Introduction

Organic solar cells (OSCs) with the bulk heterojunction (BHJ) active layer have drawn wide-spread attention because of their multiple advantages such as high mechanical flexibility, light weight, and semi-transparency1,2. Driven by the rapid development of organic photovoltaic materials and device engineering, OSCs have achieved power conversion efficiencies (PCEs) of over 18%3,4,5,6,7. Despite the great success in PCE enhancement, the practical applications of OSCs are limited by their low operational stability8,9,10,11. Of note is the degradation of photoactive layer morphology under various enviromental conditions (i.e. light, oxygen and heat) after long-term exposure, which remains a grand challenge for the commercialization of OSC technology12,13,14,15,16,17,18.

Highly efficient OSCs require an optimized BHJ morphology with abundant electron donor/electron acceptor (D/A) interfaces for efficient exciton dissociation and bicontinuous D and A domains for charge transport19,20,21,22,23,24. However, the intrinsic instability of the blend photoactive layer commonly occurs during prolonged illumination or thermal aging25,26,27,28. In particular, an evolution of the blend morphology from the kinetically trapped state to the thermodynamically stable state, such as the aggregation and/or the D/A interface recession, will result in severe degradation of charge dissociation or transfer and unsatisfactory lifetime of OSCs29,30,31. For instance, Martorell32 and McGehee33 found D and A domain disordering caused by UV light, and Brabec group34 reported the major lifetime loss mechanism induced by poor D/A miscibility. Besides, Lidzey et al. demonstrated that thermal annealing above the glass transition temperature of the photoactive layers will also accelerate the device degradation35,36. Therefore, it is imperative to develop effective strategies to achieve stable optimized BHJ morphology and thus increase the device stability37.

OSCs based on polymer/polymer blends as photoactive layers (all-polymer solar cells, all-PSCs) exhibited higher morphological stability than other kinds of OSCs due to the strong interchain entanglements in all-polymer blend38,39,40,41,42,43,44,45. However, their all-polymer BHJ blends suffered from thermodynamically unfavorable miscibility and large-scale phase separation under severe surroundings46. To this end, a few studies have demonstrated that the stability issues could be overcome by using well-tailored dual functional single-component photovoltaic materials, in which two or more D and A components are covalently bonded in one material, e.g. molecular dyads, double-cable type polymers and conjugated block copolymers consisting of D and A as blocks47,48. These materials can form thermodynamically stable phase-separated domains with appropriate length scales. Thereinto, block copolymers are considered as one of the most promising candidates due to the rapidly increased PCEs for their corresponding OSCs. Nevertheless, the free volumes in the block copolymer networks will lead to the chain segment motion49, resulting in difficulty to overcome the trade-off between initial photovoltaic performance and long-term operation stability for OSCs. More importantly, single-component photovoltaic materials suffer from either synthetic challenges or complex purifications.

Herein, we proposed a strategy to develop multicomponent photoactive layers via a facile one-pot polymerization without the complex synthesis and post-treatment, achieving highly efficient and stable OSCs. It should be noted that additional steps and time costs for the polymer purification and device fabrication are obviously reduced in our work. The multicomponent photoactive layer (named S11) is dominated by a block-conjugated polymer PM6-b-L15 featuring a donor1-acceptor1-acceptor2-acceptor3 (d1-a1-a2-a3) backbone architecture. The multicomponent photoactive layer is synthesized by adding the monomers in sequence during polymerization (for details see Supporting Information), followed by simple flash column chromatography to remove unreacted monomers and small amounts of catalysts. Compared to its analogue S9 with a dominant traditional d1-a1-d2-a2 type block polymer, OSCs fabricated from S11 yielded a much higher performance with an optimal PCE approaching 12%, which is enabled by incorporating more electron-deficient building blocks (a.k.a. a units) into the polymer backbones. In parallel, the S11-based devices exhibited much superior light and thermal stabilities (maintaining > 80% of their initial PCEs after over 1000 h) to those of the corresponding single-component and two-component photoactive layer-based devices. Through the careful analysis of the opto-electrical and morphological properties, we assume that the block copolymer interlaced networks with backbone entanglements enabled by block polymer PM6-b-L15 were formed in the S11 multicomponent system. The small amount of PM6 and L15 polymers in S11 matrix will fill into the nanovoids of the interlaced nanostructures, restricting the chain motion of block copolymers. Meanwhile, the networks formed by PM6-b-L15 blocks could in turn suppress the excessive aggregates of PM6 and L15 polymers in the multi-component system. As a result, the favorable frozen film morphology of the networks with well-balanced charge transfer affords an excellent device performance and lifetime prolonging for the OSCs. Hence, this work provides a promising synthetic avenue via one-pot polymerization for the development of the multicomponent photoactive layers in the OSCs with both excellent device efficiency and operational stability.

