shown in Figures 5b, S5, and S6 and Table S3. Parameter χ is calculated by the following equation42

Table 2. Summary of the Sub-Tg and Tg of Pristine Polymers and Binary and Ternary Blends samples PBDB-T PBDB-T:J71

J71 PNDI-2T-TR(5) PBDB-T:PNDI-2T-TR(5) PBDB-T:J71:PNDI-2T-TR(5)

J71:PNDI-2T-TR(5)

weight ratio (wt/wt) 95:5 90:10 80:20 50:50

150:100 142.5:7.5:100 135:15:100 120:30:100 150:100

Tg (°C) 51.6 52.5 53.4 55.2 57.0 61.6 −19.8 57.9 57.4 33.5 27.6 41.2

sub-Tg (°C)

122.6 144.0 161.5 57.0/124.4 134.3 143.0 50.7/155.2 168.0

PNDI-2T-TR(5), demonstrating that PBDB-T might be less miscible than J71 with PNDI-2T-TR(5). In ternary blends, the shape of the relaxation peak of PBDBT:PNDI-2T-TR(5) is significantly dependent on the J71 content. The PBDB-T:J71:PNDI-2T-TR(5) (142.5:7.5:100) specimen exhibits a similar relaxation peak to that of PBDBT:PNDI-2T-TR(5), illustrating that the phase structure of PBDB-T:PNDI-2T-TR(5) is not obviously influenced at the low J71 content. As the J71 content increases, the intensity of the relaxation peak at the temperature region from −50 to 0 °C becomes stronger. In addition, an obvious peak at 50.7 °C could be discerned and the shape of the relaxation curve at the high temperature region is similar to that of J71. It is suggested that the relaxation at the low-temperature region from −50 to F

DOI: 10.1021/acs.macromol.9b00411 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 5. (a) 1/Tg versus J71 content in PBDB-T:J71 blends; the red dots represent the data calculated via the Fox equation and the black squares are experimental data. (b) DSC cooling and heating curves of PNDI-2T-TR(5) and its blends with PBDB-T and J71.

ÄÅ ij 1 1 1 R ν2 ÅÅÅÅ ln ϕ2 1 yzz − 0 =− + jjj − z ÅÅ jm Tm m1 zz{ ΔHf ν1 ÅÅÅÇ m2 Tm k 2 yz × (1 − ϕ2zzzz + χ (1 − ϕ2)2 ] {

Figure 6. Water contact angles of the pristine polymers and binary and ternary blend films.

mentioned about the importance of miscibility in optimizing the active layer morphology, the role of the third polymer working as the compatibilizer has been reported less previously. AFM and TEM were further performed to explore changes in film morphology (Figures S8−S10). According to AFM images, no obvious differences of surface morphology could be discerned, illustrating that the polymers could be well mixed in tens of nanometer scale. It is noted that the root-mean-square (RMS) roughness of the PBDB-T:PNDI-2T-TR(5) binary blend film is determined to be 2.23 nm, which first decreases and then increases with the J71 content increasing from 5 to 20 wt %. The RMS values of the ternary blend film are 2.19, 1.91, 2.12 nm, respectively. The film showing the best photovoltaic performance exhibits an RMS value of 1.91 nm at the J71 content of 10 wt %. In contrast, the J71:PNDI-2T-TR(5) specimen exhibits a roughness of 1.34 nm, which is much smaller than that of the other binary and ternary blend films. Similarly, no discernible phase structure is observed in TEM graphs, despite the more obvious phase separation in the PBDB-T:PNDI-2T-TR(5) film than that in J71:PNDI-2TTR(5). According to the above analysis for miscibility, the difference in RMS might be originating from two aspects, including the change in crystallization upon incorporation of J71 into PBDB-T:PNDI-2T-TR(5) and vertical phase separation induced by miscibility. J71 exhibits better miscibility with PNDI-2T-TR(5) than that of PBDB-T with PNDI-2TTR(5), leading to the stronger restriction in crystallization of PNDI-2T-TR(5) by J71, which is consistent to the smaller

