A >25 kDa Star‐Shaped Dodecamer Acceptor as a Heterogeneous Nucleation Agent Enables 20.50%‐Efficient Organic Solar Cells

ABSTRACT

Star‐shaped molecules possess strong cohesive energy and welldefined molecular weights compared to polymers, making them promising additives for regulating microphase separation in bulk‐heterojunction blends. Herein, we report the synthesis of an ultrahigh–molecular‐weight star‐shaped electron acceptor that serves as a heterogeneous nucleation agent to precisely control donor/acceptor blend morphology for high‐performance organic solar cells (OSCs). The star‐shaped acceptor, denoted SP12, was synthesized via an arm‐first strategy and comprises 12 Y‐type acceptor arms, affording a molecular weight exceeding 25 kDa—the highest reported to date for star‐shaped electron acceptors. Owing to its pronounced pre‐aggregation behavior in solution and three‐dimensional (3D) architecture, SP12 acts as an effective heterogeneous nucleation center during film formation. This leads to bulk‐heterojunction thin films with enlarged fiber diameters and enhanced crystallinity, as confirmed by multiple advanced characterization techniques. Consequently, OSCs incorporating SP12 as an additive deliver a high‐power conversion efficiency of 20.50% in green‐solvent‐processed devices. Moreover, the high glass transition temperature and 3D topology of SP12 kinetically stabilize the optimized morphology by suppressing acceptor diffusion and re‐aggregation, resulting in markedly improved morphological and operational stability. This work establishes a topology‐driven molecular design and morphology‐engineering strategy for achieving efficient and stable organic photovoltaic devices.

A star‐shaped dodecamer acceptor SP12 with 12 Y‐type arms is synthesized via an arm‐first strategy. Acting as a heterogeneous nucleation additive, SP12 precisely regulates bulk‐heterojunction morphology, enabling 20.50% efficiency and enhanced stability in organic solar cells.

1. Introduction

Organic solar cells (OSCs) have emerged as a promising photovoltaic technology in the pursuit of sustainable energy owing to their intrinsic advantages, including mechanical flexibility, low weight, semi‐transparency, and compatibility with solution‐based manufacturing [1, 2, 3, 4, 5]. These attributes endow OSCs with unique application potential beyond conventional photovoltaics, such as building‐integrated systems, portable power supplies, and wearable electronics [6, 7, 8]. In recent years, the rapid development of non‐fullerene acceptors (NFAs) has fundamentally reshaped the performance landscape of OSCs, driving the power conversion efficiency (PCE) of single‐junction devices beyond the 20% benchmark [9, 10, 11, 12, 13, 14, 15]. While this milestone signifies the arrival of a high‐efficiency era for organic photovoltaics, the translation of laboratory‐scale performance into commercial viability remains impeded by two long‐standing challenges: the precise control of nanoscale active‐layer morphology and the achievement of long‐term operational stability under realistic environmental conditions [9, 16].

The performance of OSCs is governed by the bulk heterojunction (BHJ) active layer, which consists of an interpenetrating network of electron donor (D) and acceptor (A) materials. An ideal BHJ morphology requires a finely balanced phase separation on the length scale of the exciton diffusion length (∼10–20 nm) to maximize donor–acceptor interfacial area for efficient exciton dissociation [17, 18]. Concurrently, both phases must exhibit high molecular ordering and crystallinity to facilitate efficient charge transport and suppress non‐radiative recombination losses [19, 20, 21, 22]. However, the solution deposition of BHJ films is inherently a non‐equilibrium kinetic process, often resulting in metastable morphologies characterized by inhomogeneous domain sizes and disordered molecular packing [23, 24, 25]. Such morphological disorder introduces substantial energetic disorder, which is a major origin of a non‐radiative voltage loss and fundamentally limits the open‐circuit voltage (V _OC) and fill factor (FF) [26]. Moreover, metastable morphologies are intrinsically vulnerable to thermal, optical, and environmental stresses, leading to progressive phase coarsening, molecular diffusion, and ultimately rapid device degradation [27].

To address these issues, extensive efforts have focused on external morphology‐regulation strategies, including thermal annealing and solvent additive engineering. Although effective in laboratory settings, these approaches are highly sensitive to processing conditions, often suffer from limited reproducibility in large‐area manufacturing, and may introduce long‐term instability due to residual additives [25, 28, 29, 30]. Consequently, there is a growing consensus that, intrinsic material design—where the molecular structure itself governs film‐forming kinetics and final morphology—represents a more robust and fundamentally scalable solution for achieving simultaneously high efficiency and stability.

From the perspective of polymer physics, the morphological evolution of BHJ films can be understood as a coupled phase separation and crystallization process in a multicomponent polymer/small‐molecule blend under non‐equilibrium conditions [24]. According to the Flory‐Huggins theory [31, 32, 33], the miscibility and phase behavior of such blends are dictated by the mixing free energy (ΔG _m = ΔH _m—TΔS _m, here ΔG _m, ΔH _m, T, and ΔS _m are the mixing Gibbs free energy, enthalpy, temperature, and entropy, respectively). In typical NFA‐based OSCs, the small‐molecule nature of acceptors results in a large ΔS _m, favoring excessive mixing during film formation. This entropy‐driven overmixed state is thermodynamically unstable and tends to undergo spontaneous demixing over time, thereby compromising long‐term morphological and device stability [23, 34, 35, 36].