Results

Material properties

The multicomponent photoactive layers were facilely synthesized via addition polymerization in one pot50,51. The expected chemical structures of main composition in the multicomponent photoactive layers are shown in Fig. 1a. The monomers d1, d2 and a1 are commercially purchased, and the monomers a2 and a3 are synthesized according to our previously reported protocols52 (Supplementary Figs. 1–5). The multicomponent photoactive layers of S9 and S11 are developed in a similar way for block polymers by adding the monomers in sequence. The photographic images showing the progress of one-pot polymerization are depicted in Supplementary Fig. 6. The resulting polymers were treated with simple flash column chromatography for removing the residual palladium catalysts and unreacted monomers. Unlike the time-consuming post-treatment such as Soxhlet extraction, the procedures for material synthesis are evidently simplified. For the synthesis of the conjugated diblock or triblock polymers, a small amount of low Mn oligmers or polymer blocks (PM6 and L15 or PY-IT blocks in our case) are indeed included except the majority of ambipolar block copolymers (PM6-b-PY-IT or PM6-b-L15 here) even in quasi-living polymerization and self-switchable alternating copolymerization for conjugated block polymers53,54. Therefore, the obtained multicomponent polymers S11 (or S9) from the one-pot synthesis method may contain some of PM6, L15 (or PY-IT), and PM6-b-L15 (or PM6-b-PY-IT) due to deactivated polymerization sites, but the block copolymers are in majority50,51. For clear comparison, the purified block copolymers of PM6-b-PY-IT and PM6-b-L15 were also obtained via the reported post-processing method of column chromatography and Soxhlet extraction46,55. The number-average molecular weight (Mn) and polydispersity index (PDI) of the intermediate PM6 block and target block materials measured by gel permeation chromatography (GPC) are summarized in Supplementary Figs. 7–9 and Supplementary Table 1. The Mn of the PM6 block was 6.9 kDa with a PDI of 1.85 while the Mn values of S9 and S11 were 16.3 kDa and 18.6 kDa with PDIs of 2.69 and 2.71, respectively, which confirms that the target products were successfully synthesized. Furthermore, the structural characteristics of PM6 block and the synthesized materials were determined by the hydrogen nuclear magnetic resonance (1H NMR) (Supplementary Figs. 10–15). The molar ratios of PM6 and L15 (or PY-IT) units in the block copolymers and multicomponent materials were estimated by X-ray photoelectron spectroscopy (XPS) and energy-dispersive X-ray spectroscopy (EDX)56. On the basis of the XPS spectra of S9 and S11 in Supplementary Fig. 16, the ratios of F and N concentration are 1:2.70 and 1:3.03, demonstrating that the ratios of PM6 block:PY-IT block and PM6 block:L15 block in the final systems are around 1:0.92 and 1:0.91, respectively. In addition, the molar ratios of the PM6 block:PY-IT block and PM6 block:L15 block were determined to be ~1:0.89 and 1:0.91 in S9 and S11 systems from the EDX data (Supplementary Fig. 17 and Supplementary Table 2), respectively, agreeing well with the XPS results. Moreover, we have further determined the multicomponent composition and structure of S11 via the available technologies such as sequential purification, XPS, and the elemental analysis (EA) (for details see Supporting Information, Supplementary Fig. 18). The mass contents of the dissociative (or residual) PM6 polymer and L15 polymer in S11 were determined to be 1.94% and 5.29% (Supplementary Tables 3 and 4). Thus, our S11 materials are multicomponent mixtures including PM6 polymer (~2%), L15 polymer (~5%) and PM6-b-L15 block polymer (~93%).

