Rational design of conjugated side chains for high-performance all-polymer solar cells

Abstract

With their rapid development in recent years, organic solar cells hold the exciting potential to be a groundbreaking form of solar harvesting technology. Among these, all-polymer solar cells (all-PSCs) stand out owing to various advantages such as complementary absorption and superior stability. However, the advance of all-PSCs has been greatly impeded by energy level mismatches and unfavorable morphology of the active layer. Here, we report a molecular engineering approach featuring asymmetrical 4-methoxythiophene/thiophene as conjugated side chains of the donor polymer to fine-tune the energy level alignment and phase separation. The corresponding polymer, namely, poly{4-[5-(2-ethylhexyl)-4-methoxythiophen-2-yl]-8-[5-(2-ethylhexyl)thiophen-2-yl]benzo[1,2-b:4,5-b′]dithiophene}-alt-[bis(5-thiophene-2-yl)-5,6-difluoro-2-(2-hexyldecyl)-2H-benzo[d][1,2,3]triazole-4,7-diyl] (PMOT32) exhibited a power conversion efficiency that exceeded 8.5% in all-PSCs with poly{[N,N′-bis(2-octyldodecyl)-1,4,5,8-naphthalenedicarboximide-2,6-diyl]-alt-5,5′-(2,2′-bithiophene)} (N2200) as the acceptor. PMOT32:N2200 also maintained good photovoltaic performance when processed as a thick film or from non-halogenated solvents. Detailed comparisons with two other polymers with pure thiophene or 4-methoxythiophene as their side chains revealed how these side chains affected the photovoltaic performance via energy level alignment and phase separation.

---

Design, System, Application

To optimize the photovoltaic performance of all-polymer solar cells (all-PSCs), it is desirable to achieve an appropriate energy level alignment between the donor polymer and the acceptor polymer to afford a high open-circuit voltage and a sufficient energy offset for charge separation. Control of the morphology of the active layer is also important, as it can significantly affect charge transport. Here, we report the design of efficient donor polymers with suitable energy levels and favorable morphology for all-PSCs via fine-tuning their conjugated side chains. A new asymmetrically substituted benzo[1,2-b:4,5-b′]dithiophene (BDT) monomer with one 4-methoxythiophene (MOT) chain and one thiophene chain (BDT-MOT/T) and the corresponding polymer PMOT32 were synthesized. In comparison with the thiophene chain, the MOT chain lowered the energy levels of the donor polymer. When blended with N2200 as the acceptor polymer, PMOT32 solar cells exhibited a better match of energy levels and thus higher photovoltaic performance than two other polymers with pure thiophene or 4-methoxythiophene as their side chains. The photovoltaic performance of these polymer systems was correlated with the phase separation between the donor and acceptor polymers. This study can provide insights into molecular engineering to further enhance the performance of all-PSCs.

1 Introduction

All-polymer solar cells (all-PSCs), in which an electron-donor polymer interpenetrates an electron–acceptor polymer to form the active layer, have regained attention in recent years.^1–3 In comparison with polymer solar cells based on a fullerene acceptor, all-PSCs possess several advantages. Variation of the organic building blocks enables easy matching of the donor and acceptor polymers in terms of absorption spectra and energy levels.³ Complementary absorption ranges from the visible to the near infrared benefit solar harvesting, and appropriately aligned energy levels can minimize energy loss while providing a sufficient energy offset for exciton separation.⁴ In addition, entanglements between polymer chains can enable all-PSCs to exhibit superior stability⁴ and flexibility,⁵ which are difficult to achieve with fullerene PSCs.

Although promising, the development of all-PSCs greatly lags behind that of PSCs based on fullerene⁶ or non-fullerene small-molecule acceptors (SMAs).^7–9 In fact, most all-PSCs have performance inferior to that of their fullerene counterparts.³ Low electron mobility and non-optimal morphology are major bottlenecks in improving the photovoltaic performance of all-PSCs. Extensive efforts have been devoted to the development of efficient acceptor polymers based on electron-deficient building blocks, such as cyanated poly(phenylenevinylene) (CN-PPV),¹ benzothiadiazole (BT),¹⁰ naphthalenediimide (NDI),^4,11–17 perylenediimide (PDI)^18–20 and double B ← N bridged bipyridine (BNBP).²¹ Among these, N2200 copolymerized with NDI and dithiophene and its derivatives are attractive acceptor materials.^4,11–13,16 All-PSCs with PTB7-Th as the donor and N2200 as the acceptor have achieved a power conversion efficiency (PCE) of 5.7%.¹² By further molecular engineering of acceptor polymers, such as side chain modifications,¹⁶ fluorination of thiophene,¹⁵ selenophene substitution¹⁴ and copolymerization,¹⁷ improved photovoltaic performance has been recently demonstrated owing to enhanced electron mobility or improved phase separation in the active layer.¹⁴

In contrast to the intensive research into acceptor polymers, donor polymers in all-PSCs have received less attention. PTB7 (ref. 11 and 22) and PTB7-Th^{12,14–18,21} are widely used as donor polymers in all-PSCs, and a PCE of 7.7% has been reported in a PTB7-Th:PNDIS-HD system.¹⁴ Nevertheless, most high-performance all-PSC systems that use PTB7-Th as the donor suffer from a low open-circuit voltage (V_oc) and a low fill factor. Recently, Li et al. reported an efficient all-PSC system with an impressively high short-circuit current density (J_sc) and fill factor.⁴ The donor polymer J51,⁴ which was constructed from 5,6-difluoro-2H-benzo[d][1,2,3]triazole (FTAZ)²³ and benzo[1,2-b:4,5-b′]dithiophene (BDT), exhibited complementary absorption ranges and balanced hole/electron mobility in the active layer when blended with N2200, which afforded a high PCE that exceeded 8%. However, the relatively high-lying highest occupied molecular orbital (HOMO) of J51 causes the system to have a moderate V_oc and relatively large energy losses,²⁴ which limits further improvements in the PCE.