(2)

where subscript 1 is identified with a weakly crystalline polymer of PBDB-T or J71 and 2 with a stronger crystalline polymer of PNDI-2T-TR(5). Tm and Tm0 are the melting temperatures 0, respectively polymer blends and PNDI-2TTR(5), respectively. ΔHf is the enthalpy of crystalline PNDI2T-TR(5), and R is ideal gas constant. Parameters m, ν, and ϕ are the degree of polymerization, molar volume, and volume fraction, respectively. To neglect the entropy and ϕ2 effects, the value of χ could be obtained from the following equation χ=

Bν1 RT

(3)

where B represents the interaction energy density of the polymer pair. Combining the above equations, the value of χ could be calculated. In PBDB-T:PNDI-2T-TR(5) (135:100) and J71:PNDI-2T-TR(5) (15:100) blends, the χ values are 0.0061 and −1.0034, respectively. It is indicated that J71 exhibits better miscibility with PNDI-2T-TR(5) than that with PBDB-T, indicating that J71 could serve as the interfacial compatibilizer to slightly improve the miscibility between PBDB-T and PNDI-2T-TR(5). Although most researchers G

DOI: 10.1021/acs.macromol.9b00411 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 7. Two-dimensional (2D) GIWAXS patterns of (a) PBDB-T:PNDI-2T-TR(5), (b) PBDB-T:10% J71:PNDI-2T-TR(5), (c) J71:PNDI-2TTR(5), (d) PNDI-2T-TR(5), and (e) PBDB-T. (f) Proposed models through doping J71 into PBDB-T:PNDI-2T-TR(5) for optimized vertical distribution. (g) Corresponding in-plane and out-of-plane line cuts from the GIWAXS patterns.

DMA analysis, due to vertical phase separation of PNDI-2TTR(5) from the PBDB-T:J71 mixed phase. Additionally, the migration of PNDI-2T-TR(5) onto the film surface is beneficial to electron transport in the normal structure device, attributing to the improvement of Jsc, FF, and PCE values. Unfortunately, too much aggregation of PNDI-2T-TR(5) on the film surface eventually reduces the efficiency of exciton separation and performance of solar cells. Then, the crystalline structure of binary and ternary blend films was characterized by GIWAXS, and the corresponding two-dimensional (2D) patterns and one-dimensional curves are shown in Figure 7. In the out-of-plane curve, the pristine PBDB-T exhibits diffraction peaks at 0.27, 0.58, and 0.82 Å−1, ascribing to (100), (200), and (300) lattice planes, respectively. The diffraction peak at 1.52 Å−1 attributed to the (010) lattice plane illustrates both edge-on and face-on orientations. According to the previous literature, J71 has (100) and (010) peaks, showing a preference for the face-on orientation.48PNDI-2T-TR(5) shows obvious diffraction peaks at 0.42 and 1.42 Å−1 in the in-plane curve, revealing the preferential face-on orientation. After blending with PBDB-T or J71, the intensity of the peak at 0.42 Å−1 significantly reduced, revealing good miscibility between the polymers. Especially, the peak intensity in J71:PNDI-2T-TR(5) is obviously lower than that of PNDI-2T-TR(5). It is suggested that the miscibility between J71 and PNDI-2T-TR(5) is better than that between PBDB-T and PNDI-2T-TR(5), which is consistent to the previous Flory−Huggins interaction parameter and DMA analyses. After incorporation of 10 wt % J71 into the PBDB-T:PNDI-2T-TR(5) blend, the (010) peak in