Fig. 1: Design, synthesis and opto-electrical properties of multicomponent system S9 and S11.
figure 1

a Chemical structures of S9 and S11 synthesized via one-pot polymerization (for details see Supporting Information). b UV-vis absorption spectra of S9 and S11 in chloroform solution and film. c Energy level diagrams of multicomponent systems and electrodes used in the OSCs.

Full size image

Furthermore, the corresponding UV-vis absorption spectra of S9 and S11 in chloroform solutions and thin films are presented in Fig. 1b. Both films exhibit red-shifted absorption compared to their solutions due to the improved molecular ordering from solution to solid film. As expected, S11 film shows a broad absorption in the range of 400–900 nm with a narrow Eg of ~1.38 eV and contains the obvious features of the corresponding PM6 and L15 moiety, e.g., three distinct absorption peaks at 536, 631, and 802 nm, and no new absorption peaks were recorded. By contrast, S9 film exhibits similar absorption ranges but with lower absorption in 450–600 nm and slightly higher absorption in the wavelength range of 650-800 nm. The LUMO levels (ELUMO) and HOMO levels (EHOMO) were probed through cyclic voltammetry (CV) measurements. As shown in Fig. 1c and Supplementary Table 5, from S9 to S11, the ELUMO/EHOMOs were altered from −3.68/−5.64 to −3.74/−5.71 eV. It is found that the frontier molecule orbital energy levels are slightly increased due to the replacement of thiophene (d2) by bithiophene imide (a3) as co-unit. As we mentioned above, the multicomponent polymers S11 (or S9) from the one-pot method via simple post processing contained a small fraction of PM6, L15 (or PY-IT), but the block copolymers were indeed in majority. As shown in Supplementary Fig. 19, the multiple oxidation and reduction states in CV plots have also corroborated the multicomponent nature.

Photovoltaic performances

To evaluate the photovoltaic performance of multicomponent photoactive layers S9 and S11, OSCs were fabricated with a device structure of ITO/PEDOT:PSS/S9 (or S11)/PNDIT-F3N/Ag. The electron transporting layer PNDIT-F3N used in this work was measured to have a Mn of 4.0 kDa with a PDI of 1.4657 (Supplementary Fig. 20). The devices based on the purified block copolymer of PM6-b-L15 were also prepared as control. As illustrated in Fig. 2a and summarized in Table 1, the S11-based OSCs yield an optimal PCE of 11.78 % with a fill factor (FF) of 70.11%, an open-circuit voltage (Voc) of 0.95 V and a short-circuit currenty (Jsc) of 17.62 mA cm−2. In comparison, the S9-based devices show a much lower PCE of 9.34 % mainly due to the low FF value of 52.15%. In addition, the average PCE values reach 9.05 % and 11.43 % for the S9- and S11-based OSCs, respectively. The PCE values of the S11-based OSCs obtained from 15 different devices show narrow distribution (Fig. 2b), demonstrating good device reproducibility. Moreover, the repeatability of photovoltaic materials is also critical. To this end, to probe into the effects of molecular weight of donor and acceptor blocks on device performances, various batches of S11 are synthesized by changing the reaction time of PM6 (S11-D1, S11-D2 and S11-D3) and L15 blocks (S11-A1, S11-A2, and S11-A3) (Supplementary Figs. 21–26 and Supplementary Tables 6–9). It was found that the variable polymerization time of PM6 donor blocks can exert a significant influence on the material properties and the device performances. As the polymerization time of PM6 block prolonged, the Jsc and FF values increased from Jsc = 11.36 mA cm−2 and FF = 56.47% for the S11-D1-based device to Jsc = 17.51 mA cm−2 and FF = 70.32 % for the S11-D2-based one. As a result, the S11-D2-based OSCs achieved the highest PCE of 11.72% among three multicomponent systems. By contrast, almost little impact on the OSC efficiencies (11.76% for S11-A1, 11.53% for S11-A2 and 11.06% for S11-A3) was observed when the polymerization periods of L15 acceptor blocks were greatly changed but with the fixed optimal reaction time for PM6 blocks (the details can be found in the Supporting Information). Furthermore, the S11-based device with 0.8 cm2 area yielded a PCE of 11.02 % while the analogue with 1.0 cm2 achieved a PCE of 10.80 % (Supplementary Fig. 27), demonstrating its feasibility in large area preparation. It should also be noted that the PCE of 14.3 % was attained for our purified PM6-b-L15 system, which is among the highest PCE values for single component-based OSCs46 (Supplementary Figs. 28, 29 and Supplementary Table 10). The external quantum efficiency (EQE) spectra and integrated Jscs of the optimized OSCs are shown in Fig. 2c. The S11-based devices exhibit stronger photon response from 650 to 800 nm but weaker one from 400 to 600 nm than the S9-based OSCs, consistent with the absorption properties of S9 and S11. The Jscs integrated from the EQE spectra are 18.09 and 17.06 mA cm−2 for the S9 and S11-based OSCs, respectively, which are close to Jsc values from J-V curves with a mismatching below 5%.