In this context, conjugated star‐shaped polymers offer a compelling yet relatively unexplored topological alternative to conventional linear materials. Their unique architecture, consisting of multiple arms emanating from a central core, imposes pronounced conformational entropy constraints [37, 38]. During blending, this star‐shaped topology induces a distinct “shielding effect,” which effectively reduces unfavorable ΔH _m and entropy ΔS _m relative to small‐molecule acceptors, thereby, promoting more controlled phase separation and morphologies closer to thermodynamic steady states. In addition, the compact molecular conformation and multidimensional π–π interactions intrinsic to star‐shaped systems favor the formation of well‐defined pre‐aggregated nuclei in solution [39, 40, 41, 42, 43]. This characteristic provides a natural foundation for implementing heterogeneous nucleation within the BHJ (Figure 1a), a highly effective yet underutilized mechanism for morphology control. When incorporated as built‐in nucleating agents, star‐shaped conjugated polymers can offer several synergistic advantages. First, their symmetric multi‐arm topology and high molecular weight promote strong, temperature‐dependent pre‐aggregation in green solvents [44], which can serve as stable nucleation templates for epitaxial and ordered growth of the host acceptor matrix, leading to enlarged crystalline domains and enhanced molecular ordering. Second, the multiple π‐conjugated arms function as a multi‐point anchoring platform [42], enabling efficient capture and guided assembly of acceptor molecules while simultaneously modulating donor–acceptor interfacial energy to achieve optimized phase separation. Third, the typically elevated glass transition temperature (T _g) associated with star‐shaped architectures kinetically “freezes” the optimized morphology [39, 40, 42], effectively suppressing molecular diffusion and re‐aggregation of small‐molecule acceptors during device operation (Figure 1a), thereby markedly improving morphological and operational stability.

FIGURE 1.

(a) Schematic illustration of the star‐shaped topology and its role in BHJ morphology control. The star‐shaped architecture acts as a heterogeneous nucleating agent, guiding the formation of an optimal phase‐separated network with minimized isolated domains, stabilized film morphology, and enhanced overall crystallinity and device stability. (b) Chemical structure of the synthesized star‐shaped dodecamer acceptor, SP12. For clarity, the long alkyl chains are denoted as R₁ and R₂, and the electron‐withdrawing end groups are abbreviated as IC. (c) Correlation between PCE and molecular weight for reported star‐shaped acceptors in OSCs. The performance of the device based on SP12 (this work) is highlighted with a red star.

Guided by these considerations, we herein design and synthesize a 12‐arm Y‐type acceptor (SP12) with a well‐defined star‐shaped topology via an “Arm‐First” synthetic strategy (Figure 1b and Scheme S1) [37]. SP12 is deliberately engineered to function as an efficient built‐in heterogeneous nucleating agent in OSCs. Its star‐shaped topology induces a pronounced “shielding effect” during blending, enabling precise regulation of phase separation and crystallinity, while the multi‐arm structure provides a multi‐point anchoring platform to guide ordered molecular assembly and shorten nucleation time through strong solution pre‐aggregation. As a result, ternary OSCs based on D18:BTP‐eC9:SP12 achieve a high PCE of 20.50% in green‐solvent‐processed devices, which is among the highest PCEs based on star‐shaped electron acceptors with the highest molecular weight (Figure 1c). Furthermore, the three‐dimensional (3D) topology and high T _g of SP12 effectively stabilize the optimized morphology, suppressing detrimental acceptor diffusion and aggregation and leading to significantly enhanced device stability. This work establishes a topology‐driven molecular design paradigm and morphology engineering strategy for next‐generation organic photovoltaic materials that combine high efficiency, superior stability, and excellent processability.