Recently, we developed a feasible method of lowering the HOMO of donor polymers by introducing 4-methoxythiophene (MOT) as conjugated side chains on BDT.^25,26 This approach has proved to be effective in increasing the V_oc and PCE in PSCs based on both fullerene and fused-ring electron acceptors (FREAs). Here, we report two new FTAZ-based p-type polymers with MOT-modified BDT and investigate their application in high-performance all-PSCs. In order to fine-tune the energy levels, we designed and synthesized a new asymmetrically substituted BDT monomer with one MOT chain and one thiophene chain (BDT-MOT/T). When the number of MOT substituents on BDT increased, the polymer exhibited a lower HOMO level and thus a higher V_oc in all-PSCs. However, the highest PCE was achieved with PMOT32 prepared from BDT-MOT/T, instead of PMOT34 prepared from BDT-MOT, when blended with N2200 in all-PSCs. The PMOT32:N2200 system demonstrated an acceptable PCE of 8.59% from chloroform (CF), 7.98% from o-xylene and 8.56% from 2-methyltetrahydrofuran (MeTHF), which are among the best performances for all-PSCs. The influence of the MOT substituent on the photovoltaic performance of all-PSCs is correlated with the energy level alignment and blend morphology.

2 Experiments

2.1 Synthesis of monomers and polymers

General. Reagents were purchased from Sigma-Aldrich or J&K and were used as received without further purification. Organic solvents used for reactions were purified using a solvent purification system (Innovative Technology, Inc.) before use. ¹H NMR spectra in CDCl₃ were recorded with a Bruker Avance 500 system (500 MHz). Gel permeation chromatography (GPC) was performed at 170 °C on a PL-GPC220 chromatograph using polystyrene as the standard and trichlorobenzene as the eluent. The synthesis procedures of the BDT-MOT/T organotin monomer are described in Scheme S1.† The BDT-MOT organotin monomer was synthesized according to our previous work.²⁵ The BDT-T organotin monomer and 4,7-bis(thiophen-2-yl)-5,6-difluoro-2-(2-hexyldecyl)-2H-benzo[d][1,2,3]triazole dibromo monomer were prepared according to previous reports.^23,27 Poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS, Clevios PVP AI4083) was obtained from NCM International.

BDT-MOT/T (4-(5-(2-ethylhexyl)-4-methoxythiophen-2-yl)-8-(5-(2-ethylhexyl)thiophen-2-yl)benzo[1,2-b:4,5-b′]dithiophene). A pre-dried 250 mL flask was charged with 2-(2-ethylhexyl)thiophene (2.744 g, 14.0 mmol) and anhydrous THF (28 mL) and the solution was cooled to 0 °C, after which a solution of n-BuLi (2.4 M in hexane, 5.82 mL) was added dropwise under the protection of argon. The mixture was kept at room temperature for 1.5 hours. After that, the reaction mixture was transferred into a flask in which benzo[1,2-b:4,5-b′]dithiophene-4,8-dione (3.42 g, 15.5 mmol) was dissolved in anhydrous THF (31 mL). The reaction was conducted at 60 °C for 2.5 hours, during which the reaction of MOT (7.2 g, 31.8 mmol) in anhydrous THF (60 mL) with a solution of n-BuLi (2.4 M in hexane, 5.82 mL) was performed at 0 °C for 1.5 hours. After cooling to room temperature, the solution of lithiated MOT was introduced and the reaction was continued for a further 1.5 hours at 80 °C. SnCl₂·2H₂O (24.53 g, 108.7 mmol) in 10% HCl (98.5 mL) was introduced at 0 °C, and the mixture was stirred for a further 2 hours at 80 °C. The mixture was poured into ice water after cooling to ambient temperature. The organic layer was extracted with EtOAc and washed several times with H₂O. Further purification was performed by column chromatography with CH₂Cl₂/hexane (1/9, v/v) as the eluent to yield a yellow oil (3.15 g, yield: 33.2%). ¹H NMR (400 MHz, CDCl₃, δ): 7.70 (d, 1H), 7.65 (d, 1H), 7.47 (dd, 2H), 7.30 (d, 1H), 7.24 (s, 1H), 6.90 (d, 1H), 3.90 (s, 3H), 2.87 (d, 2H), 2.77 (d, 2H), 1.68 (m, 2H), 1.49–1.27 (m, 16H), 0.98–0.88 (m, 12H).

BDT-MOT/T-Sn₂: ((4-(5-(2-ethylhexyl)-4-methoxythiophen-2-yl)-8-(5-(2-ethylhexyl)thiophen-2-yl)benzo[1,2-b:4,5-b′]dithiophene-2,6-diyl)bis(trimethylstannane)). BDT-MOT/T (0.78 g, 1.28 mmol) was dissolved in anhydrous THF (20 mL) in a 100 mL argon-purged flask, and then a solution of LDA (2.0 M in THF, 2.57 mL) was added at −78 °C. The reaction mixture was then stirred for 2.0 hours at this temperature. Subsequently, trimethylstannyl chloride (1.0 M in THF, 5.65 mL) was added and the mixture was stirred overnight at ambient temperature. The organic layer was extracted by diethyl ether, washed several times with water, and concentrated to obtain the crude product. The target compound was obtained by recrystallization from isopropanol (0.31 g, yield: 51.8%). ¹H NMR (400 MHz, CDCl₃, δ): 7.71 (d, 2H), 7.31 (d, 1H), 7.24 (s, 1H), 6.90 (d, 1H), 3.90 (s, 3H), 2.87 (dd, 2H), 2.76 (dd, 2H), 1.68 (m, 2H), 1.49–1.25 (m, 16H), 0.98–0.88 (m, 12H), 0.40 (s, 18H).

