Water is a significant factor in the degradation of flexible electronics1,2. Water condensation can occur on the device surfaces because of rain or high humidity outdoors, and everyday indoor activities like washing hands can result in water attached to cloth and skin3,4. Ultraflexible organic photovoltaics (OPVs) are light and can be deformed by using a substrate with a thickness of several µm or less, making them compatible with wearable and skin electronics5,6. Hence, waterproofness becomes a critical parameter when utilizing ultraflexible OPVs as energy-harvesting solutions for wearable electronics that are integrated into daily life7,8.

Typically, enhancing the waterproofness of OPVs involves embedding them in a thick or rigid encapsulation layer9. For example, flexible OPVs with double-sided, 500 μm-thick encapsulating layers maintained approximately 95% of their initial power conversion efficiency (PCE, ~7.9%) after being immersed in water for 120 min10. Another textile-based OPV (PCE ~ 7.2%) encapsulated on both sides by a tens of microns thick composite layer of atomic-layer-deposited aluminum oxide and SiO2–polymer barriers demonstrated a 2% loss in efficiency after being washed for 20 cycles (10 min per cycle) under stirring at 200 rpm using a magnetic stirrer11.

However, achieving improved waterproofness without compromising mechanical flexibility and conformability remains a challenge. Covering with thick or rigid encapsulation layers to enhance waterproofness inevitably impair conformability, stretchability, and even output performance of wearable electronics12. On the other hand, the water sensitivity of conventional hole-transporting layer (HTL) materials would induce performance degradation of OPVs in the presence of water13,14. Additionally, deformation typically increases moisture transmission rate of elastic or flexible encapsulating materials15, thereby accelerating the deterioration of the devices in the presence of water or moisture.

Here, we present an ultraflexible OPV featuring an in-situ growth of AgOx HTL, ensuring significant waterproofness even under mechanical deformation by strengthening the electrode and active layer interface adhesion. The in-situ growth procedure entails the direct deposition of Ag on active layers, followed by annealing in air at 85 °C for 24 h. Consequently, our 3 μm-thick OPVs exhibit significantly improved waterproofness compared to those using conventional HTL materials, retaining 89% and 96% of the initial efficiency after 240 min of immersion and 300 stretching/releasing cycles at 30% strain underwater, respectively. These features enable our ultraflexible OPVs to withstand machine-washing processes. With a champion efficiency of 14.3% under one sun illumination, our developed OPVs outperform existing waterproof OPVs (Supplementary Table 1)10,11,16,17. Furthermore, our approach demonstrates universality in achieving waterproof OPVs with different active layers. The strategy significantly improves the waterproofness from the fundamental structure, enhancing waterproofness of OPVs without compromising mechanical durability and conformability. Our developed approach opens possibilities for wearable energy systems and propels the progress of waterproof electronics.


Device with in-situ growth of HTL

The waterproof and ultraflexible OPVs were realized by growing of AgOx HTL in situ with a structure of transparent polyimide (tPI)/ITO/PEI-Zn/PM6:Y6/AgOx/Ag/Parylene (Fig. 1a). The Ag electrode was directly deposited onto an active layer of PM6:Y6 (Supplementary Fig. 1), followed by annealing at 85 °C in air. The devices were encapsulated with a 1 μm-thick parylene layer after annealing treatment. Before annealing, the fabricated OPVs showed poor efficiency—open-circuit voltage (VOC) of 0.05 V, short-circuit current density (JSC) of 15.8 mA cm−2, fill factor (FF) of 0.27, PCE of 0.2% (Fig. 1b), and an average PCE of 0.2% (Table 1). After 12 h of annealing, the device performance improved—VOC = 0.77 V, JSC = 25.8 mA cm−2, FF = 0.66, PCE = 13.1%, and an average PCE of 12.5%. Extending the annealing to 24 h further improved the performance—VOC = 0.77 V, JSC = 26.5 mA cm−2, FF = 0.71, PCE = 14.3%, and an average PCE of 13.6%. Supplementary Fig. 2 displays the PCE distribution based on 30 separate solar cell devices after 24 h of annealing at 85 °C.

Fig. 1: Structure, characteristics, and mechanism analysis of waterproof OPVs.
figure 1

a Photograph of the ultraflexible and waterproof OPVs floating on water. Schematic device structure of the waterproof OPVs. Scale bar: 1 cm. b J–V characteristics of devices with 0 h, 12 h, and 24 h annealing treatment at 85 °C in air. c The evolution of EQE of the OPVs under different annealing times at 85 °C. d Depth profiles of O element determined by D-SIMS for initial and annealed samples. e, Energy level diagram of annealed samples.