2. Results and Discussion

The star‐shaped dodecamer acceptor SP12 was synthesized via a modular “arm‐first” approach, as schematically illustrated in Figure 1b and Scheme S1. The synthetic route commenced with the preparation of the key Y‐shaped arm precursor. Specifically, the intermediate M2 was synthesized via a dehydration reaction following established literature procedures [45, 46, 47]. Subsequently, a three‐step protocol was employed to construct the hydroxy‐terminated 2Y‐arm: (i) Williamson ether synthesis between M2 and 1,3,5‐tris(bromomethyl)benzene, followed by (ii) a Vilsmeier–Haack reaction to introduce the aldehyde functionality on the side unit, (iii) Suzuki coupling with 4‐hydroxyphenylboronic acid. The core‐arm coupling, identified as the pivotal step for constructing the star‐shaped architecture, was efficiently achieved through another Williamson ether synthesis, reacting the 2Y‐arm with a brominated‐octyl terminated core molecule. Finally, the electron‐accepting end‐groups were installed Knoevenagel condensation with 2‐(5,6‐difluoro‐3‐oxo‐2,3‐dihydro‐1H‐inden‐1‐ylidene)malononitrile (difluorinated IC), yielding the target molecule SP12 [48, 49]. The successful synthesis and high purity of SP12 were unambiguously confirmed by comprehensive spectroscopic characterization. Nuclear magnetic resonance (NMR) spectroscopy revealed a highly symmetric structure, as evidenced by the sharp singlet peak at 5.43 ppm in the ¹H NMR spectrum, which corresponds to the equivalent protons in the two‐Y segments of each arm. Furthermore, high‐resolution matrix‐assisted laser desorption/ionization time‐of‐flight (MALDI‐TOF) mass spectrometry showed a dominant peak at m/z 25,704.28, which matches the theoretical molecular weight of SP12 (Figure S31). The final product exhibited excellent solubility in common organic solvents such as chloroform and chlorobenzene, which is crucial for solution processing. Detailed synthetic procedures, reaction conditions, yields for each step, and full spectroscopic data (NMR and MS) for all intermediates and final products are provided in the Supporting Information to ensure reproducibility. Moreover, we have summarized the reported star‐shaped acceptors in OSCs, comparing their PCE against molecular weight (Figure 1c and Table S1). Notably, SP12, with a record‐high molecular weight exceeding 25 kDa, demonstrates state‐of‐the‐art photovoltaic performance among high‐molecular‐weight star‐shaped acceptors.

The molecular weight of SP12 was further characterized via high‐temperature gel permeation chromatography (GPC) at 150° C using 1,2,4‐trichlorobenzene as the eluent and polystyrene (PS) standards for calibration. As presented in (Figures 2a, S3, and Table S2), the number‐average molecular weight (M _n), mass‐average molecular weight (M _w), and the polydispersity index (Đ) were determined to be 16.4, 17.5, and 1.07 kDa, respectively. The measured M _n and M _w values are lower than the theoretical molecular weight and the result obtained from the MALDI‐TOF MS measurement. This discrepancy can be attributed to the denser, more compact spherical or ellipsoidal conformation of the star‐shaped polymer, which leads to a smaller hydrodynamic volume and thus a greater deviation in the apparent molecular weight determined by GPC [37].

FIGURE 2.

(a) High‐temperature GPC trace of SP12 measured in TCB at 150 °C using PS standards. (b) UV‐vis‐NIR absorption spectra of BTP‐eC9 and SP12 in chlorobenzene solution (dashed lines) and as thin films (solid lines). (c) Temperature‐dependent UV‐vis‐NIR absorption spectra of SP12 in chlorobenzene solution (20 to 140° C). (d) Two‐dimensional GIWAXS pattern of an as‐cast SP12 film processed from o‐xylene/carbon disulfide (o‐XY/CS₂) solution. The diffuse halo at the (100) lamellar stacking position indicates an amorphous packing motif. (e) DMTA curves of an SP12 film supported on glass fibers. The peak in the tanδ curve at 183.8° C corresponds to the T _g of SP12. (f) Energy level diagram illustrating the HOMO and LUMO levels of the donor polymer D18, the small‐molecule acceptor BTP‐eC9, and the star‐shaped acceptor SP12, as determined by CV measurement.

The ultraviolet‐visible‐near‐infrared (UV‐vis‐NIR) absorption spectra of BTP‐eC9 and SP12 acceptors in chlorobenzene (CB) solutions and as thin films are compared in Figure 2b. In CB solution, comparing with BTP‐eC9, despite possessing a similar degree of conjugation, SP12 exhibits a more pronounced red‐shifted and broadened absorption band in the range of approximately 600–850 nm, with the maximum absorption peak (λ _max) and onset absorption (λ _onset) observed at 764 nm and 841 nm, respectively. This spectral feature indicates pre‐aggregation behavior of the star‐shaped polymer in solution, likely due to its inherently high local chain segment density and restricted conformational entropy [37, 44]. When transforming from solutions to films, SP12 exhibits a smaller absorption red shift (∼50 nm) compared to that of the linear acceptor BTP‐eC9 (∼100 nm). This difference suggests that the crowded, 3D topology of the star‐shaped architecture alters molecular aggregation and packing motif [50]. The pre‐aggregation behavior is further corroborated by the temperature‐dependent absorption spectra of SP12 in chlorobenzene solution (Figure 2c), where a blue shift is observed with increasing temperature, indicative of disaggregation. Besides, the absorption spectra of the blended acceptor films were investigated, as shown in Figure S4a. Rather than displaying a simple superposition of individual peaks, the spectra exhibit a continuous blue‐shift toward the SP12 absorption feature with increasing SP12 weight ratio. This optical behavior strongly implies the formation of an alloy‐like phase between the two acceptors. The molecular packing of the as‐cast SP12 film was investigated using grazing‐incidence wide‐angle x‐ray scattering (GIWAXS), as shown in Figure 2d. The analysis reveals a mixed “face‐on” and “edge‐on” orientation. A (010) π–π stacking diffraction signal is observed in the out‐of‐plane (OOP) direction at an azimuthal angle of ∼45°, while the (100) lamellar packing exhibits a diffuse halo spanning from the OOP to the in‐plane (IP) direction, indicating the predominantly amorphous nature of the SP12 film.