RMS value in J71:PNDI-2T-TR(5) in contrast to that in PBDB-T:PNDI-2T-TR(5). Upon incorporation of J71 into PBDB-T:PNDI-2T-TR(5), the J71 tends to disperse in the interfaces between PBDB-T and PNDI-2T-TR(5), as discussed in the DMA analysis, restricting crystallization of PNDI2T-TR(5) showing the reduced roughness. As J71 weight ratio reaches 20 wt %, the excessive J71 tends to form a miscible phase with PBDB-T, driving off the PNDI-2T-TR(5) from the PBDB-T and PNDI-2T-TR(5) mixed phase. The enrichment of PNDI-2T-TR(5) on the film surface facilitates the crystallization of PNDI-2T-TR(5), showing an increased RMS value, which is probably due to the relatively low surface energy of PNDI-2T-TR(5). The water contact angles of pristine films and the binary and ternary blend films is shown in Figure 6, in addition to the surface energy data illustrated in Table S4. It is observed that the surface energy of PNDI-2T-TR(5) is lower than that of PBDB-T and J71, suggesting that PNDI-2T-TR(5) tends to enrich on the active layer surface, whereas J71 facilitates to penetrate into the inner region of films. In the PBDB-T:PNDI2T-TR(5) binary blend film, the surface energy of the film is 21.50 mN/m, which is close to that of pristine PBDB-T. Similarly, the J71:PNDI-2T-TR(5) film exhibits a surface energy of 21.26 mN/m, which is between that of J71 and PNDI-2T-TR(5). Upon incorporation of J71 into PBDBT:PNDI-2T-TR(5), the surface energy of the ternary blend film continuously decreases, suggesting that more and more PNDI-2T-TR(5) enriches on the film surface. Therefore, the enrichment of PNDI-2T-TR(5) on the film surface leads to the appearance of relaxation at low temperature, as revealed in the H

DOI: 10.1021/acs.macromol.9b00411 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules the out-of-plane curve and the diffraction peak at 0.78 Å−1 in the in-plane curve ascribing to PNDI-2T-TR(5) become sharper and stronger, demonstrating that the crystallization of PNDI-2T-TR(5) enhances. The result could further support the AFM and DMA speculation that J71 facilitates to mix well with PBDB-T to form a miscible phase, leading to enrichment of PNDI-2T-TR(5) on the film surface due to its low surface energy, as shown in Figure 7f. Figure S11 exhibits the RSoXS profiles and TEM graphs of PBDB-T:PNDI-2T-TR(5) and J71:PNDI-2T-TR(5) blend films. The phase separation dimension of J71:PNDI-2T-TR(5) is smaller than that of PBDB-T:PNDI-2T-TR(5), as illustrated in Figure S8. Therefore, based on the above analysis, the miscibility in different polymer blends should follow the sequence PBDBT:J71>J71:PNDI-2T-TR(5)>PBDB-T:PNDI-2T-TR(5). Due to the similar surface energies of PBDB-T and PNDI-2TTR(5), vertical phase separation may not happen in this blend. After the incorporation of a small amount of J71 as the third additive into PBDB-T:PNDI-2T-TR(5), the J71 tends to serve as a compatibilizer to optimize morphology in the PBDBT:PNDI-2T-TR(5) film, reducing the phase domain size for more efficient all-PSCs. As the J71 content increases, the amount of J71 staying at the interfacial region becomes saturated and the excessive J71 tends to penetrate into PBDBT to form the miscible phase. Due to the discernible difference in surface energies of J71 and PNDI-2T-TR(5), the PNDI-2TTR(5) tends to enrich on film surface during the solvent evaporation process, leading to optimized vertical distribution of the donor and acceptor for obviously improved performance. Despite absorption and energy-level alignment, the miscibility between the third additive and the donor (or acceptor) influencing on the active layer morphology especially of the vertical distribution should be urgently taken into consideration to guide the selection of an appropriate third additive for organic solar cells showing high performance. Due to the lack of related research in the active layer morphology from the viewpoint of polymer physics, the exploration of the mechanism concerning the morphological evolution in both horizontal and vertical directions could provide useful guidelines to effectively control the phase structure as well as the corresponding photovoltaic performance in multicomponent organic solar cells.

PBDB-T:PNDI-2T-TR(5). As the J71 content increases, the excessive J71 tended to mix well with the PBDB-T phase and PNDI-2T-TR(5) facilitated to migrate to the film surface due to the lower surface energy. The well-optimized morphology eventually led to the maximal performance of all-PSCs, which provide useful guidelines to select the appropriate third polymer from the viewpoint of miscibility, despite the consideration of absorption and energy-level alignment.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.macromol.9b00411. 1



H NMR spectra; gel permeation chromatography curves; J−V curves; storage modulus versus temperature curves; DSC heating and cooling curves; AFM and TEM images; M n , M w , and PDI values; photovoltaic parameters; surface energy data; and RSoXS profiles (PDF)

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected] (W.Z.). *E-mail: [email protected] (F.L.). *E-mail: [email protected] (Y.C.). ORCID

Yiwang Chen: 0000-0003-4709-7623 Author Contributions ⊥

S.L. and D.C. contributed equally to this work.

Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS W.Z. thanks the support from the National Natural Science Foundation of China (NSFC) (21764009 and 51563016). Y.C. thanks the support from the National Natural Science Foundation of China (NSFC) (51833004) and the National Science Fund for Distinguished Young Scholars (51425304). X-ray data was acquired at beamlines 7.3.3 and 11.0.1.2 at the Advanced Light Source, Berkley.



4. CONCLUSIONS In summary, we successfully fabricated high-performance ternary all-PSCs by adding J71 into the PBDB-T:PNDI-2TTR(5) system, affording a maximal PCE value of 9.12%. The absorption of J71 is complementary to the absorption gap of PBDB-T and PNDI-2T-TR(5), showing a cascade energy-level alignment and Förster resonance energy transfer between PBDB-T and J71. Based on the glass transition temperature analysis, J71 and PBDB-T are demonstrated to be miscible, which is consistent to the Fox equation description. In addition, J71 and PNDI-2T-TR(5) are also miscible, showing better miscibility than that of the PBDB-T and PNDI-2TTR(5) pair, with Flory−Huggins interaction parameters of −1.0034 and 0.0061, respectively. Due to difference in miscibility, the incorporation of J71 into PBDB-T and PNDI-2T-TR(5) could not only optimize the morphology in the horizontal direction but also induce the vertical distribution of the donor and acceptor as revealed by AFM surface observation and surface energy characterization. J71 served as a compatibilizer to improve the morphology of

REFERENCES

(1) Yu, G.; Gao, J.; Hummelen, J. C.; Wudl, F.; Heeger, A. J. Polymer photovoltaic cells-enhanced efficiencies via a network of internal donor−acceptor heterojunctions. Science 1995, 270, 1789− 1791. (2) Jin, Y.; Chen, Z.; Dong, S.; Zheng, N.; Ying, L.; Jiang, X-F.; Liu, F.; Huang, F.; Cao, Y. A novel naphtho[1,2-c:5,6-c′]bis([1,2,5]Thiadiazole)-based narrow-bandgap π-conjugated polymer with power conversion efficiency over 10%. Adv. Mater. 2016, 23, 9811− 9818. (3) Zhao, W.; Qian, D.; Zhang, S.; Li, S.; Inganäs, O.; Gao, F.; Hou, J. Fullerene-free polymer solar cells with over 11% efficiency and excellent thermal stability. Adv. Mater. 2016, 28, 4734−4739. (4) Liao, X.; Wang, J.; Chen, S.; Chen, L.; Chen, Y. Diketopyrrolopyrrole-based conjugated polymers as additives to optimize morphology for polymer solar cells. Chin. J. Polym. Sci. 2016, 34, 491−504. (5) Dou, L.; Liu, Y.; Hong, Z.; Li, G.; Yang, Y. Low-bandgap near-IR conjugated polymers/molecules for organic electronics. Chem. Rev. 2015, 115, 12633−12665. I