Table 1 Photovoltaic parameters of the OPVs based on different annealing times

The external quantum efficiency (EQE) spectrum showed significant enhancement after annealing (Fig. 1c). Devices with 24 h annealing did not exhibit hysteresis for the forward and reverse scans (Supplementary Fig. 3), indicating efficient extraction and collection of charge carriers. Free-standing devices were delaminated from the supporting glass and seamlessly attached to human skin (Supplementary Fig. 4). The devices demonstrated excellent storage stability, preserving 92% of the pristine PCE after 1369 h of air storage in darkness at room temperature (20–25 °C) and 93% of their original PCE after 1371 h of aging at 85 °C in air (Supplementary Fig. 5). These results are comparable to recently reported organic photovoltaics with HTLs (Supplementary Table 2)10,18,19,20,21,22,23,24,25. Additionally, the devices maintained 87.7% of their initial efficiency under AM 1.5 G illumination with a filter cutting off light below 400 nm in a N2-filled glovebox for 334 h (Supplementary Fig. 6), surpassing recently reported ultrathin OPVs (Supplementary Table 3)24,26,27,28,29,30. The temperature and annealing time optimizations of in-situ growth of AgOx HTL are shown in Supplementary Fig. 7. The devices annealed for 24 h at 85 °C demonstrated the best average performance (Supplementary Table 4), highlighting the crucial role of the appropriate annealing in the formation of an efficient AgOx HTL. Supplementary Fig. 8 illustrates the evolutions of different annealing times of photocurrent density (Jph)–effective voltage (Veff) curves and light intensity-dependent JSC. These results suggest that devices annealed for 24 h at 85 °C experienced a significant improvement in the collection and extraction of charge carriers, as well as the suppression of bimolecular recombination. However, the excessive annealing can cause performance deterioration24,31. Consequently, the prolonged heating could not yield further enhancements in device performance. Additionally, the proposed thermal strategy for in-situ growth of AgOx HTL exhibited much better performance compared with natural oxidation treatment32 reported in a previous work (Supplementary Fig. 9 and Supplementary Table 5), indicating that moderate thermal stress could promote the growth of AgOx.

Dynamic secondary ion mass spectrometry (D-SIMS) measurement was conducted to analyze the in-situ growth of AgOx HTL. Fig. 1d displays the depth profiles of O signal intensities of glass/ITO/PM6:Y6/Ag before and after annealing. The Ag and PM6:Y6 layer regions are identified based on F element distribution (Supplementary Fig. 10). After 24 h of annealing at 85 °C, the O signal intensities in Ag layer improved significantly. A peak at the Ag and PM6:Y6 interface suggests the oxygen diffused from the interface to form the AgOx HTL layer32,33. In addition, annealing the devices with an Ag electrode at 85 °C for 24 h in an N2-filled glovebox did not show any improvement in PCE (Supplementary Fig. 11), indicating that annealing treatment in air is necessary for AgOx HTL in-situ growth.

Subsequently, we checked the work function of the Ag electrode before and after annealing. Notably, the work function of top surface of the annealed Ag electrode is almost the same as that of a fresh Ag electrode, and the conductivity remains unchanged (Supplementary Fig. 12). These indicate that the 24 h annealing does not cause significant oxidation on the Ag surface, thereby not affecting the performance. Here, we applied a removable electrode34 to assess the work function of the bottom surface of the Ag electrode in contact with the active layer (Supplementary Fig. 13a). The annealed removable Ag electrode exhibited a work function of approximately 5.0 eV (Supplementary Fig. 13b), similar to the reported value35. Notably, silver oxidizes faster in the Ag-organic interface during annealing compared with the top surface of the Ag electrode because oxygen diffusion across grain boundaries is quicker than lattice diffusion36. The silver electrode generally grows in the Volmer–Weber mode (‘island’ mode) on an active layer37, leading to a rougher surface and higher grain-boundary density in the Ag–organic interface than at the top surface of Ag electrode38. The energy-level diagram of the optimized devices is shown in Fig. 1e.