The mechanical and thermal properties of the SP12 film were characterized by dynamic mechanical thermal analysis (DMTA), as illustrated in Figure 2e. The results indicate that the film exhibits predominantly elastic behavior, evidenced by the storage modulus (G′) being approximately one order of magnitude higher than the loss modulus (G′′) across the measured temperature range. The T _g was identified from the peak of the tanδ (tanδ = G′′/G′) curve. While a peak around 50 °C corresponds to the T _g of the glass fiber substrate, a distinct peak at 183.8 °C is assigned to the T _g of the SP12 film. To further corroborate this thermal property, deviation matric method was additionally employed (Figure S4b) [51]. The thermally induced spectral evolution revealed a structural transition point at approximately 190 °C, which is highly consistent with the DMTA measurement. This value is significantly higher than those reported for other Y‐series acceptors [46, 47, 52, 53]. Electrochemical cyclic voltammetry (CV) measurements were performed to determine the frontier orbital energy levels of SP12 (Figure S5). The highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) energy levels were calculated to be −5.33 and −3.89 eV, respectively (Figure 2f). These energy levels are not well‐aligned with those of the donor polymer D18, specifically the mismatched HOMO levels, which may lead to insufficient driving force for charge separation.

To elucidate the dynamic structural evolution during film formation, in situ UV‐vis‐NIR spectroscopy was employed to monitor the solution‐to‐film transformation process of the photoactive blends in o‐xylene/carbon disulfide (o‐XY/CS₂) solvent, as shown in Figures 3a–f. Figures 3a–c display the contour plots of the spectral evolution for the D18:SP12, D18:BTP‐eC9, and D18:BTP‐eC9:SP12 blends, respectively. The transformation can be delineated into three distinct stages: initial dissolution, aggregation during solvent evaporation, and final solid‐film formation. The absorption peak of the D18 donor polymer in the 500–600 nm region exhibited negligible spectral shift over time, consistent with its pre‐aggregation behavior in solution [52, 54]. In stark contrast, the absorption bands corresponding to the acceptors (700–850 nm) displayed a pronounced bathochromic shift, signaling the progressive molecular assembly and crystallization of the acceptors during solvent drying. The kinetics of this process were quantified by tracking the temporal shift of the acceptor's peak wavelength, as summarized in Figure 3d–f. The D18:SP12 binary blend completed its spectral transition (from ∼770 to ∼795 nm) within approximately 1560 ms, significantly faster than the 3000 ms required for the D18:BTP‐eC9 blend. This accelerated aggregation kinetics for SP12 can be attributed to its compact, 3D star‐shaped architecture, which promotes pre‐aggregation in solution via enhanced multi‐dimensional π–π interactions [39, 40, 41, 42, 43]. Notably, for the D18:BTP‐eC9:SP12 ternary blend, the transformation time was recorded at 2040 ms—an intermediate value between the two binary systems. Based on these observations, we propose that the star‐shaped SP12 macromolecule acts as a heterogeneous nucleating agent within the ternary blend. Its pre‐formed, ordered aggregates provide a template for the epitaxial growth of BTP‐eC9 molecules. The multiple conjugated arms of SP12 serve as a multi‐point anchoring platform, efficiently capturing BTP‐eC9 and guiding its assembly into a more ordered structure. This proposed mechanism is schematically illustrated in Figure 3g: Initially, in the solution state, the components are molecularly dispersed. As solvent evaporation commences, aggregation initiates in the sequence of D18>SP12>BTP‐eC9, leading to the preferential formation of a D18/SP12 interface. Subsequently, SP12 aggregates function as nucleation sites, templating the ordered assembly of BTP‐eC9. This kinetic evolution is highly consistent with the thermodynamic compatibility of the system. According to the Flory‐Huggins interaction parameters (χ) derived from contact angle measurements (Figure S7 and Table S4), the χ value between SP12 and BTP‐eC9 is merely 0.28 k, which is significantly lower than that between SP12 and D18 (χ = 3.37k). Such a pronounced thermodynamic contrast dictates that SP12 will be excluded from the D18 donor phase and tightly intermix with BTP‐eC9 to form an alloy‐like phase. In the final solid film, SP12 is postulated to reside at the D18/BTP‐eC9 interface, acting as a molecular adhesive that enhances the continuity of the acceptor phase and improves the overall morphological stability of the blend [55]. Moreover, the star‐shaped acceptor SP12 exhibits lower miscibility with the D18 donor compared to BTP‐eC9 (with relative values of 3.37k versus 1.71k, respectively). This indicates that the introduction of SP12 does not thermodynamically stabilize the system by increasing phase miscibility. Instead, the stability enhancement is governed by kinetic arrest due to high T _g value according to the Ade‐O'Connor‐Ghasemi [23, 36] model.

FIGURE 3.

The waterfall plots of spectral evolution during the film‐drying process for (a) D18:SP12, (b) D18:BTP‐eC9, and (c) D18:BTP‐eC9:SP12 blends from o‐XY/CS₂ solution, respectively. The derived fitting results from the waterfall plots for (d) D18:SP12, (e) D18:BTP‐eC9, and (f) D18:BTP‐eC9:SP12 with simulated drying times of 1,560, 3,000, and 2,040 ms, respectively. (g) Schematic diagrams of the film‐forming process of the ternary D18:BTP‐eC9:SP12 blends.