DOI: 10.1021/acs.macromol.9b00411 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules (6) Huo, L.; Liu, T.; Sun, X.; Cai, Y.; Heeger, A.; Sun, Y. Singlejunction organic solar cells based on a novel wide-bandgap polymer with efficiency of 9.7%. Adv. Mater. 2015, 27, 2938−2944. (7) Zhao, J.; Li, Y.; Yang, G.; Jiang, K.; Lin, H.; Ade, H.; Ma, W.; Yan, H. Efficient organic solar cells processed from hydrocarbon solvents. Nat. Energy 2016, 1, No. 15027. (8) Meng, L.; Zhang, Y.; Wan, X.; Li, C.; Zhang, X.; Wang, Y.; Ke, X.; Xiao, Z.; Ding, L.; Xia, R.; Yip, H.; Cao, Y.; Chen, Y. Organic and solution-processed tandem solar cells with 17.3% efficiency. Science 2018, 361, 1094−1098. (9) Lu, L.; Zheng, T.; Wu, Q.; Schneider, A. M.; Zhao, D.; Yu, L. Recent advances in bulk heterojunction polymer solar cells. Chem. Rev. 2015, 115, 12666−12731. (10) Facchetti, A. Polymer donor-polymer acceptor (all-polymer) solar cells. Mater. Today 2013, 16, 123−132. (11) Jung, I.; Lo, W.; Jang, J.; Chen, W.; Zhao, D.; Landry, E.; Lu, L.; Talapin, D.; Yu, L. Synthesis and search for design principles of new electron accepting polymers for all-polymer solar cells. Chem. Mater. 2014, 26, 3450−3459. (12) Sharma, S.; Kolhe, N.; Gupta, V.; Bharti, V.; Sharma, A.; Datt, R.; Chand, S.; Asha, S. Improved all-polymer solar cell performance of n-type naphthalene diimide-bithiophene P(NDI2OD-T2) copolymer by incorporation of perylene diimide as coacceptor. Macromolecules 2016, 49, 8113−8125. (13) Jung, J.; Jo, J.; Chueh, C.; Liu, F.; Jo, W.; Russell, T.; Jen, A. Fluoro-substituted n-type conjugated polymers for additive-free allpolymer bulk heterojunction solar cells with high power conversion efficiency of 6.71%. Adv. Mater. 2015, 27, 3310−3317. (14) Shi, S.; Yuan, J.; Ding, G.; Ford, M.; Lu, K.; Shi, G.; Sun, J.; Ling, X.; Li, Y.; Ma, W. Improved all-polymer solar cell performance by using matched polymer acceptor. Adv. Funct. Mater. 2016, 26, 5669−5678. (15) Liu, S.; Firdaus, Y.; Thomas, S.; Kan, Z.; Cruciani, F.; Lopatin, S.; Bredas, J.; Beaujuge, P. Isoindigo-3,4-difluorothiophene polymer acceptors yield “all-polymer” BHJ solar cells with >7% efficiency. Angew. Chem., Int. Ed. 2018, 8, 540−544. (16) Wang, G.; Melkonyan, F. S.; Facchetti, A.; Marks, T. J. Allpolymer solar cells: Recent progress, challenges, and prospects. Angew. Chem., Int. Ed. 2019, 58, 4129−4142. (17) Cho, H.; Kim, S.; Kim, T.; Sree, V.; Jin, S.; Kim, F.; Kim, B. Design of cyanovinylene-containing polymer acceptors with large dipole moment change for efficient charge generation in highperformance all-polymer solar cells. Adv. Energy Mater. 2017, 8, No. 1701436. (18) Zhang, Z-G.; Yang, Y.; Yao, J.; Xue, L.; Chen, S.; Li, X.; Morrison, W.; Yang, C.; Li, Y. Constructing a strongly absorbing lowbandgap polymer acceptor for high-performance all-polymer solar cells. Angew. Chem., Int. Ed. 2017, 56, 13503−13507. (19) Guo, Y.; Li, Y.; Awartani, O.; Zhao, J.; Han, H.; Ade, H.; Zhao, D.; Yan, H. A vinylene-bridged perylenediimide-based polymeric acceptor enabling efficient all-polymer solar cells processed under ambient conditions. Adv. Mater. 2016, 28, 8483−8489. (20) Guo, Y.; Li, Y.; Awartani, O.; Han, H.; Zhao, J.; Ade, H.; Yan, H.; Zhao, D. Improved performance of all-polymer solar cells enabled by naphthodiperylenetetraimide-based polymer acceptor. Adv. Mater. 2017, 29, No. 1700309. (21) Li, Z.; Ying, L.; Zhu, P.; Zhong, W.; Li, N.; Liu, F.; Huang, F.; Cao, Y. A generic green solvent concept boosting the power conversion efficiency of all-polymer solar cells to 11%. Energy Environ. Sci. 2019, 12, 157−163. (22) Gao, L.; Zhang, Z.; Xue, L.; Min, J.; Zhang, J.; Wei, Z.; Li, Y. All-polymer solar cells based on absorption-complementary polymer donor and acceptor with high power conversion efficiency of 8.27%. Adv. Mater. 2016, 28, 1884−1890. (23) Li, Z.; Xu, X.; Zhang, W.; Meng, X.; Ma, W.; Yartsev, A.; Inganas, O.; Andersson, M.; Janssen, R.; Wang, E. High performance all-polymer solar cells by synergistic effects of fine-tuned crystallinity and solvent annealing. J. Am. Chem. Soc. 2016, 138, 10935−10944.