Stability in water

We tested the waterproofness of the free-standing OPVs by immersing them in water (Fig. 2a). The OPV with an AgOx/Ag electrode exhibited only a 11% decrease in PCE after 4 h of water immersion (Fig. 2b). Additionally, we explored the influence of water temperature on the waterproofness. All devices based on AgOx HTL demonstrated significant waterproofness under 0 °C, 20 °C, and 40 °C water immersion (Supplementary Fig. 14). The results indicate that the waterproof performance remains largely unaffected by water temperatures in the range of 0 °C–40 °C. For comparison, devices having a structure of tPI/ITO/PEI-Zn/PM6:Y6/MoOx/Ag/Parylene were also fabricated (Supplementary Fig. 15). However, the device with MoOx/Ag experienced a complete performance drop after only 2 h of water immersion. Then, we examined the electrode conductivity of the waterproof devices before and after water immersion. Both AgOx/Ag and MoOx/Ag electrodes maintained a stable conductivity after water immersion (Supplementary Fig. 16a, b), implying that the electrode conductivity does not impact the water stability of the devices. To delve deeper, impedance tests were conducted, and the impedance spectra of the device before and after water immersion are shown in Supplementary Fig. 16c, d. The transport resistance39 (intersecting point of the curve and the lateral axis at low frequencies) of the device based on AgOx/Ag slightly increased after water immersion, whereas that of the device with MoOx/Ag sharply increased only after 1 h immersion. The increased resistance indicates an increase in charge recombination after water immersion39,40,41. Because both devices share the same structure except for the HTL layer, the observed difference can be attributed to the charge transportation within the HTL layer. The OPV device with the AgOx/Ag electrode also works when fully immersed in water under light illumination (Supplementary Fig. 17). Additionally, we compared the long-term operation of free-standing OPVs in water with different HTLs (Supplementary Fig. 18). Fig. 2c shows the JSC of the device with AgOx/Ag and MoOx/Ag electrode under light illumination in water. The current of the developed OPV with AgOx/Ag electrode remained above 91% of its initial value after operation in water for 64 min. In contrast, the device with MoOx/Ag electrode exhibited a rapid deterioration, with a loss of 35% after operation in water for 24 min.

Fig. 2: Stability of free-standing devices in water.
figure 2

a Photograph of the free-standing devices being immersed in water. Scale bar: 1 cm. b The evolution of PCE for PM6:Y6 devices based on AgOx/Ag and MoOx/Ag electrodes with different immersion times in water. c, The evolution of JSC for PM6:Y6 devices based on AgOx/Ag and MoOx/Ag electrodes as a function of dipping time in water under light illumination. d Photograph of the free-standing devices under deformation stress in water. Scale bar: 1 cm. e The evolution of PCE for PM6:Y6 devices based on AgOx/Ag and MoOx/Ag electrodes under cyclic compression test in water. f The evolution of PCE for PM6:Y6 devices based on AgOx/Ag and MoOx/Ag electrodes under washing process by washing machine.

Subsequently, we subjected the devices to a cyclic compression test to assess their mechanical durability underwater. A photograph of free-standing devices under deformation stress in water is shown in Fig. 2d. The cyclic compression test under water involves a combination of stresses, including water exposure and deformation pressure, which impose stringent demands on device stability. The testing equipment for cyclic compression test is depicted in Supplementary Fig. 19. The OPV with AgOx/Ag electrode retained 96% of its initial value under 30% strain after 300 cycles of stretching–compressing in water (Fig. 2e). In contrast, the device with MoOx/Ag electrode saw its PCE plummet to 0% after just 100 stretching–compressing cycles. The sharp decline indicates that the mechanical stress underwater severely compromises the waterproofness of devices with MoOx/Ag electrodes. The results emphasize that OPVs with AgOx/Ag electrodes maintain stability even when subjected to mechanical deformation underwater, highlighting their suitability for wearable electronics.

Furthermore, given the excellent water stability of our devices even under deformation, we conducted the machine-washing test. Fig. 2f illustrates the machine-washing test results for free-standing OPVs without detergent. OPVs with AgOx/Ag electrodes remained 90% of their original PCE after two washing cycles in a washing machine. However, devices with MoOx/Ag electrodes experienced a complete PCE drop after two washing cycles. Notably, the concept of ultraflexible OPVs enduring machine-washing has not been reported in previous literature. These outcomes underscore that the developed OPVs meet the rigorous requirements of machine-washing tests. Furthermore, these findings suggest that OPVs with AgOx/Ag electrode can maintain continuous operation underwater, thereby enabling long-running power source for underwater autonomous systems and sensors42. The substitution of MoOx with AgOx, achieved through a novel thermal annealing strategy, advanced the waterproofness of OPVs. With HTL optimization, the water stability of OPVs improved from the aspect of fundamental structure perspective. We believe that the combination of AgOx HTL and thermal annealing presents an alternative potential strategy for designing waterproof electronics.