The film morphology and molecular packing of these blends were analyzed using atomic force microscopy (AFM) in tapping mode and the nanoscale infrared spectroscopy coupled with AFM (nano IR‐AFM) measurements [56]. As shown in the Figures 4a, c and S8, a fibrillar network morphology was observed in both binary and ternary blends with the root‐mean‐square (RMS) roughness of 0.77 and 0.72 nm for D18:BTP‐eC9 and D18:BTP‐eC9:SP12 films, respectively. And an increased fiber‐like grain size was observed in D18:BTP‐eC9:SP12 film comparing with D18:BTP‐eC9 film in Figure S9. Then, we recorded nano IR‐AFM signals of blended films by detecting wavenumber at 1696 cm⁻¹, which is the stretching vibration of the C═O bond in the acceptors [14, 56]. As shown in the Figure 4b, d, it is obvious that interpenetrating fibrillary networks at nanoscale can be observed, which are conducive to effective charge carrier transport, and thus enable the devices to achieve higher J _SC and FF. And an increased phase separation was observed in D18:BTP‐eC9:SP12 film comparing with D18:BTP‐eC9 film. The GIWAXS analysis of the molecular orientation and packing motif have been shown in Figure 4e–g, in which both D18:BTP‐eC9 and D18:BTP‐eC9:SP12 blended films exhibited a “face‐on” molecular orientation, which is indicative of anisotropic molecular alignment. However, D18:SP12 exhibited similar molecular orientation with the pristine SP12 film in Figure 2d. The corresponding line‐cuts extracted from the IP and OOP directions for D18:SP12, D18:BTP‐eC9, and D18:BTP‐eC9:SP12 blended films have been shown in Figure 4h. The D18:SP12 exhibited a larger (100) laminar d‐spacing of 21.44 Å at q = 0.293 Å⁻¹ than 20.11 Å at q = 0.312 Å⁻¹ of the D18:BTP‐eC9 blended film in the IP direction due to the large molecular size of SP12. Sequentially, the ternary D18:BTP‐eC9:SP12 exhibited a moderate (100) laminar d‐spacing of 20.29 Å at q = 0.310 Å⁻¹ in the IP direction. Similarly, the introduction of SP12 into D18:BTP‐eC9 blends slightly increase the (010) π–π stacking from 3.68 to 3.69 Å at the OOP direction. Then, we have calculated the coherence lengths (CCL) for (100) and (010) peaks using the Scherrer formula [57, 58], as shown in Table S5. The CCL values increased from 57.38 Å for D18:BTP‐eC9 to 61.16 Å for D18:BTP‐eC9:SP12 films of (100) peak and from 16.67 Å for D18:BTP‐eC9 to 16.69 Å for D18:BTP‐eC9:SP12 films of (010) peak, respectively. The increased grain sizes in ternary blended films are ascribed to the heterogeneous crystallization kinetics through the introduction of star‐shaped dodecamer acceptor SP12. Then, we have calculated the relative degree of crystallinity (rDoC) for these blended films by integrating intensity versus azimuth for specific diffraction peaks, as shown in Figures 4i, S10. The rDoC for both (100) and (010) diffraction peaks also increase from D18:BTP‐eC9 to D18:BTP‐eC9:SP12 blends. The enhanced crystallinity and lattice rigidity are expected to mitigate charge recombination decay, thereby unlocking simultaneous improvements in efficiency, charge dynamics, and operational stability. Finally, the grazing incidence small‐angle x‐ray scattering (GISAXS) has been employed to acquire the nanoscale structural information of these blended films [59], as shown in Figure 4j and Table S6. The 2D‐scattering pattern of these thin films measured at an incidence angle of 0.15° was shown in Figure S11 and the horizontal IP linecuts intensity profiles at the Yoneda peak positions of each incident angle. Thus, the domain sizes (2R _g) in the IP direction estimated from the GISAXS IP profiles are 30.24 and 32.23 nm for D18:BTP‐eC9 and D18:BTP‐eC9:SP12 films, respectively, which is consistent with the observation from AFM measurement.

FIGURE 4.

The AFM phase images (1 × 1 µm) and AFM‐IR images of (a‐b) D18:BTP‐eC9 and (c‐d) D18:BTP‐eC9:SP12 blended films, respectively. The GIWAXS patterns of (e) D18:SP12, (f) D18:BTP‐eC9, and (g) D18:BTP‐eC9:SP12 blended films, respectively. (h) The corresponding line‐cuts extracted from the IP and OOP directions for D18:SP12, D18:BTP‐eC9, and D18:BTP‐eC9:SP12 blended films, respectively. (i) Comparison of the rDoC values extracted from the laminar (100) and π–π stacking (010) diffraction peaks of the blend films. Normalize the rDoC to 1 for films with D18:BTP‐eC9:SP12 blended films. (j) 1D GISAXS intensity linecuts along the IP direction extracted at the Yoneda peak for D18:BTP‐eC9 and D18:BTP‐eC9:SP12 blended films with domain sizes 2R _g in the legend.