PMOT32. A 25 mL pre-dried flask was charged with BDT-MOT/T-Sn₂ (164.73 mg, 0.1763 mmol), 4,7-bis(5-bromothiophen-2-yl)-5,6-difluoro-2-(2-hexyldecyl)-2H-benzo[d][1,2,3]triazole (123.66 mg, 0.1763 mmol) and Pd(PPh₃)₄ (7 mg). The flask was successively evacuated and refilled with argon for 3 cycles. Then toluene (7 mL) and DMF (1 mL) were added. The mixture was reacted at 120 °C for 8 hours. The crude product was collected by precipitation from MeOH. The solid was then rinsed in a Soxhlet extractor with MeOH, EtOAc, hexane and chloroform successively. The solution in chloroform was concentrated and then precipitated in MeOH. After drying at a reduced pressure, 168 mg of the copolymer was obtained. Yield: 82.8%. M_n: 38.5 kDa; PDI: 2.16.

PMOT34. The copolymer was obtained via the same procedure as PMOT32 by replacing the BDT-MOT/T-Sn₂ monomer with the BDT-MOT-Sn₂ monomer. Yield: 97.3%. M_n: 32.8 kDa; PDI: 1.91.

J52. The copolymer was obtained via the same procedure according to the literature.²⁷ Yield: 97.4%. M_n: 31.4 kDa; PDI: 1.79.

2.2 Cyclic voltammetry

Cyclic voltammetry (CV) measurements were performed with an electrochemical workstation (CHI 600E) to determine the ionization potential (IP) and electron affinity (EA) of the polymers. Polymer films were dip-coated from chloroform solutions on glassy carbon working electrodes (diameter of 3 mm). CV curves were recorded under an argon atmosphere in CH₃CN solutions containing 0.1 M Bu₄NPF₆ with a Pt wire as the counter electrode and Ag/Ag⁺ as the reference electrode. The redox potential of ferrocene/ferrocenium (Fc/Fc⁺) under the same conditions is located at 0.073 V, which is assumed to have an absolute energy level of −4.8 eV with respect to vacuum. The IP and EA values were calculated by the following equations:

E_IP = E_ox + [4.8 − 0.073] (eV)

E_EA = E_red + [4.8 − 0.073] (eV)

2.3 Device fabrication

OPV devices were fabricated with a traditional structure, namely, indium tin oxide (ITO)/poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS)/donor polymer:N2200/perylenediimides functionalized with amine oxide (PDINO)²⁸/Al. The ITO substrates were cleaned in an ultrasonic bath with detergent water, deionized water, acetone, and isopropanol, and then treated with UV-ozone for 15 min. A layer of PEDOT:PSS was spin-coated onto ITO glass in air at 2500 rpm for 30 s to give an average thickness of ∼50 nm. After annealing at 150 °C for 15 min, the substrates were transferred into a N₂-filled glove box. The active layer was spin-coated onto PEDOT:PSS and then vacuumed for 3 hours to remove residual DIO. Two kinds of solvent were used. The polymers were dissolved in CF + 0.7% DIO or o-xylene + 1% DIO. The concentrations were 6.5 mg mL⁻¹ for J52, PMOT32 and PMOT34 and 3 mg mL⁻¹ for N2200. The active layers were spin-coated at 2500 rpm for 30 s when CF was used and 1500 rpm for 30 s when o-xylene was used. For the MeTHF solution, the concentration of PMOT32 was 4.5 mg mL⁻¹ and the donor/acceptor ratio was 2 : 1 (w/w). This solution was heated at 80 °C overnight and was then spin-coated at 1500 rpm for 30 s to give a blend film with a thickness of ∼110 nm. Subsequently, the as-cast layers were annealed at 100 °C for 10 min. PDINO was dissolved in ethanol to give a concentration of 1.1 mg mL⁻¹ and was then spin-coated onto the active layers at 3000 rpm for 30 s. Finally, 80 nm Al was evaporated on the top surface under a vacuum of 3 × 10⁻⁶ mbar. The active area of the devices was 0.045 cm².

2.4 Characterization of PSC devices

The thicknesses were measured using an XP-200 stylus profilometer (KLA-Tencor). Absorption spectra were recorded with a UV-vis spectrophotometer (UV-3600, Shimadzu). J–V measurements of the devices under illumination by an AM1.5G solar simulator (Enli Technology, 100 mW cm⁻², calibrated by a silicon reference cell) were performed using a computer-controlled Keithley 2400 Source Measure Unit. EQE measurements were performed using an Enli Technology QE-R system with a bromine tungsten lamp as the light source. AFM images of the active layers were acquired with an Asylum MFP-3D Stand Alone microscope in tapping mode. TEM images were acquired with an FEI Helios Nanolab 600i system. PL was recorded with a PerkinElmer LS55 fluorescence spectrometer. GIWAXS and RSoXS measurements were performed at the Advanced Light Source. GIWAXS measurements were performed at beamline 7.3.3.²⁹ Samples were prepared on Si substrates using identical blend solutions to those used in the devices. The 10 keV X-ray beam was incident at a grazing angle of 0.11–0.15°, which was selected to maximize the intensity of scattering from the samples. The scattered X-rays were detected using a Dectris Pilatus 2M photon-counting detector. RSoXS transmission measurements were performed at beamline 11.0.1.2.³⁰ Samples used for RSoXS measurements were prepared on a PSS-modified Si substrate under the same conditions as those used for device fabrication and were then transferred by floating in water to a 1.5 mm × 1.5 mm Si₃N₄ membrane with a thickness of 100 nm supported by a 5 mm × 5 mm Si frame with a thickness of 200 μm (Norcada, Inc.). 2-D scattering patterns were collected with an in-vacuum CCD camera (Princeton Instrument PI-MTE). The distance of the sample detector was calibrated from the diffraction peaks of the triblock copolymer poly(isoprene-b-styrene-b-2-vinylpyridine), which has a known spacing of 391 Å. The beam size at the sample was approximately 100 μm by 200 μm.

