Table of Contents
Adv. Mater.: Acad. Yongfang Li and Dr. Yaowen Li’s Research Group Achieves 15.2% Efficient Flexible OSCs with 6,000 Bending Cycles
Research Team
First Author: Xiaobin Chen
Corresponding Authors: Hongwei Gu, Yaowen Li

Key Highlights
Achieving ultrahigh mechanical flexibility and efficiency exceeding 15% in flexible organic solar cells through a “welding” flexible transparent electrode.
- Efficiency Challenge: The power conversion efficiency (PCE) of flexible organic solar cells (OSCs) still lags behind rigid devices. Their mechanical stability cannot meet the demands of flexible electronics due to the lack of high-performance flexible transparent electrodes (FTEs).
- Innovative Concept: This paper introduces a “welding” concept to design FTEs with tight integration between the top electrode and bottom substrate.
- Technical Breakthrough: The top electrode, composed of solution-processed aluminum-doped zinc oxide (AZO) and silver nanowire (AgNW) networks, is well-welded through AZO’s capillary force effects and secondary growth, reducing AgNWs junction resistance.
- Substrate Enhancement: Polyethylene terephthalate was modified by embedding AgNWs, which were then used to connect with AgNWs in the top hybrid electrode, enhancing electrode-substrate adhesion.
- Superior Performance: Single-junction flexible organic solar cells based on this welded FTE demonstrated high performance, achieving a record-breaking 21% power conversion efficiency. Additionally, the PCE of flexible OSCs showed minimal dependence on device area and exhibited robust bending durability even under extreme test conditions.
- Bottleneck Breakthrough: Through this welding strategy, critical bottleneck issues related to FTEs in both photoelectric and mechanical performance aspects were comprehensively resolved.
Introduction
In February 2020, Advanced Materials published research by Academician Yongfang Li and Professor Yaowen Li et al. on achieving ultrahigh mechanical flexibility and efficiency exceeding 15% in flexible organic solar cells through a “welding” flexible transparent electrode. To demonstrate its viability as an electrode for flexible OSCs, the study examined fullerene and non-fullerene active layer materials with various bandgaps. All flexible OSCs demonstrated PCE comparable to their rigid counterparts, achieving a record-high efficiency of 15.21%. Importantly, the PCE of flexible organic solar cells showed minimal dependence on device area and exhibited robust bending durability even under extreme test conditions.
Background
Metal nanowires, particularly silver nanowires (AgNWs), are considered the most promising FTE materials due to their excellent transmittance, high conductivity, and flexibility. However, solution-processed AgNW networks typically have low coverage (less than 40%) and high junction resistance, substantially reducing device conductivity and even operational stability, as AgNW junctions concentrate heat locally through current radiation. Furthermore, randomly stacked AgNWs with poor adhesion to plastic substrates may induce device short circuits and reduce mechanical peeling stability. Although this issue can be partially mitigated by coating conductive polymers (such as PH1000) on AgNW network films, high parasitic absorption in the long-wavelength region and the acidity of PH1000 reduce light collection and deteriorate device stability.
Given the advantages of combining AgNW networks with PH1000 in enhancing coverage, adhesion, and conductivity, some researchers have attempted to replace PH1000 with zinc oxide (ZnO) to achieve high transmittance, favorable interface energetics, and low-temperature solution processability. Under the synergistic effects of capillary force and electrical bridging, ZnO solution can easily fill the AgNW network and weld connection points to achieve full-coverage AgNWs: ZnO hybrid electrodes with reduced junction resistance. To further improve the junction resistance and transmittance of such hybrid electrodes, precise silver grid patterns were dry-etched on pre-deposited metal films without junctions (AgNN). The resulting 35.2 Ω sq-1 low sheet resistance and 91.6% high transmittance (excluding polyethylene terephthalate (PET) substrate) were achieved with the 15 nm thick AgNN/ZnO hybrid electrode. However, the complexity and high energy consumption of manufacturing processes such as vacuum thermal evaporation, electrospinning, and ion beam etching would undoubtedly limit their practical applications.
To simplify the process, Chen et al. proposed an ionic electrostatic repulsion method by doping polyelectrolytes into AgNW-water solution to reduce AgNW aggregation and junction resistance. The resulting hybrid AgNWs:ZnO electrode containing grid-like AgNW patterns showed excellent optoelectronic properties and smooth surface. Therefore, it is desirable to manufacture high-performance and low-cost hybrid electrodes through solution processing with precise control of composition and structure, which is urgently needed to promote the development of flexible electronics.
This paper presents an integrated FTE design welding strategy, including top electrodes and bottom substrates, to match state-of-the-art non-fullerene near-infrared absorption and robust flexibility active layer materials. For the top electrode, a solution-processed AgNW-based hybrid was adopted, precisely controlling composition through capillary force effects and secondary growth of aluminum-doped zinc oxide (AZO) to weld unfavorable AgNW junctions. The insulating bottom PET substrate was also modified by embedding AgNWs in UV-curable resin, enabling connection between AgNWs in the top hybrid electrode and the bottom substrate. Through this approach, improvements were achieved not only in optoelectronic properties (such as conductivity and transmittance) but also in adhesion between the top electrode and substrate, as well as the previously poor morphology before the top hybrid electrode. The resulting welded AgNW-based FTE exhibited a low sheet resistance (Rsh) of ≈18 Ω sq-1, maximum transmittance of about 95% at 550 nm (excluding PET substrate), smooth surface, and good mechanical stability in bending and peeling tests.
Key Results