The photovoltaic performance of the star‐shaped acceptor SP12 was evaluated using a conventional device architecture of ITO/2PACz [60]/D18:acceptors/PNDIT‐F3N [61]/Ag. All photoactive layers were processed from a mixed green solvent system of o‐XY/CS₂, enabled by the excellent solubility of SP12 in environmentally benign solvents. The current density‐voltage (J‐V) characteristics and corresponding photovoltaic parameters of the D18:acceptor‐based OSCs are presented in Figure 5a and summarized in Tables 1 and S7–S9. Binary OSCs based on D18:SP12 delivered a maximum PCE of 7.58%, with a relatively high V _OC of 0.964 V, a short‐circuit current density (J _SC) of 13.92 mA/cm², and an FF of 56.45% (Figure S12). The modest PCE of the D18:SP12 binary system is likely associated with the suboptimal HOMO energy‐level alignment between D18 and SP12, which limits efficient hole transfer. In contrast, the benchmark binary D18:BTP‐eC9 devices exhibited a significantly higher PCE of 19.49%, with a V _OC of 0.861 V, a J _SC of 28.53 mA/cm², and an FF value of 79.33% (Figure 5a and Table 1). Upon incorporating SP12 as a third component into the D18:BTP‐eC9 system, a champion PCE of 20.50% was achieved, accompanied by simultaneous enhancements in V _OC (0.868 V), J _SC (29.09 mA/cm²), and FF (81.17%). The J _SC values of all binary and ternary OSCs were independently verified by external quantum efficiency (EQE) measurements (Figure 5b). The calculated J _SC values derived from the EQE spectra (J _cal Table 1) are in good agreement with those obtained from J‐V test, with deviations within 5%. Notably, the ternary D18:BTP‐eC9:SP12 devices exhibit a pronounced EQE enhancement in the 650–850 nm region compared with the binary D18:BTP‐eC9 devices, which can be attributed to the complementary absorption contributions of BTP‐eC9 and SP12, consistent with the absorption spectra shown in Figure 2b. To verify the universality of SP12, it was further integrated into other representative host system, specifically D18:L8‐BO, D18:L8‐BO‐X, and PM6:L8‐BO shown in Figure S13–S15 and Table S10–S18. Relative to the corresponding binary references, the SP12‐based ternary devices exhibited consistent performance improvements across all systems. These results demonstrate the general applicability of SP12 as a versatile third component compatible with diverse blend morphologies. To further probe energetic disorder in the photoactive layers, Fourier‐transform photocurrent spectroscopy EQE (FTPS‐EQE) was employed to quantify the Urbach energy (E _U) [62]. As shown in Figure 5c, incorporation of SP12 into the D18:BTP‐eC9 blend results in a sharper absorption onset and a reduced E _U value, decreasing from 24.83 to 23.98 meV. This reduction indicates suppressed energetic disorder and a lower density of sub‐gap trap states at the donor–acceptor interface, which is beneficial for minimizing voltage losses in OSCs [63].

FIGURE 5.

The (a) J‐V characteristics under AM1.5G illumination (100 mW/cm²), (b) corresponding EQE spectra, and (c) FTPS‐EQE spectra at the absorption onset region of binary D18:BTP‐eC9 and ternary D18:BTP‐eC9:SP12 blends, respectively. The (d) EQE_EL spectra, (e) breakdown of energy losses (ΔE ₁, ΔE ₂, and ΔE ₃) calculated from an established method, and (f) PL spectra of the binary BTP‐eC9 and D18:BTP‐eC9, and ternary D18:BTP‐eC9:SP12 blend films, respectively. (g) The column plots of τ _extract and τ _recomb of OSCs based on D18:BTP‐eC9 and D18:BTP‐eC9:SP12 extracted from the TPC and TPV measurements, respectively. (h) The P _light‐V _OC plots of OSCs based on the binary and ternary blends. (i) Thermal stability tests under constant 70°C stress in the dark condition of OSCs based on D18:BTP‐eC9 and D18:BTP‐eC9:SP12.

TABLE 1.

Photovoltaic parameters of binary and ternary OSCs.

Active layer	V OC (V)	J SC (mA/cm2)	J cal a (mA/cm2)	FF (%)	PCE b (%)
D18:SP12	0.964	13.92	13.51	56.45	7.58 (7.27 ± 0.13)
D18:BTP‐eC9	0.861	28.53	27.35	79.33	19.49 (19.36 ± 0.09)
D18:BTP‐eC9:SP12	0.868	29.09	27.73	81.17	20.50 (20.13 ± 0.10)

^a Calculated from EQE spectra.

^b Average PCEs were obtained from 6 devices as summarized in Table S7–S9.