(24) Li, Z.; Zhang, W.; Xu, X.; Genene, Z.; Rasi, D.; Mammo, W.; Andersson, A.; Janssen, R.; Wang, E.; et al. High-performance and stable all-polymer solar cells using donor and acceptor polymers with complementary absorption. Adv. Energy Mater. 2017, 7, No. 1602722. (25) Liu, X.; Zhang, C.; Duan, C.; Li, M.; Hu, Z.; Wang, J.; Liu, F.; Li, N.; Brabec, C.; Janssen, R.; Bazan, G.; Huang, F.; Cao, Y. Morphology optimization via side chain engineering enables allpolymer solar cells with excellent fill factor and stability. J. Am. Chem. Soc. 2018, 140, 8934−8943. (26) Fan, B.; Ying, L.; Wang, Z.; He, B.; Jiang, X.; Huang, F.; Cao, Y. Optimisation of processing solvent and molecular weight for the production of green-solvent-processed all-polymer solar cells with a power conversion efficiency over 9%. Energy Environ. Sci. 2017, 10, 1243−1251. (27) Fan, B.; Ying, L.; Zhu, P.; Pan, F.; Liu, F.; Chen, J.; Huang, F.; Cao, Y. All-polymer solar cells based on a conjugated polymer containing siloxane-functionalized side chains with efficiency over 10%. Adv. Mater. 2017, 29, No. 1703906. (28) Chen, D.; Yao, J.; Chen, L.; Yin, J.; Lv, R.; Huang, B.; Liu, S.; Zhang, Z.; Yang, C.; Chen, Y.; Li, Y. Dye-incorporated polynaphthalenediimide acceptor for additive free high-performance allpolymer solar cells. Angew. Chem., Int. Ed. 2018, 57, 4580−4584. (29) Li, Z.; Ying, L.; Xie, R.; Zhu, P.; Li, N.; Zhong, W.; Huang, F.; Cao, Y. Designing ternary blend all-polymer solar cells with an efficiency of over 10% and a fill factor of 78%. Nano Energy 2018, 51, 434−441. (30) Fan, B.; Zhu, P.; Xin, J.; Li, N.; Ying, L.; Zhong, W.; Li, Z.; Ma, W.; Huang, F.; Cao, Y. High-performance thick-film all-polymer solar cells created via ternary blending of a novel wide-bandgap electrondonating copolymer. Adv. Energy Mater. 2018, 8, No. 1703085. (31) Watts, B.; Belcher, W. J.; Thomsen, L.; Ade, H.; Dastoor, P. C. A quantitative study of PCBM diffusion during annealing of P3HT:PCBM blend films. Macromolecules 2009, 42, 8392−8397. (32) Collins, B. A.; Gann, E.; Guignard, L.; He, X.; McNeill, C. R.; Ade, H. Molecular miscibility of polymer-fullerene Blends. J. Phys. Chem. Lett. 2010, 1, 3160−3166. (33) Chen, D.; Liu, F.; Wang, C.; Nakahara, A.; Russell, T. P. Bulk heterojunction photovoltaic active layers via bilayer interdiffusion. Nano Lett. 2011, 11, 2071−2078. (34) Treat, N. D.; Varotto, A.; Takacs, C. J.; Batara, N.; Al-Hashimi, M.; Heeney, M. J.; Heeger, A. J.; Wudl, F.; Hawker, C. J.; Chabinyc, M. L. Polymer-fullerene miscibility: a metric for screening new materials for high-performance organic solar cells. J. Am. Chem. Soc. 2012, 134, 15869−15879. (35) Vakhshouri, K.; Kozub, D. R.; Wang, C.; Salleo, A.; Gomez, E. D. Effect of miscibility and percolation on electron transport in amorphous poly(3-Hexylthiophene)/phenyl-C61-butyric acid methyl ester blends. Phys. Rev. Lett. 2012, 108, No. 026601. (36) Kouijzer, S.; Michels, J. J.; Berg, M. van den; Gevaerts, V. S.; Turbiez, M.; Wienk, M. M.; Janssen, René A. J. Predicting morphologies of solution processed polymer: fullerene blends. J. Am. Chem. Soc. 2013, 135, 12057−12067. (37) Li, N.; Perea, J. D.; Kassar, T.; Richter, M.; et al. Abnormal strong burn-in degradation of highly efficient polymer solar cells caused by spinodal donor−acceptor demixing. Nat. Commun. 2017, 8, No. 14541. (38) Liu, F.; Chen, D.; Wang, C.; Luo, K.; Gu, W.; Briseno, A. L.; Hsu, J. W. P.; Russell, T. P. Molecular weight dependence of the morphology in P3HT:PCBM solar cells. ACS Appl. Mater. Interfaces 2014, 6, 19876−19887. (39) Kozub, D. R.; Vakhshouri, K.; Orme, L. M.; Wang, C.; Hexemer, A.; Gomez, E. D. Polymer crystallization of partially miscible polythiophene/fullerene mixtures controls morphology. Macromolecules 2011, 44, 5722−5726. (40) Ye, L.; Collins, B. A.; Jiao, X.; Zhao, J.; Yan, H.; Ade, H. Miscibility-function relations in organic solar cells: signifcance of optimal miscibility in relation to percolation. Adv. Energy Mater. 2018, 8, No. 1703058. J