Interface adhesion strength

To identify the primary source of degradation in devices with MoOx/Ag electrodes exposed to water, we fabricated devices with the structure of glass/ITO/PEI-Zn/PM6:Y6/MoOx/Ag and immersed them in water. As shown in Fig. 3a, the top electrode is completely removed from the active layer after 1 h of immersion. To investigate further, we separately redeposited the Ag and MoOx/Ag electrodes onto the active layer. The device with the redeposited Ag electrode exhibited a poor efficiency of 0.6% (Fig. 3b), whereas the device with the redeposited MoOx/Ag electrode exhibited a PCE comparable to the initial device (Fig. 3c and Table 2). This indicates that delamination occurred at the interface between MoOx and the active layer during water immersion, leading to PCE deterioration.

Fig. 3: Mechanism analysis for enhancing waterproofness.
figure 3

a Photograph of the device with initial MoOx/Ag (left), after immersion in water for 1 h (middle), and redeposited metal electrode (right). Scale bar:5 mm. b J–V characteristics of PM6:Y6 devices after 1 h of immersion with redeposited Ag. c J–V characteristics of PM6:Y6 devices with initial and redeposited MoOx/Ag. d J–V characteristics of PM6:Y6 devices with AgOx/Ag electrode before and after immersing in water for 1 h. e Force stroke curves of initial samples with AgOx/Ag and MoOx/Ag. f Force stroke curves of samples with AgOx/Ag and MoOx/Ag after being immersed in water for 1 h. g Comparison of adhesion for AgOx/Ag and MoOx/Ag electrodes on the PM6:Y6 active layer. The error bars in the plots of adhesion indicate standard deviations (based on three samples).

Table 2 Photovoltaic parameters of OPVs before and after immersion in water under 100 mW cm-2 AM 1.5 G illumination

Fig. 3d presents the J–V characteristic curves of the device with structure of glass/ITO/PEI-Zn/PM6:Y6/AgOx/Ag before and after immersion in water for 1 h. The device exhibited a loss in performance of only 2.2% after immersion (Table 2). In addition, we conducted a waterproofness test on naturally oxidized devices (Supplementary Fig. 20). The naturally oxidized devices demonstrated significant PCE deterioration (10.3% loss) after 1 h of water immersion (Supplementary Table 6). This could be attributed to the improved bonding strength between the active layer and electrode due to thermal annealing, compared to natural oxidation treatment (Supplementary Fig. 21), thereby highlighting the importance and advancement of the thermal annealing strategy. A thermally stable interface between the electron-transporting layer (ETL) and active layer is essential for the success of this strategy. Fortunately, our previous work successfully established a thermally stable interface using the PEI-Zn ETL24. Additionally, we fabricated devices with the structure of glass/ITO/PEI-Zn/PM6:Y6/PEDOT:PSS/Ag. Similar to devices with MoOx/Ag electrodes, water exposure resulted in delamination between the top electrode and active layer (Supplementary Fig. 22). These results clearly demonstrate the superior reliability of the AgOx/Ag structure compared to MoOx/Ag or PEDOT:PSS/Ag when exposed to water.

To quantitatively compare delamination forces between different HTLs and active layer, a stretching machine was employed for a tensile test (Supplementary Fig. 23). Cross-sectional scanning electron microscopy (SEM) images revealed that the delamination occurred at the interface between the active layer and HTLs during peeling-off test (Supplementary Fig. 24), indicating that the tensile force can be used to evaluate the strength between HTLs and the active layer. The tensile force between AgOx and active layer was 51.0 N, whereas the force between MoOx and active layer was 24.5 N (Fig. 3e). After water immersion, the tensile force between AgOx and active layer remained at 47.6 N, whereas the force between MoOx and active layer decreased to 11.5 N (Fig. 3f). The slight decrease in the tensile force for AgOx/Ag after immersion can be attributed to the swelling of organic materials in water. Importantly, even after 1 h of immersion, the tensile force between AgOx and the active layer remained higher than that of MoOx without immersion.