2.5 Mobility measured by space-charge-limited current method

To study the charge transport properties, both hole-only and electron-only devices were fabricated with architectures of ITO/PEDOT:PSS/polymer blend/MoO_x/Ag and ITO/ZnO/polymer blend/PDINO/Al, respectively. The hole/electron mobility was determined by plotting and fitting current–voltage curves. The SCLC is determined by:

J = (9/8)ε_rε₀μ(V²/L³)

where ε₀ is the permittivity of free space, ε_r is the dielectric constant of the polymer, μ is the hole or electron mobility, V is the voltage drop across the device and L is the polymer thickness. The dielectric constant ε_r was assumed to be 3, which is a typical value for conjugated polymers. The thicknesses of the films were measured with a Tencor D120 profilometer.

3 Results

The asymmetric BDT monomer with one thiophene side chain and one MOT side chain, namely, BDT-MOT/T, was prepared using a similar method to BDT-MOT but by subsequently reacting benzo[1,2-b:4,5-b′]dithiophene-4,8-dione with (5-alkylthiophen-2-yl)lithium and (4-methoxy-5-alkylthiophen-2-yl)lithium in a molar ratio of 1 : 1 : 1 (see Scheme S1, ESI†). PMOT32 was subsequently synthesized by a Stille polycondensation reaction between the BDT-MOT/T distannous monomer and 4,7-bis(5-bromothiophen-2-yl)-5,6-difluoro-2-(2-hexyldecyl)-2H-benzo[d][1,2,3]triazole. Two other polymers, namely, J52 (ref. 27) with thiophene conjugated side chains and PMOT34 with MOT conjugated side chains, were also synthesized for comparison (ESI†). Scheme 1 illustrates the molecular structures of these three polymers, as well as N2200. The number-average molecular weight (M_n) and polydispersity index (PDI), which were estimated by high-temperature gel permeation chromatography (GPC), were 38.5 kDa and 2.16 for PMOT32, 32.8 kDa and 1.91 for PMOT34, and 31.4 kDa and 1.79 for J52, respectively (Table S1†). Both PMOT32 and PMOT34 exhibited high solubility in CF, chlorobenzene (CB), o-dichlorobenzene (o-DCB) and o-xylene at room temperature.

Scheme 1 Chemical structures of the related materials in this work.

Fig. 1a and S1a† show the absorption spectra of the donor polymers and N2200 in film and solution states, respectively. The donor polymers exhibited similar absorption spectra, which suggested that the MOT substituent had a similar impact on the optical properties when compared with the thiophene substituent. For the pristine films, the replacement of thiophene with MOT side chains resulted in a slight blue shift in the absorption edge, which was 631 nm for J52, 629 nm for PMOT32 and 624 nm for PMOT34, respectively (Fig. 1a). The optical band gap was thus calculated to be 1.96 eV for J52, 1.97 eV for PMOT32 and 1.99 eV for PMOT34, respectively. The absorption ranges of these donor polymers can complement that of the low-band-gap polymer acceptor N2200 (Fig. 1a), which favors solar harvesting from the ultraviolet-visible to the near-infrared region.⁴

Fig. 1 (a) Normalized absorbance of the pristine polymer films. (b) Schematic diagram of the energy levels of the polymers as determined by cyclic voltammetry. The ionization potential of N2200 was also calculated on the basis of the optical band gap and the measured electron affinity (number in red).

Cyclic voltammetry (CV) was used to determine the energy levels of the polymers. Using ferrocene as an internal standard, the ionization potential (IP) and electron affinity (EA) were 5.12 eV/3.17 eV for J52, 5.18 eV/3.25 eV for PMOT32, 5.26 eV/3.33 eV for PMOT34 and 5.84 eV/3.79 eV for N2200, respectively (Fig. 1b and S2†). It is clear that the IP increases (HOMO is lowered) in proportion to the increase in the ratio of MOT to thiophene in the conjugated side chains from J52 to PMOT32 and PMOT34, which is consistent with our previous work.^25,26 In contrast to the donor polymers, the band gap of N2200 measured by CV is much larger than the optical band gap. We deduced that the IP of N2200 might be overestimated by CV, as N2200 is a strong electron acceptor. Therefore, an IP of 5.25 eV was calculated by adding the optical band gap to the EA of N2200 (Fig. 1b). It is noted that the calculated IP of N2200 is very similar to the measured IP of PMOT34, which might cause problems with charge separation and recombination in the blend after photoexcitation. The photophysical properties of the four polymers are summarized in Table S1.†

The photovoltaic performance of the all-PSCs was investigated in a conventional structure of ITO/PEDOT:PSS/active layer/PDINO²⁸/Al. Typical current density–voltage (J–V) curves of the all-PSCs are displayed in Fig. 2, and the corresponding device parameters are summarized in Table 1, S2 and S3.† When processed from CF with 0.7% (by volume) 1,8-diiodooctane (DIO) as an additive, PMOT32:N2200 exhibited a V_oc of 0.871 V, a J_sc of 13.84 mA cm⁻², a fill factor of 71.2% and a PCE of 8.59% in the best-performing device. In contrast, the counterparts displayed inferior performance, with a V_oc of 0.798 V and a PCE of 7.51% for J52 and a V_oc of 0.912 V and a PCE of 7.25% for PMOT34. The significant increase in the V_oc mainly accounted for the superior performance of PMOT32 in comparison with J52, which corresponded to the lower HOMO level of PMOT32. For PMOT34:N2200, however, the highest V_oc among these systems did not result in the highest PCE owing to the substantial decline in the J_sc (11.83 mA cm⁻² for PMOT34).

Fig. 2 (a) J–V characteristics of polymer:N2200 solar cells fabricated from CF with 0.7% DIO under AM1.5G solar radiation and (b) EQE spectra of the corresponding devices in (a). (c) Power conversion efficiency of all-PSCs with various thicknesses of the active layer. (d) J–V characteristics of polymer:N2200 solar cells fabricated from o-xylene with 1% DIO under AM1.5G solar radiation.