Figure 1 a) Schematic diagram of electrode fabrication. b) Sheet resistance, c) conductivity statistics, and d) optical transmission spectra of Em-Ag/PH1000, Em-Ag/AgNWs:AZO-SG, and AgNWs:AZO-SG electrodes, along with normalized absorption spectra of the active layer. All transmittance values for FTEs include PET substrate. e) Transmittance (λ = 550 nm) plotted as a function of sheet resistance, inset: photograph of Em-Ag/AgNWs:AZO-SG FTE.
The schematic diagram of the FTE manufacturing process is shown in Figure 1a. First, the PET substrate was modified by coating with an AgNW film, followed by applying UV-curable resin to protect the coating from air exposure and promote AgNW adhesion to the PET substrate, forming an embedded AgNW substrate (Em-Ag). Interestingly, the resulting Em-Ag substrate still exhibited significantly improved conductive performance (sheet resistance = 130 Ω sq-1, conductivity = 7.7 × 104 S m-1), indicating that incompletely embedded AgNWs provided additional charge transport channels.
To demonstrate the feasibility of using this welded FTE in flexible OSCs, comprehensive evaluations were conducted on conduction mechanisms, morphology, electrical and optical properties, and mechanical stability. The conductive polymer PH1000 was coated onto the Em-Ag substrate (Em-Ag/PH1000), and the AgNWs:AZO-SG hybrid electrode was coated onto bare PET substrate (AgNWs:AZO-SG) for comparison. In the case of Em-Ag/AgNWs:AZO-SG FTE, the calculated Rsh based on statistical results was 18 Ω sq-1 (average) with a standard deviation of 0.66, comparable to conventional glass/ITO electrodes and significantly lower than Em-Ag/PH1000 (90 Ω sq-1) and AgNWs:AZO-SG (28 Ω sq-1) FTEs (Figure 1b). This result is also consistent with their respective conductivity values, as shown in Figure 1c.
As shown in Figure 1d, at an Rsh of 18 Ω sq-1, the welded FTE exhibits an average transmittance of up to 84% in the 500-1000 nm range without any degradation at longer wavelengths. This characteristic matches well with state-of-the-art non-fullerene active layer materials’ absorption extending into the near-infrared region. In contrast, when using traditional conductive polymer PH1000 as the top electrode (Em-Ag/PH1000), the average transmittance was only 75% (with Rsh of 90 Ω sq-1) and could even drop to 70% at 1000 nm. The research team also noted that AgNWs:AZO-SG FTE without Em-Ag showed slightly enhanced transmittance, although its Rsh was 1.5 times higher compared to the welded FTE, indicating Em-Ag’s crucial role in enhancing conductivity.