A comprehensive energy loss analysis was conducted to elucidate the voltage loss mechanisms. The charge‐transfer state energy (E _CT) was determined via sensitive EQE (sEQE) measurements based on the Marcus theory [62]. The fitted E _CT values were 1.34, 1.35 eV for D18:BTP‐eC9 and D18:BTP‐eC9:SP12 blends (Figure S16). Consequently, the total energy loss (E _loss) was calculated to be 0.477, 0.520, and 0.512 eV for the respective blends (Table S19), indicating a progressive reduction in voltage loss with the incorporation of SP12. The electroluminescence EQE (EQE _EL) spectra (Figure 5d) for the D18:BTP‐eC9:SP12 ternary blend is significantly higher than that of the binary D18:BTP‐eC9, pointing to a substantially lower non‐radiative energy loss (ΔE _nr). As summarized in Figure 5e, while all devices exhibit comparable radiative energy losses (ΔE ₁ and ΔE ₂, indicating inevitable and avoidable parts of the radiative energy loss, respectively) [64], the non‐radiative loss component ΔE ₃ (representing the ΔE _nr) decreased from 0.217 eV for the binary blend to 0.208 eV for the ternary blend. This suppression of ΔE _nr is a key factor contributing to the enhanced V _OC, aligning with strategies that focus on mitigating non‐radiative recombination through molecular design. Steady‐state photoluminescence (PL) quenching measurements (Figure 5f) revealed an enhanced quenching efficiency of 89.81% for the D18:BTP‐eC9:SP12 ternary film, compared to 88.93% for the D18:BTP‐eC9 binary film, indicating more efficient charge transfer and separation at the donor‐acceptor interface.

Transient photocurrent (TPC) and transient photovoltage (TPV) measurements were performed to investigate charge carrier dynamics [65]. As illustrated in Figures 5g and S18, the charge extraction time (τ _extract) decreased from 0.57 µs for the binary device to 0.42 µs for the ternary device, suggesting faster and more efficient charge collection. Concurrently, the charge recombination lifetime (τ _recomb), derived from TPV decay, increased markedly from 3.63 to 7.52 µs upon SP12 incorporation. This prolonged carrier lifetime signifies effective suppression of non‐geminate (bimolecular) recombination, which is critical for achieving high FF and J _SC values. Analysis of photocurrent density (J _ph) versus effective voltage (V _eff) (Figure S19) yielded charge dissociation probabilities (P _diss) of 0.990 for the ternary blend versus 0.986 for the binary, and charge collection efficiencies (P _coll) [66, 67] of 0.915 versus 0.908, confirming that SP12 improves both charge generation and transport processes. Further insight into recombination mechanisms was gained from light intensity (P _light) dependence studies [68, 69, 70]. The slope (n) of the V _OC versus ln(P _light) plot (Figure 5h) decreased from 1.16k _B T/q for D18:BTP‐eC9 to 1.09k _B T/q for D18:BTP‐eC9:SP12 (where k _B is Boltzmann’s constant, T is temperature, and q is elementary charge). All values are significantly below the threshold for dominant trap‐assisted recombination (approaching 2k _B T/q) [68, 69, 70], indicating that bimolecular recombination is the primary loss mechanism, and its influence is further suppressed in the ternary system. This is corroborated by the J _SC‐ P _light analysis (Figure S19), where the exponent α increased from 0.993 to 0.994 for the ternary blend, suggesting reduced bimolecular recombination. The charge transport properties were evaluated using the space‐charge‐limited‐current (SCLC) method (Figure S20, Table S21). The hole mobility (µ _h) and electron mobility (µ _e) for the D18:BTP‐eC9:SP12 ternary blend were measured to be 5.87 × 10⁻⁴ cm² V⁻¹ s⁻¹ and 5.10 × 10⁻⁴ cm² V⁻¹ s⁻¹, respectively, both higher than those of the binary blends. More importantly, the µ _h/µ _e ratio improved from 1.35 (binary) to 1.15 (ternary), indicating a more balanced charge transport. This balanced and enhanced mobility, coupled with the extended carrier lifetime, synergistically promotes efficient charge extraction and contributes to the superior FF and J _SC parameters in the corresponding OSCs.

Finally, the operational stability of the devices was assessed under thermal stress at 70° C. The ternary device based on D18:BTP‐eC9:SP12 exhibited significantly improved thermal stability, with the T ₈₀ lifetime (time to retain 80% of initial PCE) extending from 554 h for the binary device to 820 h (Figure 5i). This enhancement is attributed to the unique 3D star‐shaped architecture of SP12. The large molecular size and high T _g of SP12 effectively suppress excessive molecular aggregation and diffusion within the active layer, thereby, stabilizing the nanomorphology against thermal degradation and leading to superior long‐term performance [46].

In summary, the templated, epitaxial growth of BTP‐eC9 facilitated by SP12 leads to a more ordered and interconnected acceptor domain. This optimized morphology synergistically addresses several key limitations in binary OSCs: 1) It enhances charge transport pathways, corroborated by the increased and more balanced hole/electron mobility, which translates to high FF. 2) It reduces energetic disorder at the donor‐acceptor interface, as evidenced by the lower E _U value, contributing to reduced non‐radiative voltage loss and a higher V _OC. 3) It improves charge generation and collection efficiency, supported by the enhanced PL quenching, prolonged carrier lifetime and higher charge dissociation probability. Furthermore, the proposed interfacial localization of SP12 and the resulting stabilized nanomorphology provide a coherent explanation for the significantly enhanced thermal stability of the ternary device. In conclusion, the incorporation of the star‐shaped acceptor SP12 transcends the role of a mere third component; it acts as a morphology‐directing agent that orchestrates a more favorable and robust bulk‐heterojunction structure, thereby unlocking simultaneous improvements in efficiency, charge dynamics, and operational stability.