Table 1 Summary of the photovoltaic performance of all-PSCs based on donor polymer:N2200 (2.17/1, wt/wt) under AM1.5G solar radiation (100 mW cm⁻²)

Donor	Solvents	V oc [V]	J sc [mA cm^−2]	FF [%]	PCE [%]
a Calculated by integrating the EQE spectra. b Average PCEs were obtained from more than 10 devices.
PMOT32	CF + 0.7% DIO	0.871	13.84 [13.27]^a	71.2	8.59 [8.50]^b
	o-Xylene + 1.0% DIO	0.864	13.21 [12.47]^a	69.9	7.98 [7.87]^b
	MeTHF	0.862	13.96 [13.38]^a	71.1	8.56 [8.41]^b
J52	CF + 0.7% DIO	0.798	14.05 [13.90]^a	67.0	7.51 [7.40]^b
	o-Xylene + 1.0% DIO	0.805	13.54 [12.92]^a	67.8	7.39 [7.28]^b
PMOT34	CF + 0.7% DIO	0.912	11.83 [11.15]^a	67.2	7.25 [7.11]^b
	o-Xylene + 1.0% DIO	0.915	11.10 [10.00]^a	68.4	6.95 [6.85]^b

---

Fig. 2b shows the corresponding external quantum efficiency (EQE) spectra of the all-PSC devices. PMOT32:N2200 was very efficient in converting photons into electrons with an EQE of greater than 60% in the range of 400–620 nm, but exhibited mediocre performance beyond 650 nm with an EQE of less than 40%. On the basis of the absorption data (Fig. S1b†), the optical absorption of the active layer in the range of 400–650 nm was mainly contributed by the donor polymers, whereas the acceptor polymer N2200 was dominant in the range of 650–800 nm, and both the donor polymer and N2200 made similar contributions in the range of 300–400 nm. With an increase in the acceptor : donor ratio from 1 : 2.17 to 1 : 1, the blend exhibited a gradual increase in absorption in the range of 650–800 nm (Fig. S3a†) and an increase in EQE in the corresponding region (Fig. S3b†), which was closely correlated with the slight increase in J_sc at higher acceptor contents (Table S2†). However, an acceptor : donor ratio of greater than 1 : 2.17 in the blend was detrimental to the fill factor and thus the PCE, which probably resulted from inferior phase separation and charge transport in the active layer.⁴ The PMOT34 system displayed a significantly lower EQE response when compared with the PMOT32 and J52 systems, and the percentage decrease was more obvious in the ranges of 350–450 nm and 650–800 nm, where the absorption of N2200 played a major role. It should be noted that there were negligible differences between the PMOT34 blend and the PMOT32 and J52 blends in terms of absorption in these ranges (Fig. S1b†). Interestingly, PMOT34:N2200 exhibited an EQE that was almost identical to those of the other two systems at a wavelength of 450 nm, where N2200 has low absorption. These results suggest that the inferior EQE response and thus the low J_sc of the PMOT34:N2200 system may be ascribed to the ineffective conversion of photons by N2200 in the active layer.

It has been proposed that thick active layers can be favorable for printing processes³¹ and practical applications.³² Thus, we fabricated devices with various thicknesses of the active layer to study their performance in a thick film (Fig. 2c and Table S4†). As can be observed from Fig. 2c, the all-PSCs exhibited a slow decline in PCE when the thickness exceeded the optimal value of ∼120 nm, which was possibly due to the high mobility of FTAZ-based polymers. PMOT32:N2200 still maintained a PCE of around 6% at an active layer thickness of 250 nm. Processing solvents represent another important issue for the practical application of PSCs. The chlorinated solvents that are generally used are toxic and carcinogenic. However, few PSCs can demonstrate high performance when processed from non-halogenated solvents, which is possibly due to their non-optimal morphology.⁶ Very few all-PSCs with a PCE that exceeded 7% processed from non-halogenated solvents have been reported thus far.^31,33,34 When processed from o-xylene, PMOT32:N2200 exhibited an impressive PCE of 7.98% with a V_oc of 0.864 V, a J_sc of 13.21 mA cm⁻², and a fill factor of 69.9% in the best-performing device (Fig. 2d, Tables 1 and S4†). In comparison, J52:N2200 and PMOT34:N2200 achieved a PCE of 7.39% and 6.95%, respectively (Fig. 2d and Table 1). The significant increase in V_oc and the slight enhancement in the fill factor contributed to the better photovoltaic performance of the PMOT32:N2200 system in comparison with J52:N2200, which was similar to that of the systems processed from CF. The corresponding EQE spectra of the respective devices are shown in Fig. S4† and resemble those of the systems processed from CF. Very recently, Huang et al. reported a PTzBI:N2200 system that had an efficiency of ∼9% when processed from MeTHF.³⁴ We also experimented with MeTHF as the processing solvent for the PMOT32:N2200 system, and an acceptable PCE of 8.56% was achieved (V_oc = 0.862 V, J_sc = 13.96 mA cm⁻² and FF = 71.1%) (Fig. S5†). It should be pointed out that a hot MeTHF solution is required for optimal processing. Preliminary studies of the stability of the PMOT32:N2200 devices were also performed (Fig. S6†). Unencapsulated devices stored in a glove box exhibited relatively high stability, with a decline in efficiency of about 12% over 9 days. Encapsulated devices in air were less stable, with a decline in efficiency of 22% over 9 days. These unsatisfactory stabilities were probably due to the acidic and hygroscopic properties of the hole transport material PEDOT:PSS in the devices. Increases in the stability of these polymer solar cell devices are expected if the device structure can be further optimized. The synthesis of PMOT32 can be less complicated than that of the well-known donor polymer PTB7-TH, and further optimization of synthesis and scale-up will produce these materials at acceptable costs.