Adhesion (σ) is defined as \({{{{{\rm{\sigma }}}}}}=\frac{P}{\pi {a}^{2}}\), where P is the critical force at interfacial failure and a is the radius of the stud43. Fig. 3g illustrates that the adhesion between AgOx and the active layer demonstrates a value of 1.53 MPa, which is twice that between MoOx and the active layer (0.76 MPa). After 1 h of water immersion, the adhesion between AgOx and the active layer remained relatively stable at 1.36 MPa, whereas that between MoOx and active layer decreased to 0.30 MPa. The interface adhesion strength between the active layer and anodes (AgOx/Ag and MoOx/Ag) can be quantified using the thermodynamic work of adhesion (Wa). Wa represents the work needed to separate two in-contact materials44 and is related to the surface energy (γ) of these materials. The γ of a film can be characterized by measuring the contact angles of liquid drops on the film45. We measured the contact angles of deionized water and glycerol drops on films (Supplementary Fig. 25), and we utilized average surface energies of water and glycerol for accurate calculations46. The dispersive surface energy (γ d) and polar surface energy (γ p) of AgOx are 29.67 mJ/m2 and 1.67 mJ/m2, respectively, resulting in a surface energy of 31.34 mJ/m2. For MoOx, the γd and γp are 3.41 mJ/m2 and 80.21 mJ/m2, respectively, yielding a surface energy of 83.62 mJ/m2 (Supplementary Table 7). Additionally, the different polarities of AgOx (γ d > γ p) and MoOx (γ d < γ p) indicate that the calculation of Wa should be different (See Methods section for detailed information). These differences in polarities and surface energies result in varying Wa values between the PM6:Y6 active layer and AgOx/Ag (44.08 mJ/m2) and MoOx/Ag (24.55 mJ/m2) anodes. A higher Wa between the active layer and anode indicates a stronger interface adhesion. Hence, the replacement of the anode from MoOx/Ag to AgOx/Ag significantly improves the interface adhesion. Moisture permeation from the interface could lead to increased defect density47 and delamination48, resulting in efficiency deterioration49,50,51. The strengthened interface plays a significant role in preventing the delamination of the electrode and active layer. Additionally, the water contact angle of MoOx is 6.3°, whereas that of AgOx is 92.8°. A higher water contact angle indicates better moisture resistance52. These findings underscore the superior water stability of the AgOx/Ag structure, making it the preferred choice for water-exposed environments.

Universality of the waterproof strategy

To improve the waterproofness of the proposed ultrathin freestanding OPVs10, additional 1 μm-thick parylene encapsulation layer was added to the side of the tPI substrate (Fig. 4a). Following this, the double-sided encapsulated PM6:Y6 OPVs with AgOx/Ag electrode maintained over 92% of their original PCE after 4 h of immersion. In contrast, devices with MoOx/Ag electrodes exhibited a decline of their pristine PCE to 28% (Fig. 4b). It is noteworthy that while double-sided encapsulation indeed proves efficacious in enhancing the waterproofness of OPVs with MoOx/Ag electrodes, it still falls short in waterproofing OPVs incorporating conventional HTLs. The encapsulation layer of parylene has a water vapor transmission rate (WVTR) of approximately 90 g/m2 per day18, inadequate for fully preventing the impact of water on the devices53. Moreover, we conducted the maximum power point tracking (MPPT) test with devices immersed in water. The device with MoOx/Ag electrode diminished the performance to almost zero within 20 min. In contrast, the device with AgOx/Ag electrode maintained 85.8% of its initial efficiency after operation in water for 70 min (Supplementary Fig. 26). These results underscore the exceptional waterproofness exhibited by our devised OPVs under operational conditions. In addition, the operational stability in water fluctuates considerably and decreases faster than that in air. The result may be attributed to the heat generated during maximum power point operation, causing the expansion of conductive tape connections, thereby weakening the waterproof capabilities of the encapsulation layer.

Fig. 4: Universality of developed structure in waterproofness improvement.
figure 4

a Schematic diagram of device with double-sided encapsulation layer. b–f The evolution of PCE for double-side encapsulated devices based on different active layers after being immersed in water. b Free-standing OPVs based on PM6:Y6 active layer with AgOx/Ag and MoOx/Ag electrodes. c Free-standing OPVs based on PBDTTT-OFT:IEICO-4F active layer with AgOx/Ag and MoOx/Ag electrodes. d Free-standing OPVs based on PM6:BTP-eC9 with AgOx/Ag and MoOx/Ag electrodes. e Free-standing OPVs based on PM6:ITCC active layer with AgOx/Ag and MoOx/Ag electrodes. f Free-standing OPVs based on ternary active layer of PM6:Y6:PC71BM, PBDTTT-OFT:IEICO-4F:PC71BM, and PM6:BTP-eC9:PC71BM with AgOx/Ag electrodes. The error bars in the plots indicate standard deviations (based on six devices).