DOI: 10.1021/acs.macromol.9b00411 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules (41) Ye, L.; Hu, H.; Ghasemi, M.; Wang, T.; Collins, B. A.; Kim, J.; Jiang, K.; Carpenter, J. H.; Li, H.; Li, Z.; McAfee, T.; Zhao, J.; Chen, X.; Lai, J. L. Y.; Ma, T.; Bredas, J.; Yan, H.; Ade, H. Quantitative relations between interaction parameter, miscibility and function in organic solar cells. Nat. Mater. 2018, 17, 253−260. (42) Nishi, T.; Wang, T. T. Melting point depression and kinetic effects of cooling on crystallization in poly(vinylidene fluoride)poly(methyl methacrylate) mixtures. Macromolecules 1975, 8, 909− 915. (43) Peng, Z.; Jiao, X.; Ye, L.; Li, S.; Rech, J. J.; You, W.; Hou, J.; Ade, H. Measuring Temperature-dependent miscibility for polymer solar cell blends: An Easily Accessible Optical Method Reveals Complex Behavior. Chem. Mater. 2018, 30, 3943−3951. (44) Sharma, A.; Pan, X.; Campbell, J. A.; Andersson, M. R.; Lewis, D. A. Unravelling the thermomechanical properties of bulk heterojunction blends in polymer solar cells. Macromolecules 2017, 50, 3347−3354. (45) Fox, T. G.; Flory, P. J. Second-order transition temperatures and related properties of polystyrene. I. Influence of molecular weight. J. Appl. Phys. 1950, 21, 581−591. (46) Jung, J.; Lee, W.; Lee, C.; Ahn, H.; Kim, B. J. Controlling molecular orientation of naphthalenediimide-based polymer acceptors for high performance all-polymer solar cells. Adv. Energy Mater. 2016, 6, No. 1600504. (47) Zhang, Z. G.; Qi, B.; Jin, Z.; Chi, D.; Qi, Z.; Li, Y.; Wang, J. Perylene diimides: a thickness-insensitive cathode interlayer for high performance polymer solar cells. Energy Environ. Sci. 2014, 7, 1966− 1973. (48) Bin, H.; Gao, L.; Zhang, Z. G.; Yang, Y.; Zhang, Y.; Zhang, C.; Chen, S.; Xue, L.; Yang, C.; Xiao, M.; Li, Y. 11.4% Efficiency nonfullerene polymer solar cells with trialkylsilyl substituted 2Dconjugated polymer as donor. Nat. Commun. 2016, 7, No. 13651.

K

DOI: 10.1021/acs.macromol.9b00411 Macromolecules XXXX, XXX, XXX−XXX