Fig. 2: Photovoltaic properties of OSCs.
figure 2

a J–V characteristics for OSCs and the corresponding device structure used in this work. b Histograms of the PCE counts for 15 individual devices. c EQE spectra and integrated Jsc for devices based on S9 and S11. d Transfer characteristics of the S11-based top-gate/bottom-contact (TGBC) organic field-effect transistors (OFETs). e J-V curves of the hole-only and electron-only devices containing the optimal active layer in dark. f Mobility ratios for the corresponding devices.

Full size image
Table 1 Photovoltaic performance parameters of the OSCs incorporating S9 and S11 photoactive layers under the illumination of AM 1.5 G,100 mW cm−2
Full size table

Charge dissociation, transport and collection mechanism

To evaluate the charge transport properties of S9 and S11, top-gate/bottom-contact (TGBC) organic field-effect transistors (OFETs) were then fabricated55. As shown in Fig. 2d, Supplementary Figs. 30, 31 and Supplementary Table 11, both multicomponent photoactive layers exhibit ambipolar transport characteristics. After annealing at elevated temperature to optimize the morphology, S11 shows hole mobility (μh,OFET) of 1.43 × 10−2 cm2  V−1 s−1 and electron mobility (μe,OFET) of 6.19 × 10−3 cm2 V−1 s−1, which is much higher than S9 with a μh,OFET of 5.90 × 10−3 cm2 V−1 s−1 and a μe,OFET of 8.65 × 10−4 cm2 V−1 s−1, suggesting p-type dominant transport behavior. Please note that the μe,OFET of S11 is more than an order of magnitude higher than that of S9, corroborating a larger fraction of acceptor units in S1158,59. The trend correlates well with that of the results obtained from the space-charge limited current (SCLC) method (Fig. 2e and Supplementary Table 12). The SCLC hole and electron mobilities (μh,SCLCs and μe,SCLCs) are measured to be 1.93 × 10−4 and 0.59 × 10−4 cm2 V−1 s−1 for S9, and 3.58 × 10−4 and 2.29 × 10−4 cm2 V−1 s−1 for S11, respectively. The more balanced mobilities with a μh,SCLC/μe,SCLC ratio of 1.56 were obtained in the S11-based devices (Fig. 2f), which explains the higher FF values.

The photocurrent density (Jph) versus the effective applied voltage (Veff) and the relationship between Jsc, Voc and light intensity (Plight) were probed to understand the charge generation and dissociation process. As depicted in Supplementary Fig. 32a, the Jphs reach the saturation value (Jsat) under high Veff (>1.0 V), indicating that each photo-generated exciton dissociated into free electron and hole. The calculated maximum exciton generation rates (Gmax) of the devices are included in Supplementary Table 13. The S11-based device achieved slightly higher Gmax value (1.20 × 1028 m−3 s−1) than the S9-based analogue (1.16 × 1028 m−3 s−1). The exciton dissociation probability (Pdiss) can be obtained from the ratio between Jpha (Jph under short circuit condition) and Jsat, while the the charge collection efficiency (Pcoll) from the ratio between Jphb (Jph under the maximal power output condition) and Jsat. Hence, the Pdiss and Pcoll for the S9-based OSCs are 86.0% and 74.8%, whereas the corresponding values are 91.9% and 83.2% for S11-based ones (Supplementary Table 14), indicating a significantly improved exciton dissociation efficiency in the S11-based OSCs. Photoluminescence (PL) quenching measurements were subsequently conducted (Supplementary Fig. 32b). The S11 film showed a higher PL quenching efficiency of 88%, indicating more efficient charge transfer. In addition, to better understand the charge recombination process, the dependence of J–V characteristics on light intensity (Plight) was studied. The exponential factor (α) close to 1 demonstates negligible bimolecular recombination according to the power-law relevance Jsc∝Plightα 60. As presented in Supplementary Fig. 33, the S11-processed OSC showed a higher α value of 0.961 than the S9-based one (0.852). Therefore, free charges generated in the S11-based devices can be more efficiently transported to the electrodes than those in the S9-based cells. These results demonstrated that the S11-based devices exhibit better exciton dissociation and charge collection, corroborating their higher photovoltaic performance parameters than the S9-based analogues.