Subsequently, we examined the water stability of the OPVs with different active layers, including PBDTTT-OFT:IEICO-4F, PM6:BTP-eC9, and PM6:ITCC (Supplementary Fig. 1). Devices with AgOx/Ag electrodes exhibited PCE values comparable to those with MoOx/Ag electrodes (Supplementary Fig. 27a–f and Supplementary Table 8). As depicted in Fig. 4c–e, all AgOx/Ag electrode-based devices demonstrate superior waterproofness compared to devices with MoOx/Ag electrodes. These results demonstrate the wide applicability of the developed structure in enhancing waterproofness. Furthermore, devices based on PM6:ITCC active layer exhibit better PCE retention after immersion in water compared to other devices based on PM6:Y6, PBDTTT-OFT:IEICO-4F, and PM6:BTP-eC9. Optical images illustrate that the variations in waterproofness among different OPVs can be attributed to the water stability of the active layer (Supplementary Fig. 28). The irreversible damage of aggregation dots could be observed in the water soaked films of PM6:Y6, PBDTTT-OFT:IEICO-4F, and PM6:BTP-eC9, whereas the water soaked film of PM6:ITCC showed only the swelling streaks that disappeared after we removed the residual water. Additionally, we verified that using PEI-Zn as an ETL supports the improved water stability compared to the conventional sol-gel processed ZnO as ETL (Supplementary Fig. 29). Optical images illustrate that ZnO induces aggregation (Supplementary Fig. 30), likely due to the chemical dissolution of ZnO layer when exposed to water54. The evolution of VOC, JSC, FF, and PCE of double-side encapsulated PM6:Y6 OPVs under MPPT is shown in Supplementary Fig. 31. The double-side encapsulated PM6:Y6 OPV device exhibited operational stability with T80 (80% of its initial value) lifetime of 27 h. Additionally, we measured the T80 lifetime of devices based on PBDTTT-OFT:IEICO-4F, PM6:BTP-eC9, and PM6:ITCC active layers. Devices based on PBDTTT-OFT:IEICO-4F active layer demonstrated T80 lifetime of approximately 110 h (Supplementary Fig. 32a), which is superior to the existing active layer solar cells (Supplementary Table 3). These results suggest that the differences in the operational stability of devices based on different active layers primarily stem from the active layer materials, and the operational stability of devices with AgOx/Ag electrode could be further improved by developing a stable active layer. The evolution of VOC, JSC, and FF of double-side encapsulated solar cells based on PBDTTT-OFT:IEICO-4F under MPPT is depicted in Supplementary Fig. 32b. Notably, the decay of VOC in PM6:Y6 solar cells is faster than that of JSC and FF, while the VOC of PBDTTT-OFE:IEICO-4F solar cell remains comparatively stable when compared to the variations in JSC and FF. The relatively swift decay of VOC in PM6:Y6 solar cells under operational stability could be ascribed to the changes in the microstructure of the active layer55.

Furthermore, we investigated the waterproofness of OPVs with ternary active layers of PM6:Y6:PC71BM, PBDTTT-OFT:IEICO-4F:PC71BM, and PM6:BTP-eC9:PC71BM (Supplementary Fig. 1). The J–V curves of ternary systems with in-situ grown AgOx HTL are shown in Supplementary Fig. 33. With the in-situ growth of AgOx HTL, these ternary OPVs demonstrated significant improvement in PCE, which is the same with the binary devices. The changes also suggest that the strategy with in-situ growth of AgOx HTL demonstrate excellent applicability for ternary solar cells with fullerene acceptor. All of these devices with ternary active layer systems also exhibited good waterproofness (Fig. 4f), further proving the universality of in-situ grown AgOx in enhancing the waterproofness. Moreover, these devices with ternary active layers demonstrated better operational stabilities compared to the devices with binary active layers (Supplementary Fig. 34). The best operational stability of free-standing OPVs was achieved using PBDTTT-OFT:IEICO-4F:PC71BM as an active layer, maintaining an estimated T80 lifetime of 149 h. These results suggest that using AgOx/Ag electrode strategy enables both waterproofness and operational stability by further optimizing the active layers.


