Broadband and omnidirectional antireflective film with bioinspired nanocone-grid hybrid structures for enhanced solar energy harvesting

Abstract

The accelerating development of solar energy conversion technology has raised an urgent demand for high-performance antireflective (AR) materials to enhance photon transmission. However, conventional AR materials are constrained by limited spectral bandwidth and pronounced angular dependence, leading to significant reductions in photon transmission under broadband and wide-angle conditions. Herein, we propose a broadband omnidirectional antireflective (BOAR) film with unique nanocone-grid-like (NCGL) hybrid structures, inspired by dragonfly wings. Optical characterization reveals that the biomimetic NOA61 film with the NCGL structure achieves approximately 96.3% transmittance in the visible spectrum at normal incidence, which is about 3.7% higher than that of the smooth NOA61 film. This transmittance enhancement remains at ∼2.9% within a 30° incidence range, and even at an extreme 75° incidence, the transmittance still reaches ∼78%. The excellent optical performance stems from the continuous, effective refractive index gradient constructed by the NCGL structure, which enables a smooth optical transition from air to the substrate by mitigating interfacial optical discontinuities. Moreover, the biomimetic BOAR film exhibits good environmental stability and mechanical properties. Notably, under AM1.5G simulated solar illumination at 25 °C, the solar panel integrated with the NOA61 film exhibits a 36.6% relative improvement in power conversion efficiency compared to that with the smooth film. In the 350–900 nm spectral range, the structured NOA61 film achieves an average 2.9% enhancement in external quantum efficiency. These results confirm that the NCGL hybrid structure enhances transmittance by suppressing Fresnel reflection, thereby increasing the photon flux into the active layer. This work provides new design principles for developing high-performance optical interfaces in photoelectric and photothermal conversion systems.

---

1. Introduction

The abrupt refractive index transition at air–substrate interfaces induces substantial Fresnel reflection, leading to significant optical energy losses. This detrimental effect critically impairs the performance of optoelectronic devices including photovoltaic cells,^1,2 photodetectors,³ and high-power laser optical components,⁴ where maximizing light transmission is critical for efficiency and functionality.^5,6 Anti-reflection (AR) technology has emerged as an indispensable solution, demonstrating remarkable performance enhancements across multiple domains.⁷ Conventional AR coatings, based on the quarter-wavelength interference principle, demonstrate notable optical performance limitations despite their technological maturity.^8,9 This severely restricts the photon manipulation capacity of these devices, ultimately degrading their optical performance in practical applications. To address these limitations, various micro/nanostructures including random pyramid arrays and nanopillars have been engineered on substrate surfaces.^10–12 They can enhance broadband omnidirectional antireflective (BOAR) performance by optimizing light–matter interactions across wider spectral and angular ranges, thereby improving device efficiency. However, due to the single structural design, the existing AR structures still have limitations in terms of incident angle adaptability and spectral bandwidth. This restricts the light modulation efficiency of the devices, indicating that there is still room for further improvement in their performance.^13,14 Over millions of years of evolutionary optimization, organisms in nature have developed various microstructures with remarkable AR properties, providing efficient solutions for nearly perfect optical management.^15,16 These biological prototypes have inspired the design of novel photonic structures. For instance, insect compound eyes (e.g., moth eyes, butterfly eyes, and fly eyes) exhibit excellent AR performance due to their unique nipple array structure. Inspired by the periodic arrays of moth eye, researchers have employed advanced techniques such as Langmuir–Blodgett deposition and nanoimprint lithography to fabricate various bioinspired high-transmission surfaces, achieving transmittance of up to 94.9% across the visible spectrum.^17–19 Chowdhury et al. fabricated large-area, uniform columnar nanovoid structures on germanium surfaces via a self-organization process, utilizing high-dose Sn⁺ ion irradiation combined with a silicon nitride capping layer technique. These structures exhibit a high aspect ratio and tunable pore depth, endowing them with excellent broadband and omnidirectional photon-trapping capabilities. Specifically, within the 350–1800 nm spectral range, the average reflectance is reduced by 67.1%, with a maximum reflectance reduction of up to 96% in a specific wavelength band around 960 nm. This work provides a novel approach for the development of high-efficiency optoelectronic devices.²⁰ In addition, to address the insufficient light absorption issue in the active layer of organic photovoltaics (OPVs), they prepared one-dimensional nanograting structures on silica substrate surfaces using laser interference lithography combined with reactive ion etching techniques. This nanostructured design enhances the light absorption efficiency of the OPV active layer by 19% across the visible and near-infrared spectral ranges, while increasing the short-circuit current density by 14%—thus achieving significantly improved broadband photon-trapping performance.²¹ Similarly, cicada wings also feature periodic nanopillar arrays analogous to those found in moth eyes, demonstrating a transmittance exceeding 91% at normal incidence in the visible range.^22,23 Drawing inspiration from this structure, they developed a bioinspired reversible AR film using a one-step imprinting technique, which achieves around 90% transmittance in the 400–900 nm wavelength range.²⁴ In recent years, research in this area has expanded to include various bioinspired optical structures, such as hollow soccer-ball-like micro–nano structures mimicking leafhopper wing brochosomal coatings,^25–27 micro–nano wrinkle patterns replicating the internal structures of dragonfly eyes,²⁸ and nano-window configurations derived from the wing scales of glasswing butterfly,^29–31 highlighting the broad potential of bioinspired photonics for AR applications. These biomimetic templates manipulate photon trajectories through precisely tuned structural dimensions and morphologies, thereby enhancing omnidirectional AR performance beyond conventional single-structure arrays. Nevertheless, significant challenges persist in matching the performance of ideal AR models while addressing fabrication complexity and scalability constraints.