In order to obtain further understanding of how the MOT conjugated side chain affects the photovoltaic properties, atomic force microscopy (AFM) and transmission electron microscopy (TEM) were employed to study the morphology of the active layer of the all-PSCs. Fig. S7† shows AFM topography images of films of the three different blends processed from CF with or without thermal annealing (TA). All the films exhibited a similar topography with a low root mean square (RMS) roughness of around 0.85–0.93 nm. However, a significant difference appeared in the TEM images (Fig. 3). Dark spots were observed for all these polymer blends before TA. These dark spots can be correlated with N2200-rich domains owing to the high electron affinity of N2200.¹¹ The dark spots became smaller from J52 to PMOT32, and the PMOT34 blend displayed the smallest and fewest dark spots, which suggests that the MOT substituent decreased the donor polymer/N2200 phase separation. After TA treatment, both J52 and PMOT32 exhibited more sharply contrasting features and dendritic morphology with an interconnected network of a dark phase,¹¹ which demonstrated greater donor/acceptor phase separation after TA. However, homogeneous fine features were observed in the PMOT34 blend film. Favorable phase separation and aggregation of N2200 in the active layers could improve charge transport and suppress geminate recombination.³⁵ Thus, the improvement in the photovoltaic performance of the J52 and PMOT32 blends after TA was more prominent than that for the PMOT34 blend (Table S3†). A similar trend in phase separation was also observed for films processed from o-xylene after thermal annealing (Fig. S8†).

Fig. 3 TEM images of three different blends prepared under two different film processing conditions: without TA (top row) and TA for 10 minutes (bottom row): (a) and (d) J52:N2200, (b) and (e) PMOT32:N2200 and (c) and (f) PMOT34:N2200, respectively, processed from CF with 0.7% DIO. The scale bars are indicated on the images.

To further investigate the micromorphology of these systems, grazing-incidence wide-angle X-ray scattering (GIWAXS) measurements were performed (Fig. 4 and S8a†).³⁶ For pristine films, the lamellar packing peaks of PMOT32 and PMOT34 were located at q ≈ 0.30 Å⁻¹, which corresponded to a lamellar spacing of ∼21 Å, whereas the π–π stacking peaks of PMOT32 and PMOT34 were located at q ≈ 1.7 Å⁻¹ (∼3.7 Å). The lamellar and π–π spacings of PMOT32 and PMOT34 were nearly identical to those of J52, which suggests that the introduction of MOT as a conjugated side chain did not change the packing model of the polymers. When blended with N2200, the lamellar peaks of the three blends shifted to q ≈ 0.27 Å⁻¹ (∼23 Å), whereas a negligible difference was observed in the π–π stacking peaks, which indicated that N2200 did not enter the crystalline phase of the donor polymers but formed a crystalline phase in the lamellar spaces instead.⁴ Because the GIWAXS results did not reveal any difference between the three polymer blends, resonant soft X-ray scattering (RSoXS)^36–38 measurements were also conducted (Fig. S9b†). As can be seen from Fig. S9b,† all these blends exhibited similar scattering profiles, which proved that domains with similar sizes of 20–30 nm were present in the three active layers. As RSoXS provides enhanced contrast between different organic components, the domain size revealed by RSoXS represents phase separation inside the donor-rich phase (bright region in TEM). However, analysis of the purity of the three domains revealed that the PMOT32 blend had higher purity (relative purity of 1) than that of the PMOT34 blend (relative purity of 0.89), which is favorable for charge transport and thus improvements in device performance.^31,37

Fig. 4 Two-dimensional GIWAXS images of pristine polymers and blends with N2200: (a) J52, (b) PMOT32, (c) PMOT34, (d) J52:N2200, (e) PMOT32:N2200, and (f) PMOT34:N2200.

The space-charge-limited current (SCLC) model was used to investigate the charge transport characteristics (Table S5 and Fig. S10†). The hole-only architecture of ITO/PEDOT:PSS/polymer blend/MoO_x/Ag and the electron-only architecture of ITO/ZnO/polymer blend/PDINO/Al were employed to measure the hole and electron mobility, respectively. As can be seen from Fig. S10 and Table S5,† all these systems exhibited high hole/electron mobility, which could account for the high fill factors of these systems. The hole mobility of the three blends was similar (∼8 × 10⁻⁴ cm² V⁻¹ s⁻¹) before TA and increased slightly after TA (∼9 × 10⁻⁴ cm² V⁻¹ s⁻¹). With respect to electron mobility, although all these blends exhibited similar values (∼2.6 × 10⁻⁴ cm² V⁻¹ s⁻¹) before TA, PMOT32 and J52 displayed significant increases after TA (∼3.8 × 10⁻⁴ cm² V⁻¹ s⁻¹), whereas that of PMOT34 remained virtually unchanged, which was probably caused by inferior phase separation in the PMOT34 blend, which was closely correlated with the unfavorable morphology in TEM and lower purity in RSoXS. The lower electron mobility could be partly responsible for the lower J_sc in PMOT34:N2200 PSCs, which was consistent with the EQE spectra. The dependence of the J_sc of the three blend films on the light intensity was also measured to further investigate bimolecular recombination in all-PSCs (Fig. S11†). The relationship between J_sc and light intensity (P) is formulated as J_sc ∝ P^S. If all the free charges are swept out and collected without bimolecular recombination, the value of S (the slope in logarithmic coordinates) should be equal to 1. The three all-PSCs gave slopes of 0.984, 0.992 and 0.992 for J52:N2200, PMOT32:N2200 and PMOT34:N2200, respectively, which indicated that bimolecular recombination was negligible in all these systems, which agrees closely with the high charge mobility in the three blends.

Photoluminescence (PL) measurements were conducted to study charge dissociation and recombination in the active layer (Fig. 5a, b and S12†).^35,39 The PMOT34 blend displayed more intense PL than the other two blends, which probably resulted from the decrease in the effectiveness of charge separation due to the adjacent HOMO energy levels of PMOT34 and N2200, although homogeneity of the two mixing phases is usually beneficial for the dissociation of excitons at the interface.³⁵

Fig. 5 Normalized PL spectra of pristine and blend films excited at a wavelength of 545 nm: (a) PMOT32 system and (b) PMOT34 system. (c) Plots of J_phversus V_eff for the three systems and (d) the corresponding curves of P(E, T) versus V_eff for the systems in (c).