Photoactive layer morphology

Atomic force microscopy (AFM) and transmission electron microscopy (TEM) measurements were firstly performed to investigate the film morphological features. As shown in Fig. 3a, b, the root-mean-square (RMS) roughness value of S9 film is 2.2 nm, which is about two fold larger than that of S11 film. Interestingly, the AFM phase (Fig. 3c, d) and TEM images (Fig. 3e, f) of the S11-based film show distinctive fibrillar textures, which should be beneficial to charge generation and transport in the corresponding OSCs. Furthermore, molecular ordering in photoactive layers was investigated through grazing incidence wide angle x-ray scattering (GIWAXS) measurements (Fig. 3g, h and Supplementary Table 15). The S11 film exhibits significantly stronger and sharper scattering peaks in the in-plane (IP) direction (qxy = 0.29 Å−1) and the out-of-plane (OOP) direction (qz = 1.66 Å−1) compared to the S9 film. To explain this difference more quantitatively, we compared the crystal coherence lengths (Lcs) estimated from the Scherrer equation15. The S11-based photoactive film shows increased crystal sizes in both lamellar (20.4 vs.19.9 nm) and π–π stackings (3.85 vs. 2.83 nm) in comparison to the S9 film, leading to enhanced charge transport properties and better performance in the S11-based OSCs61. Furthermore, the photo-induced force microscopy (PiFM) was conducted to better understand the interpenetrating networks for S9 and S11 systems62,63. The PM6:L15 blend film was also prepared for a better comparison. As depicted in Supplementary Fig. 34, the PiFM pattern of PM6:L15 blend film shows uniform but unclear phase separation with a few fibrous aggregates. In contrast, the PiFM images of both S9 and S11 display more pronounced fibrillar nanonetwork. Especially, the longer backbone and more favorable entanglement of S9 and S11 than that of PM6:L15 blend can be observed from the magnified PiFM images with scale bar at 100 nm (Supplementary Fig. 35), confirming the bi-continuous network structures in S9 and S11 multicomponent systems.

Fig. 3: Morphological characteristics of the films.
figure 3

a, b AFM height, c, d phase and e, f TEM images of S9 and S11 films. g 2D GIWAXS patterns and h scattering profiles of out-of-plane and in-plane for their corresponding films.

Full size image

Device stability

The operational stabilities of the multicomponent photoactive layer S11-based OSCs were monitored by tracking the J-V characteristics under dark, illumination, and thermal aging conditions for over 1000 h, respectively. For a better comparison, we also measured the corresponding stabilities of two-component PM6:L15-based devices as the control. Please note that the OSCs based on the PM6:L15 blend were fabricated with the donor: acceptor weight ratio of 1:1.2, achieving a PCE of 15.38%, which is consistent with the previous report (15.20%)52. To probe the differences between fresh PM6:L15 binary blend and S11 multicomponent system, the AFM, TEM, and GIWAXS characterizations were employed to compare the morphology, and the transient absorption spectroscopy (TAS) was correspondingly performed to better understand the hole transfer kinetics in both systems. The detailed discussion is included in Supporting Information (Supplementary Figs. 36–37 and Supplementary Table 16). The periodically monitored photovoltaic performances are depicted in Fig. 4, and the changes in the J-V curves and the parameters are depicted in Supplementary Fig. 38 and Supplementary Tables 17–22. Regarding the shelf-life stability, the S11-based devices exhibit a slight PCE decrease and retain ~90% of original PCEs after 1000 h storage at room temperature (RT), which mainly stems from the burn-in degradation of electrodes and interfaces55. Nevertheless, the analogs with PM6:L15 suffer from ~28% degradation of PCEs during the same intervals. Likewise, 83% of its initial efficiencies were preserved in the multicomponent photoactive layer S11-based OSCs, while the performance for the two-component devices drops to 70 % after light-soaking with the irradiation intensity of 100 mW cm−2 white LED for more than 1000 h. Furthermore, unencapsulated devices under continuous thermal stress of 85 °C demonstrate a similar trend in these two systems. The two-component-based OSCs show dramatically decreased PCE with only 65% of the initial PCEs retained. On the contrary, the multicomponent S11-based OSCs exhibit excellent thermal stability, maintaining 80% of its original PCEs after 1000 h thermal aging. It is found that the multicomponent S11-based OSCs show more stable Jsc and FF values than the binary ones during the same period of device aging. Hence, our multicomponent OSCs reach a better balance between the PCEs and device stability compared to the single-component OSCs (Supplementary Fig. 39). These results unravel the significant effects on mitigating material property and film morphology degradation via the facile one-pot polymerization64,65,66.