Figure 2 a–c) Cross-sectional SEM images of: a) Em-Ag/PH1000, b) Em-Ag/AgNWs:AZO-SG, and c) AgNWs:AZO-SG FTEs. d) Sheet resistance changes in various FTEs with increasing bending cycles for inward and outward bending tests. e) Adhesion values for various FTEs, inset: schematic diagram of adhesion measurement.
As shown in Figure 2d, the Rsh/R0 of AgNWs:AZO-SG grown on bare PET substrate increased slightly after 1200 inward bending cycles, indicating minimal impact of AZO’s brittle characteristics on bending durability. In contrast, when the PET substrate was modified with Em-Ag, the Rsh/R0 change of Em-Ag/AgNWs:AZO-SG FTE was negligible regardless of bending direction, consistent with the behavior of conductive polymer-based electrodes (Em-Ag/PH1000). This result suggests that the synergistic effect of bottom Em-Ag and AZO-SG-assisted top electrode can further mitigate the impact of brittle AZO on flexibility.
To clarify this, cross-sectional SEM imaging was performed on the FTEs. Figures 2a-c show that all components, including AgNW, PET, UV-curable resin, and PH1000, can be clearly distinguished in various FTEs. In the case of Em-Ag/AgNWs:AZO-SG FTE, the authors observed that the top layer AgNWs connected with UV-curable resin-exposed AgNWs at the interface. As expected, this interface welding would enhance the bonding strength between the top electrode and substrate and reduce stress on the substrate, thus contributing to enhanced bending durability. Furthermore, this interface welding also enhanced adhesion between the electrode and substrate. Through 90° peel measurements, the adhesion between the AgNWs:AZO-SG layer and Em-Ag substrate reached 1.45 N mm-1. In contrast, Em-Ag/PH1000 and AgNWs:AZO-SG FTEs without interface welding showed reduced adhesion to 58% and 73.2%, respectively, as shown in Figure 2e. Due to comprehensive improvements in optoelectronic and mechanical performance, the welded FTE is a promising choice as an electrode in flexible OSCs. Clearly, these advantages should be accompanied by improved device performance.
Figure 3 a) Schematic diagram of device energy alignment. b) Schematic of flexible organic solar cells and donor PBDB-T-2F and acceptor Y6 in the active layer.
From the device energy level diagram in Figure 3a, Em-Ag/AgNWs:AZO-SG FTE shows good energy level alignment with non-fullerene active layer materials. The small energy offset between the electrode work function and the LUMO level of non-fullerene acceptors such as Y6 (4.1 eV) can effectively reduce device energy losses.
Consequently, flexible OSCs with an inverted structure of Em-Ag/AgNWs:AZO-SG/active layer/MoO3/Al (Figure 3b) were fabricated using PBDB-T-2F:Y6, PBDB-T-2F:IT-4F, and PTB7-Th:PC71BM as active layer materials. Notably, no additional electron transport layer was added in the devices due to the matching energy levels and high electron extraction capability of AZO in the FTE, enabling simplified device structure.

Figure 3 c-f) Photovoltaic performance of PBDB-T-2F:Y6-based flexible OSCs: c) J-V curves under AM1.5G 100 mW cm−2 illumination; d) EQE spectra and integrated Jsc, e) natural logarithm of Jsc versus light intensity with linear relationship fitting, and f) Voc versus natural logarithm of light intensity with linear relationship fitting.
As shown in Figure 3c, J-V curves of flexible OSCs using PBDB-T-2F:Y6 as the active layer were measured under AM1.5G 100 mW cm−2 illumination. To verify Jsc variations in various FTEs, External Quantum Efficiency (EQE) spectra of corresponding devices were examined to evaluate light response across the entire absorption region (Figure 3d). The EQE curves were measured using the Enlitech QE-R system. The QE-R software includes built-in functionality for calculating integrated current density Jsc(EQE) against the AM1.5G spectrum. As expected, devices based on high-transmittance AgNWs:AZO-SG and Em-Ag/AgNWs:AZO-SG FTEs showed significant light response in the 500-900 nm range, with integrated Jsc values consistent with those obtained from J-V measurements, showing less than 4% deviation.
To further understand how FTE optoelectronic properties affect device performance, related photophysical processes were studied. Carrier recombination processes in each device were evaluated by examining Jsc dependence on various light intensities, with data fitting the power law Jsc ∝ Plightα. As shown in the log-Jsc versus log Plight curves of flexible OSCs in Figure 3e, similar slopes (α values) indicate that active layer quality is minimally affected by underlying FTEs due to smooth surface and good wettability. The α value approaching unity for Em-Ag/AgNWs:AZO-SG FTE devices indicates negligible bimolecular recombination in the active layer. Carrier recombination behavior was evaluated through Voc dependence on light intensity measurements based on the relationship Voc ∝ (nkT/q)ln(Plight) (Figure 3f). The significantly reduced slope of 1.12 kT/q (approaching kT/q) for Em-Ag/AgNWs:AZO-SG FTE devices indicates lower trap states, possibly due to reduced carrier recombination between the electrode and active layer materials. Enhanced exciton dissociation efficiency further confirms this behavior.
Table 1 Solar cell performance parameters of PBDB-T-2F:Y6-based OSCs under AM1.5G 100 mW cm−2 illumination. The Jsc(EQE) [Jcal in the table above] and Jsc(IV) [Jsc in the table above] values show less than 3% comparison error.