To further elucidate the impact of SP12 incorporation on charge carrier dynamics, transient absorption spectroscopy (TAS) measurements were conducted on the binary (D18:BTP‐eC9) and ternary (D18:BTP‐eC9:SP12) blend films. Given the distinct absorption profiles of the donor and acceptor components, a pump wavelength of 830 nm was selected to preferentially excite the acceptor moieties (BTP‐eC9 and BTP‐eC9:SP12) within the respective blends. The corresponding TAS data for the blend films and their neat counterparts are presented in Figures 6a,b, 6d,e, and S21. Upon selective excitation of the acceptor, both blend films exhibited analogous spectral features. Ground‐state bleaching (GSB) signals emerged immediately, with peaks located around 500–600 nm and 650–800 nm, corresponding to the donor (D18) and acceptor (BTP‐eC9 or SP12), respectively. Following photoexcitation, the acceptor GSB signal appeared instantaneously and subsequently decayed. Concurrently, the donor GSB signal emerged and intensified over time. This temporal evolution is characteristic of hole transfer from the photoexcited acceptor to the donor, signifying the formation of a charge‐transfer (CT) state at the donor‐acceptor interface [55, 71, 72]. To quantitatively analyze the exciton and charge carrier kinetics, the TAS data were subjected to a global fitting procedure (Figures 6c,f). The derived time constants, τ ₁ and τ ₂, are associated with distinct photophysical processes. The τ ₁ (sub‐picosecond scale) is attributed to ultrafast charge transfer at the donor‐acceptor interface immediately following exciton generation. And τ ₂ (picosecond scale) is assigned to charge transfer events occurring after exciton diffusion within the pure acceptor domains prior to interfacial encounter. The ternary blend D18:BTP‐eC9:SP12 exhibits significantly accelerated kinetics, with τ ₁ = 0.41 ps and τ ₂ = 12.52 ps, compared to τ ₁ = 0.81 ps and τ ₂ = 34.29 ps for the binary D18:BTP‐eC9 blend. This marked reduction in both time constants indicates a more rapid charge transfer process in the presence of SP12. The accelerated charge separation effectively outcompetes potential geminate recombination pathways, thereby enhancing free charge carrier generation. This improvement in early‐stage photo‐physics directly contributes to the observed superior photovoltaic parameters by minimizing energy losses associated with charge recombination [73].

FIGURE 6.

Contour plots of the TAS spectra of (a) D18:BTP‐eC9 and (d) D18:BTP‐eC9:SP12 excited at 830 nm. Kinetics of the GSB signal probed at 630 nm for (b) D18:BTP‐eC9 and (e) D18:BTP‐eC9:SP12 blends excited at 830 nm. Global fitting result curves for (c) D18:BTP‐eC9 and (f) D18:BTP‐eC9:SP12 blended films.

3. Conclusions

In this study, we have successfully designed and synthesized a very high molecular weight star‐shaped dodecamer acceptor, SP12, and incorporated it as a ternary component into high‐performance OSCs. Through systematic photophysical, morphological, and device physics characterizations, we elucidated that SP12 functions not merely as a supplementary absorber, but as a multifunctional morphology‐directing agent, synergistically enhancing device efficiency, stability, and charge dynamics. The multiple π‐conjugated arms play a crucial role as a heterogeneous nucleating agent during the active layer film formation process, guiding the epitaxial and ordered assembly of the host acceptor. This action optimizes the donor‐acceptor interfacial energetics and fosters the formation of a highly crystalline BHJ morphology with appropriate phase‐separation scale, which directly translates into significant improvements in charge carrier dynamics. Benefiting from this dual optimization of morphology and charge dynamics, the ternary OSCs based on D18:BTP‐eC9:SP12 achieved a champion PCE of 20.50%. More importantly, the 3D topology and high T _g of SP12 kinetically and odynamically “lock in” this optimized morphology, significantly suppressing molecular diffusion and enhance device stability. This strategy not only provides a new perspective for addressing the inherent trade‐off between efficiency and stability in OSCs, but also opens new avenues for designing multifunctional, high‐performance organic photovoltaic materials in the future.

Author Contributions

Bo Wang:: investigation, conceptualization, writing – original draft, methodology, data curation, and formal analysis. Chengyi Xiao:: conceptualization, funding acquisition, writing – original draft, writing – review and editing, formal analysis, data curation, and supervision. Xucong Liu:: investigation, data curation. Jixiang Xie:: investigation, data curation. Shijie Liang:: investigation, data curation. Dinglong Feng:: investigation, data curation. Xinhui Lu:: investigation, formal analysis, and data curation. Zhou Lu:: investigation, data curation, formal analysis, and methodology. Weiwei Li:: conceptualization, funding acquisition, writing – review and editing, supervision, data curation, and formal analysis.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: anie72829‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that supports the findings of this study are available in the supplementary material of this article