In summary, we have successfully realized waterproof and ultraflexible OPVs with in-situ growth of AgOx HTL to strengthen the adhesion of the interface of HTL and the active layer. These OPVs offer remarkable stretchability and waterproof properties even with such thin structure, making them well-suited for wearable electronics. Notably, the efficiency degradation was limited to just 10% even after subjecting the devices to two washing cycles in a washing machine, each lasting 66 min. The optimization of waterproofness through structural design has enabled the OPVs to maintain excellent waterproof performance even with an ultrathin encapsulation layer. This approach has demonstrated broad applicability in enhancing the waterproof capabilities of ultraflexible OPVs. The successful implementation of waterproof organic OPVs holds promising implications for advancing self-powered wearable electronics and potentially underwater electronics for the Internet of Things in the foreseeable future.



PM6, Y6, BTP-eC9, IEICO-4F, and ITCC were purchased from 1-Materials (Quebec, Canada). PBDTTT-OFT was received from Toray Industries. PC71BM was purchased from Solenne BV Corporation. The precursor (ECRIOS VICT-Cz) of transparent polyimide was from Mitsui Chemicals (Japan). Zinc acetate dehydrates, 2-methoxyethanol, chloroform, and chlorobenzene were purchased from FUJIFILM Wako Pure Chemical Corporation (Japan). Ethoxylated polyethyleneimine (PEIE, 80% ethoxylated solution, 37 wt% in H2O), 1-chloronaphthalene (1-CN), and 1,8-diiodooctane (DIO) were purchased from Sigma-Aldrich. All materials were used as received without further purification.

Device fabrication

The tPI substrate with ITO electrode and PEI-Zn film were fabricated based on our previous study report24. PM6:Y6 (7.5 mg:9 mg) and PM6:Y6:PC71BM (7.5 mg:7.5 mg:1.5 mg) in chloroform with 1-chloronaphthalene (1-CN) solvent (1 mL, 99.5:0.5 volume ratio) was spin-coated at 3500 rpm for 40 s, followed by annealing at 110 °C for 10 min in an N2-filled glovebox. PBDTTT-OFT:IEICO-4F (10 mg:15 mg) and PBDTTT-OFT:IEICO-4F:PC71BM (10 mg:12 mg:3 mg) in chlorobenzene with 1-CN solvent (1 mL, 97:3 volume ratio) was spin-coated at 1400 rpm for 60 s, followed by annealing at 70 °C for 1 min in an N2-filled glovebox. PM6:ITCC (10 mg:10 mg) in chlorobenzene with 1,8-diiodooctane (DIO) solvent (1 mL, 99.5:0.5 volume ratio) was spin-coated at 2000 rpm for 60 s followed by annealing at 100 °C for 10 min in an N2-filled glovebox. PM6:BTP-eC9 (7 mg:8.4 mg) and PM6:BTP-eC9:PC71BM (7 mg:7 mg:1.4 mg) in chlorobenzene with DIO solvent (1 mL, 99.5:0.5 volume ratio) was spin-coated at 2500 rpm for 60 s followed by annealing at 100 °C for 10 min. Ag (100 nm) or MoOx/ Ag (7.5 nm/100 nm) were sequentially deposited (EX-200, ULVAC). The effective area of the solar cells was 4 mm2. The devices with Ag electrode were annealed in air at 85 °C. The parylene (diX-SR, Daisan Kasei) encapsulation layer was evaporated by chemical vapor deposition (PDS 2010, KISCO Company). The removable electrode was fabricated by evaporating Cr and Ag on poly(dimethylsiloxane) (PDMS) substrate.