Fig. 4: Shelf-life, light and thermal stability.
figure 4

Evolution of photovoltaic parameters of the optimized two-component system PM6:L15- and multicomponent system S11-based OSCs in nitrogen atmosphere for up to 1008 h storage: a under room temperature and dark condition, b under continuous illumination condition, c under 85 °C heated condition.

Full size image

The PM6:L15 and S11-based OSCs show distinct degrees of efficiency drops during the thermal aging, which triggers our interest to focus on the thermal stability in terms of property and morphology evolution for the corresponding photoactive layers. The UV-vis absorption and time-resolved photoluminescence (TRPL) aging tests during 984 h heating (at 85 °C) were tracked at regular intervals. The UV-vis absorption spectra of both photoactive layers before and after heating are shown in Fig. 5a, b. For the PM6:L15 blend film, absorption intensity gradually decreases with prolonged heating time, suggesting a low thermal stability of the two-component film. In contrast, S11 film shows negligible change of absorption spectrum after ~1000 h heating aging. This result indicates that S11 can well retain its initial microstrcuture, which accounts for the enhanced thermal stability of OSCs. Besides, the normalized TRPL decay curves of PM6:L15 and S11 films as a function of storage time are depicted in Fig. 5c, d. The fluorescence lifetime (τ) of the PM6:L15 photoactive layer is calculated to be 540 ps for 0 h and 270 ps for 984 h. The shortened lifetime is not conducive to the extion diffusion in the D/A interfaces, leading to a sharp decline of its device performance. Impressively, the multicomponent S11 film holds almost unchanged fluorescence lifetime with a τ of 360 ps for 0 h and τ of 310 ps for 984 h, ensuring efficient exciton diffusion between D/A interfaces, and, thereby yielding more stable Jsc and FF67,68.

Fig. 5: Absorption property and fluorescence lifetime as a function of aging time.
figure 5

Normalized UV-vis absorption spectra of aged a two-component system PM6:L15 and b multicomponent system S11 films with different heating duration. Normalized TRPL decay curves of c PM6:L15 and d S11 films as a function of storage time.

Full size image

To gain insights into the correlation between exciton diffusion decay and BHJ microstructure changes, the AFM and TEM images for thin films were obtained under 984 h continuous thermal aging (Fig. 6a–f, i–n). All the fresh films based on PM6:L15 and S11 exhibit distinctive fibrous nanostructures69. After over 1000 h heating, RMS roughness of the PM6:L15 blend significantly increased from 0.95 to 1.32 nm, and excessive large-sized polymer aggregates emerged in the AFM images for the aged two-component system. In contrast, the multi-component system showed no change in the RMS roughness and the same BHJ homogeneity after annealing. The negligible change has been further confirmed by TEM measurement, which proves that the presence of block polymers can effectively suppress the further aggregation of a handful of D and A in S11. To reveal the morphology stability of our multi-component system more intuitively, the PiFM technique was accordingly conducted. After 716 h thermal annealing, the multicomponent S11 film retains distinguishable phase separation and distinctive interpenetrating nanonetwork of the PM6-b-L15 block copolymers with a small fraction of PM6 and L15 surrounding the block copolymer domains, which facilitates the exciton dissociation and charge transport70. In contrast, the PM6:L15 blend film showed larger donor and acceptor domains with oversized phase separation (Supplementary Fig. 40). Therefore, segregated domains with decreased D/A interfaces led to seriously deteriorated exciton dissociation, which was confirmed by the fluorescence decay dynamics (Fig. 6g, h). Consequently, a rapid drop in device efficiency was observed in the aged two-component film.