The maximum exciton generation rates (G_max) were calculated using the equation G_max = J_sat/qL, where J_sat is the saturation photocurrent density, q is the electronic charge and L is the thickness of the active layers in the all-PSCs.¹⁵J_sat is determined from a curve of effective voltage (V_eff) (V_eff = V₀ − V, where V₀ is the voltage when J_ph = 0 and V is the applied bias) versus photocurrent density (J_ph) (J_ph is defined as J_ph = J_L − J_D, where J_L and J_D are the current densities under illumination and in the dark, respectively). The J_ph value was saturated at V_eff values close to 1 V for the three all-PSCs (Fig. 5c and Table S6†). The G_max was determined to be 8.39 × 10²⁷ m⁻³ s⁻¹ for J52:N2200, 8.23 × 10²⁷ m⁻³ s⁻¹ for PMOT32:N2200 and 7.39 × 10²⁷ m⁻³ s⁻¹ for PMOT34:N2200, respectively. Clearly, the PMOT34:N2200 system gave a lower G_max than those of the other two systems, which corresponds to the fact that the J_sc value was the lowest. On the basis of the G_max value, the exciton dissociation probability P(E, T) was subsequently obtained by the equation P(E, T) = J_ph/qLG_max. As shown in Fig. 5d and Table S6,† unlike the PMOT32 and J52 systems, PMOT34:N2200 exhibited a lower dissociation probability of 0.830. The relatively low value of P(E, T) for the PMOT34 system indicates ineffective charge dissociation and extraction caused by the insufficient energy offset and lower electron mobility, which is in good agreement with the PL results.

4 Conclusions

In summary, we have demonstrated an approach based on fine-tuning side chains on donor polymers for high-performance all-PSCs prepared from both halogenated and non-halogenated solvents. By introducing asymmetric conjugated side chains with one 4-methoxythiophene and one thiophene side chain on a BDT unit, PMOT32 exhibited increases in both V_oc and PCE, which exceeded 8.5% in all-polymer solar cells with an N2200 acceptor. In contrast, although the PMOT34 polymer, which was substituted with 4-methoxythiophene, displayed a higher V_oc, it suffered from inferior charge separation and unfavorable phase separation, which led to inferior J_sc and PCE values. Our study reveals how engineering a conjugated side chain on donor polymers affects the photovoltaic performance of all-PSCs via tuning the energy level alignment and phase separation. This provides new insights into the optimization of the molecular structure for high-performance all-PSCs.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors acknowledge the funding support from the Recruitment Program of Global Youth Experts of China and National Science Foundation of China (51503095, 21504066, 21534003). YYL and TBY thank the Shenzhen fundamental research programs (No. JCYJ20150630145302226, JCYJ20150630145302236), Shenzhen Key Lab funding (ZDSYS201505291525382) and Peacock Plan (KQTD20140630110339343) for financial support. The support from the Ministry of Science and Technology (No. 2016YFA0200700) is acknowledged. X-ray data was acquired at beamlines 7.3.3 and 11.0.1.2 at the Advanced Light Source, which is supported by the Director, Office of Science, Office of Basic Energy Sciences, of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231. The authors thank Chenhui Zhu at beamline 7.3.3 and Cheng Wang at beamline 11.0.1.2 for assistance with data acquisition.