Device characterization and measurement

The J–V and stability measurements were performed under 1 sun illumination (XES-40S3, SAN-EI ELECTRIC) and recorded using a Keithley 2400 source in forward direction under 20 °C. The D-SIMS measurement (ADEPT-1010, ULVAC-PHI, Inc.) was performed in TORAY Research Center, Inc. The light intensity was calibrated by using a silicon reference diode (BS-520BK, Bunkoukeiki). The work function was measured using photoemission spectroscopy (RIKEN KEIKI AC-3). The ultraviolet (UV) photons emitted by a UV lamp of KEIKI were focused on the surfaces of the Ag electrode. The energy of UV light can be controlled by a spectrometer. The photoelectrons escaped from the Ag electrode surface because of photoelectric effect. The escaped photoelectrons were counted by an open counter. The work function is the energy of a point of interest between the baseline and extrapolated line in the figure of square root plots of the photoemission yield. The EQE measurements were performed using monochromatic light (SM-250F, Bunkoukeiki) calibrated using a silicon reference diode. The stretching–compressing test was conducted by laminating the free-standing OPVs on a pre-stretched elastomer (VHB Y-4905J, 3 M). A screw machine controlled by a program with pre-designed parameters was developed to control the strain of compression and stretching. The period for one compressing–stretching cycle is approximately 4 s. The adhesion measurement was performed by a high-precision tensile tester (AG-X, Shimadzu). The holder was gradually moved at 5 mm/min until the samples were completely separated from the stud. The contact angles were measured using a contact angle meter (DMe-211, Kyowa Interface Science). The dispersive energy (\({{{{{{\rm{\gamma }}}}}}}_{{{{{{\rm{L}}}}}}}^{{{{{{\rm{d}}}}}}}\)) and polar surface (\({{{{{{\rm{\gamma }}}}}}}_{{{{{{\rm{L}}}}}}}^{{{{{{\rm{p}}}}}}}\)) of deionized water are 21.8 mJ/m2 and 51.0 mJ/m2, respectively. The \({{{{{{\rm{\gamma }}}}}}}_{{{{{{\rm{L}}}}}}}^{{{{{{\rm{d}}}}}}}\) and \({{{{{{\rm{\gamma }}}}}}}_{{{{{{\rm{L}}}}}}}^{{{{{{\rm{p}}}}}}}\) values of glycerol are 37.0 mJ/m2 and 26.4 mJ/m2, respectively. Here, the surface energies of water and glycerol are average values46. Owing to the similar polarities between PM6:Y6 and AgOxd > γp), but contrasting polarities compared to PEDOT:PSS and MoOxd < γp), the calculation model for Wa between AgOx and MoOx (or PEDOT:PSS) should be different. When calculating Wa between PM6:Y6 and AgOx, the Wa can be calculated using the following equation56:

$${{{{{\rm{Wa}}}}}}=2{({{{{{{\rm{\gamma }}}}}}}_{1}^{{{{{{\rm{d}}}}}}}{{{{{{\rm{\gamma }}}}}}}_{2}^{{{{{{\rm{d}}}}}}})}^{1/2}{+2({{{{{{\rm{\gamma }}}}}}}_{1}^{{{{{{\rm{p}}}}}}}{{{{{{\rm{\gamma }}}}}}}_{2}^{{{{{{\rm{p}}}}}}})}^{1/2}$$

The Wa between PM6:Y6 and MoOx (or PEDOT:PSS) should be calculated using the following equation56:

$${{{{{{\rm{W}}}}}}}_{{{{{{\rm{a}}}}}}}=\frac{4{{{{{{\rm{\gamma }}}}}}}^{{{{{{\rm{d}}}}}}}{{{{{{\rm{\gamma }}}}}}}_{{{{{{\rm{L}}}}}}}^{{{{{{\rm{d}}}}}}}}{{{{{{{\rm{\gamma }}}}}}}^{{{{{{\rm{d}}}}}}}+{{{{{{\rm{\gamma }}}}}}}_{{{{{{\rm{L}}}}}}}^{{{{{{\rm{d}}}}}}}}+\frac{4{{{{{{\rm{\gamma }}}}}}}^{{{{{{\rm{p}}}}}}}{{{{{{\rm{\gamma }}}}}}}_{{{{{{\rm{L}}}}}}}^{{{{{{\rm{p}}}}}}}}{{{{{{{\rm{\gamma }}}}}}}^{{{{{{\rm{p}}}}}}}+{{{{{{\rm{\gamma }}}}}}}_{{{{{{\rm{L}}}}}}}^{{{{{{\rm{p}}}}}}}}$$

The washing test was performed using a Panasonic washing machine (NA-FA80H7). The period of one standard washing cycle is 66 min, including washing, dewatering, and water injection procedures. The cross-sectional morphologies were observed by Thermo Fisher’s SEM (QuattroS, Thermo Fisher Scientific) under 10 kV after Os coat (2.5 nm). The cross sections were obtained through Ar ion milling for 12 h processing under 5.5 kV at 4 °C ± 5 °C (Leica, EM TIC 3X). The conductivity measurements were performed under dark and recorded using a Keithley 2400 source. The impedance tests were conducted using a potentiostat (AMETEK VersaSTAT 4) from 1 Hz to 500 kHz in the dark, at a bias voltage of 0.5 V.

Reporting summary

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