Fig. 6: Morphology evolution during the aging test.
figure 6

The OSCs based on PM6:L15 and S11 active layers under continuous heating of 984 h: a, d, i, l AFM height, b, e, j, m phase and c, f, k, n TEM and combined PiFM images for two-component system PM6:L15 and multicomponent system S11 films before and after aging, respectively. Schematic illustration of morphology changes for g, h two-component system and o, p multicomponent system before and after thermal aging, respectively.

Full size image

The stability of the purified PM6-b-L15-based devices was also periodically monitored and founded to be far lower than that of our multicomponent system after 1000 h accelerated thermal aging (71% vs. 80% maintained, Fig. 7). We considered it is mainly due to morphological instability caused by the chain motion through the free volumes between the block copolymer networks in single PM6-b-L15 system49. Thus, the contact angle analysis was then implemented to investigate the possible intermolecular interactions in the S11 multicomponent system (Supplementary Fig. 41). The calculated Flory-Huggins interaction parameter (χ) of PM6:PM6-b-L15 is very close to that of L15:PM6-b-L15 (Supplementary Table 23), indicating a good miscibility for these combinations71. Therefore, the PM6 and L15 polymer tended to interact with PM6 block and L15 block of PM6-b-L15 backbones in S11 system, respectively. Driven by these favorable interactions, the individual PM6 or L15 polymers can easily fill into the nanovoids of the PM6-b-L15 block copolymer interlaced networks, restricting the molecular motion of block copolymers56. Based on the above considerations, we assume that the abundant block polymers with long d1-a1-a2-a3 backbones firstly tend to self-assemble and form fibrillar nanonetwork structures, freezing the film morphology by embeding the interconnected D and A components into these dense networks (Fig. 6o). Under continuous thermal annealing, the presence of block polymer networks can effectively suppress the migration of D and A matrix (Fig. 6p). Meanwhile, the small amount of residual PM6 and L15 polymers in S11 system can favorably fill free volumes of the entangled block copolymer backbones, thus achieving more compact molecular packing and enhanced morphology stability even under continuous thermal annealing conditions56 and, therefore, realizing the efficient charge transfer in the multicomponent system72,73,74. With the synergistic effects of these two factors, reduced charge recombination and facilitated exciton dissociation can be sustained during the aging. Hence, the presence of dominant block polymer networks and small fraction of PM6 and L15 polymers in S11 can affect both the surface and interface characteristics of photoactive layers, and improve the long-term stability of OSCs in a prolonged aging period. To further confirm the assumption, a small amount of polymers PM6 (2%) and L15 (5%) polymers were intentionally incorporated into the purified block copolymer PM6-b-L15 as the multicomponent active layer with the exact composition and structure to fabricate the OSC devices. The stability results after 1000 h accelerated thermal aging for the PM6-b-L15:PM6:L15 system were comparable to the S11 system prepared by our one-pot method (78% vs. 80% retained, Fig. 7). Apparently, our one-pot method is simpler and lower-cost for the device fabrication than the traditional multicomponent system. This result highlights the feasibility of one-pot method to boost the long-term stability of the OSC devices.

Fig. 7: The Comparison of thermal stability.
figure 7

Evolution of a Voc, Jsc, and b FF, PCE c retained photovoltaic parameters of the optimized single component system PM6-b-L15-, the ternary system PM6-b-L15:PM6:L15- and the multi-component system S11-based OSCs under 85 °C heated condition.