EQE is defined as the ratio of output electrons to incident photons. Jcal can be calculated through the EQE curve and photon flux spectrum. The formula is as follows:

Notably, flexible OSCs based on Em-Ag/AgNWs:AZO-SG FTE achieved a PCE of 15.21%, with an open-circuit voltage (Voc) of 0.832 V, Jsc of 25.05 mA cm-2, and fill factor (FF) of 72.97%. To our knowledge, the 15.21% PCE was the highest reported value for single-junction flexible OSCs at the time of publication.

Figure 4 a) Photograph of flexible OSC with a 4 mm bending radius on cyclic bending machine. b) Schematic diagram of inward and outward bending tests. c-e) Relative PCE degradation of flexible OSCs based on: c) Em-Ag/PH1000, Em-Ag/AgNWs:AZO-SG, and AgNWs:AZO-SG FTE versus bending cycles with 4 mm radius; d) Em-Ag/AgNWs:AZO-SG FTE versus bending radius after 1200 inward bending cycles, and e) Em-Ag/AgNWs:AZO-SG FTE versus bending cycles with 4 mm inward bending radius.
Bending durability is another crucial factor for high-performance flexible OSCs. Therefore, the obtained flexible OSCs underwent bending tests in both directions, with 1200 consecutive bending cycles at a 4 mm radius, to evaluate their stability under mechanical bending (Figure 4a, b). As is well known, the conductive polymer PH1000 exhibits excellent mechanical flexibility due to its plastic nature. As shown in Figure c, flexible OSCs based on Em-Ag/PH1000 FTE retained 90.8% of their initial efficiency after 1200 outward bending cycles and 89.5% after inward bending cycles.
To further understand the mechanical stability of flexible OSCs, bending tests were also conducted at different bending radii. As shown in Figure 4d, devices based on Em-Ag/AgNWs:AZO-SG FTE maintained robust efficiency (PCE) after 1200 bending cycles under 0-8 mm bending radii. Even under extreme conditions where the device was completely folded inward (Rc = 0 mm), it maintained 81.7% of its initial PCE. Furthermore, flexible OSCs with Em-Ag/AgNWs:AZO-SG FTE surprisingly maintained 75% of their initial PCE values after 6000 bending cycles with an inward bending radius of 4 mm (Figure 4e).
Summary
This work successfully developed a welded FTE with Em-Ag/AgNWs:AZO-SG structure to address the mismatch between FTE transmission/absorption spectra and state-of-the-art non-fullerene active layers. The fine-tuned AgNW network combined capillary force effects and secondary growth of AZO in solution to effectively avoid parasitic absorption of electrode compounds, and the deposited AZO welded AgNW junctions. Exposed AgNWs in the bottom Em-Ag further connected with AgNWs in the top AgNWs:AZO-SG layer, enhancing FTE mechanical performance. The welded FTE demonstrated good optoelectronic performance, smooth surface, and flexibility.
Consequently, based on Em-Ag/AgNWs:AZO-SG FTE, flexible organic solar cells using various bandgap active layers achieved power conversion efficiencies comparable to glass/ITO devices, with record-high PCEs of 15.21% and 12.28% for small-area and large-area single-junction flexible organic solar cells, respectively. More importantly, these devices demonstrated robust mechanical performance in both bending and peeling tests. The strategy presented in this work provides new functionality for emerging FTEs, showing promise in driving flexible electronic devices toward high performance and large area applications.
Publication Information
Realizing Ultrahigh Mechanical Flexibility and >15% Efficiency of Flexible Organic Solar Cells via a “Welding” Flexible Transparent Electrode
Xiaobin Chen, Guiying Xu, Guang Zeng, Hongwei Gu,* Haiyang Chen, Haitao Xu, Huifeng Yao, Yaowen Li,* Jianhui Hou, and Yongfang Li