References

1. J. J. M. Halls, C. A. Walsh, N. C. Greenham, E. A. Marseglia, R. H. Friend, S. C. Moratti and A. B. Holmes, Nature, 1995, 376, 498–500.
2. G. Yu and A. J. Heeger, J. Appl. Phys., 1995, 78, 4510–4515.
3. A. Facchetti, Mater. Today, 2013, 16, 123–132.
4. L. Gao, Z.-G. Zhang, L. Xue, J. Min, J. Zhang, Z. Wei and Y. Li, Adv. Mater., 2016, 28, 1884–1890.
5. T. Kim, J.-H. Kim, T. E. Kang, C. Lee, H. Kang, M. Shin, C. Wang, B. Ma, U. Jeong, T.-S. Kim and B. J. Kim, Nat. Commun., 2015, 6, 8547.
6. J. Zhao, Y. Li, G. Yang, K. Jiang, H. Lin, H. Ade, W. Ma and H. Yan, Nat. Energy, 2016, 1, 15027.
7. Y. Lin, F. Zhao, Y. Wu, K. Chen, Y. Xia, G. Li, S. K. K. Prasad, J. Zhu, L. Huo, H. Bin, Z.-G. Zhang, X. Guo, M. Zhang, Y. Sun, F. Gao, Z. Wei, W. Ma, C. Wang, J. Hodgkiss, Z. Bo, O. Inganäs, Y. Li and X. Zhan, Adv. Mater., 2017, 29, 1604155.
8. D. Baran, R. S. Ashraf, D. A. Hanifi, M. Abdelsamie, N. Gasparini, J. A. Rohr, S. Holliday, A. Wadsworth, S. Lockett, M. Neophytou, C. J. M. Emmott, J. Nelson, C. J. Brabec, A. Amassian, A. Salleo, T. Kirchartz, J. R. Durrant and I. McCulloch, Nat. Mater., 2017, 16, 363–369.
9. S. Li, L. Ye, W. Zhao, S. Zhang, S. Mukherjee, H. Ade and J. Hou, Adv. Mater., 2016, 28, 9423–9429.
10. A. C. Arias, J. D. MacKenzie, R. Stevenson, J. J. M. Halls, M. Inbasekaran, E. P. Woo, D. Richards and R. H. Friend, Macromolecules, 2001, 34, 6005–6013.
11. N. Zhou, H. Lin, S. J. Lou, X. Yu, P. Guo, E. F. Manley, S. Loser, P. Hartnett, H. Huang, M. R. Wasielewski, L. X. Chen, R. P. H. Chang, A. Facchetti and T. J. Marks, Adv. Energy Mater., 2014, 4, 1300785.
12. D. Mori, H. Benten, I. Okada, H. Ohkita and S. Ito, Energy Environ. Sci., 2014, 7, 2939–2943.
13. K. D. Deshmukh, T. Qin, J. K. Gallaher, A. C. Y. Liu, E. Gann, K. O'Donnell, L. Thomsen, J. M. Hodgkiss, S. E. Watkins and C. R. McNeill, Energy Environ. Sci., 2015, 8, 332–342.
14. Y.-J. Hwang, B. A. E. Courtright, A. S. Ferreira, S. H. Tolbert and S. A. Jenekhe, Adv. Mater., 2015, 27, 4578–4584.
15. J. W. Jung, J. W. Jo, C.-C. Chueh, F. Liu, W. H. Jo, T. P. Russell and A. K. Y. Jen, Adv. Mater., 2015, 27, 3310–3317.
16. C. Lee, H. Kang, W. Lee, T. Kim, K.-H. Kim, H. Y. Woo, C. Wang and B. J. Kim, Adv. Mater., 2015, 27, 2466–2471.
17. Z. Li, X. Xu, W. Zhang, X. Meng, W. Ma, A. Yartsev, O. Inganäs, M. R. Andersson, R. A. J. Janssen and E. Wang, J. Am. Chem. Soc., 2016, 138, 10935–10944.
18. Y. Guo, Y. Li, O. Awartani, J. Zhao, H. Han, H. Ade, D. Zhao and H. Yan, Adv. Mater., 2016, 28, 8483–8489.
19. E. Zhou, J. Cong, Q. Wei, K. Tajima, C. Yang and K. Hashimoto, Angew. Chem., Int. Ed., 2011, 50, 2799–2803.
20. Y. Zhou, T. Kurosawa, W. Ma, Y. Guo, L. Fang, K. Vandewal, Y. Diao, C. Wang, Q. Yan, J. Reinspach, J. Mei, A. L. Appleton, G. I. Koleilat, Y. Gao, S. C. B. Mannsfeld, A. Salleo, H. Ade, D. Zhao and Z. Bao, Adv. Mater., 2014, 26, 3767–3772.
21. X. Long, Z. Ding, C. Dou, J. Zhang, J. Liu and L. Wang, Adv. Mater., 2016, 28, 6504–6508.
22. Y. Liang, Z. Xu, J. Xia, S.-T. Tsai, Y. Wu, G. Li, C. Ray and L. Yu, Adv. Mater., 2010, 22, E135–E138.
23. S. C. Price, A. C. Stuart, L. Yang, H. Zhou and W. You, J. Am. Chem. Soc., 2011, 133, 4625–4631.
24. J. Liu, S. Chen, D. Qian, B. Gautam, G. Yang, J. Zhao, J. Bergqvist, F. Zhang, W. Ma, H. Ade, O. Inganäs, K. Gundogdu, F. Gao and H. Yan, Nat. Energy, 2016, 1, 16089.
25. W. Huang, M. Li, L. Zhang, T. Yang, Z. Zhang, H. Zeng, X. Zhang, L. Dang and Y. Liang, Chem. Mater., 2016, 28, 5887–5895.
26. F. Lin, W. Huang, H. Sun, J. Xin, H. Zeng, T. Yang, M. Li, X. Zhang, W. Ma and Y. Liang, Chem. Mater., 2017, 29, 5636–5645.
27. H. Bin, Z.-G. Zhang, L. Gao, S. Chen, L. Zhong, L. Xue, C. Yang and Y. Li, J. Am. Chem. Soc., 2016, 138, 4657–4664.
28. Z.-G. Zhang, B. Qi, Z. Jin, D. Chi, Z. Qi, Y. Li and J. Wang, Energy Environ. Sci., 2014, 7, 1966–1973.
29. H. Alexander, B. Wim, G. James, S. Eric, G. Eliot, K. Rick, M. Alastair, C. Matthew, R. Bruce and P. Howard, J. Phys.: Conf. Ser., 2010, 247, 012007.
30. E. Gann, A. T. Young, B. A. Collins, H. Yan, J. Nasiatka, H. A. Padmore, H. Ade, A. Hexemer and C. Wang, Rev. Sci. Instrum., 2012, 83, 045110.
31. Y. Diao, Y. Zhou, T. Kurosawa, L. Shaw, C. Wang, S. Park, Y. Guo, J. A. Reinspach, K. Gu, X. Gu, B. C. K. Tee, C. Pang, H. Yan, D. Zhao, M. F. Toney, S. C. B. Mannsfeld and Z. Bao, Nat. Commun., 2015, 6, 7955.
32. 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.
33. S. Li, H. Zhang, W. Zhao, L. Ye, H. Yao, B. Yang, S. Zhang and J. Hou, Adv. Energy Mater., 2016, 6, 1501991.
34. B. Fan, L. Ying, Z. Wang, B. He, X.-F. Jiang, F. Huang and Y. Cao, Energy Environ. Sci., 2017, 10, 1243–1251.
35. N. C. Miller, E. T. Hoke and M. D. McGehee, in Organic Photovoltaics, Wiley-VCH Verlag GmbH & Co. KGaA, 2014, pp. 421–444.
36. J. R. Tumbleston, B. A. Collins, L. Yang, A. C. Stuart, E. Gann, W. Ma, W. You and H. Ade, Nat. Photonics, 2014, 8, 385–391.
37. W. Ma, J. R. Tumbleston, M. Wang, E. Gann, F. Huang and H. Ade, Adv. Energy Mater., 2013, 3, 864–872.
38. W. Ma, J. R. Tumbleston, L. Ye, C. Wang, J. Hou and H. Ade, Adv. Mater., 2014, 26, 4234–4241.
39. G.-J. A. H. Wetzelaer, L. J. A. Koster and P. W. M. Blom, in Organic Photovoltaics, Wiley-VCH Verlag GmbH & Co. KGaA, 2014, pp. 343–376.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7me00088j

‡ These authors contributed equally to this work.

---