Full size image

Discussion

In summary, we demonstrated an effective molecular strategy based on an one-pot polymerization for improving the stability of OSCs, reducing synthetic cost and simplifying the device fabrication procedure. The multicomponent photoactive layer S11 is developed in a similar way for block polymers by adding the monomers in sequence, but only purified by simple flash column chromatography to remove small amounts of residual catalyst and monomers without complex post-treatment. The dominant block polymer component in the S11 features a d1-a1-a2-a3 backbone architecture. Compared to its analogue polymer with an alternative d-a backbone structure, incorporating more A unit into the polymer enables higher and more balanced electron and hole mobilities, achieving a decent PCE of 11.8% for the OSCs. More importantly, the multicomponent S11-based OSCs exhibit much more superior device stability than the corresponding two-component ones and the reported single-component OSCs, which endows a better trade-off between PCE and lifetime compared to single-component OSCs, i.e. maintaining > 80% of their initial PCEs even after continuous illumination and thermal annealing for over 1000 h. The systematic investigations on the opto-electrical and morphological evolution as a function of aging time were performed. We assume the majority of block copolymer interlaced networks and a small amount of PM6 and L15 polymers contribute to the excellent stability: (i) the formation of excessive aggregates in the photoactive layer is effectively suppressed while maintaining fibrillar percolations in the D and A domains enabled by the network nanostructures; (ii) the chain backbone motion is significantly restricted by the individual PM6 or L15 polymers packed into the nanovoids of the block copolymer interpenetrated networks. Moreover, the physically mixed PM6:L15:PM6-b-L15 multiple components with the exact composition were established to further prove our assumption. The stability results achieved in the PM6:L15:PM6-b-L15 multicomponent system were also comparable to the S11 system prepared by this one-pot method. Our one-pot method is simpler and cost-effective for device fabrication, and this work provides an important and practicable strategy to develop highly efficient and operationally stable OSCs toward industrial viability.

Methods

Materials

All chemicals and reagents are commercially purchased from vendors and are used without further purification unless otherwise stated. Anhydrous toluene is distilled from Na/benzophenone under argon. The synthesis of Compound a (Supporting Information) is based on the previously reported procedures52. Compound IC-Br (Supplementary Fig. 2) is purchased from Hyperchemical, Inc. Unless otherwise stated, all reactions are performed in an inert atmosphere using standard Schlenk-line techniques.

Fabrication of the OSCs

OSCs with a conventional structure of ITO/PEDOT: PSS/active layers/PNDIT-F3N/Ag are fabricated. The pre-cleaned ITO-coated glass with a sheet resistance of 12 Ω sq−1 is used as the substrate and UV-ozone (BZS250GF-TC, HWOTECH, Shenzhen) treatment for 15 min before utilization. PEDOT: PSS (Clevios P VP A1 4083) was spin-coated onto the UV-ozone-treated ITO substrates at 3000 rpm for 40 s and then annealed at 150 °C for 15 min, forming a ~35 nm film. For S9- and S11-based OSCs, the total concentration was fixed at 16 mg mL−1, and the films were obtained by spin coating solutions of the polymers in CF containing 1-chloronaphthalene (1-CN, 4 vol%). The solutions were stirred at room temperature for 6 h and heated at 45 °C for about 1 h before use. In addition, the control devices based on PM6:L15 blend processed with CF solvent were also fabricated for comparison. The photoactive layers with an optimal thickness of 100~110 nm are annealed at 100 °C for 5 min. Then PNDIT-F3N dissolved in methanol at the concentration of 1.0 mg mL−1 was spin-coated at 3500 rpm for 40 s on the photoactive layers. 100 nm Ag is sequentially deposited on the photoactive layers to complete the devices via thermal evaporation (ca. 1 × 10−5 Pa).

Characterization of the OSCs

All current-voltage (J-V) characteristics of the devices were measured under simulated AM1.5 G irradiation (100 mW/cm2) using a Xe lamp-based SS-F5-3A Solar Simulator (Enli Technology, Inc.). A Xe lamp equipped with an AM1.5 G filter was used as the white light source. The light intensity was controlled with an NREL-calibrated Si solar cell with a KG-5 filter. The external quantum efficiency (EQE) was measured by a QE-R3011 measurement system (Enli Technology, Inc.).

AFM measurements

AFM measurements of two- or multicomponent photoactive layers were conducted using a Dimension Icon Scanning Probe Microscope (Asylum Research, MFP-3D-Stand Alone) in the tapping mode.

TEM measurements

TEM micrographs were obtained on a Tecnai G2 F30 microscope operating at 300 kV.

GIWAXS measurements

GIWAXS measurements were carried out at the PLS-II 9 A U-SAXS beamline of Pohang Accelerator Laboratory, Korea.

Reporting Summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.