Through environmental adaptation, dragonflies have evolved wings exhibiting exceptional BOAR properties coupled with multifunctional characteristics including self-cleaning surfaces, antimicrobial activity, and remarkable structural rigidity.³² The dragonfly wing's structure consists of two distinct components: a macroscopic vein network that simultaneously optimizes light transmission and mechanical toughness,³³ and microscopic membrane regions containing disordered nanocone arrays that enable superior photon management across wide spectral and angular ranges.^34,35 Inspired by this natural design, researchers have successfully replicated the wing's photonic structures using advanced fabrication techniques. For instance, Ding et al. employed femtosecond laser manufacturing technology to create dragonfly wing-inspired submicron truncated cone structures on MgF₂ glass, achieving an average transmittance of 99.896% across the 3–5 μm wavelength range.³⁶ More importantly, the transmittance remained no less than 70% even at a 75° incident angle. Additionally, the disordered nanocone structures mimicking dragonfly wing were constructed on polymer deformable material substrates using the self-assembly technology, demonstrating up to a 40% increase in transmittance compared to unstructured substrates at a 75° incident angle.³⁷ Although these studies confirm dragonfly nanocones as archetypal templates for bioinspired AR materials,^38,39 current research primarily focuses on single-structure parametric optimization (height/diameter variations),^40,41 while largely overlooking the synergistic effects of binary composite nanostructures crucial for achieving true omnidirectional performance.

To date, state-of-the-art fabrication techniques including precision laser ablation,^42,43 physical/chemical vapor deposition,^44–46 and the sol–gel processing have been extensively employed to engineer biomimetic AR surfaces.^47,48 For example, inspired by the nanostructures of glasswing butterfly and cicada wings, Papadopoulos et al. fabricated tilted nanocolumn arrays on glass substrates using ultrashort laser pulse technology, significantly enhancing omnidirectional AR performance for display devices.⁴⁹ In a separate approach, Nedelcu's team developed AR coatings via sputtering and thermal evaporation deposition, achieving 91% transmittance across the visible spectrum at 30° incidence.⁵⁰ However, when fabricating more complex nanostructures, these methods encounter issues such as difficulties in size and morphology control, and high processing costs.^51,52 Fortunately, the template method provides distinct advantages for creating nano-conical structures similar to those found on the transparent wings of biological organisms.^16,22,53 This approach is characterized by low cost, high efficiency, and excellent fidelity, facilitating rapid and stable testing of the optical properties of biological micro–nano structures. More importantly, by combining multiple templates produced using this method, the challenge of small template size due to the limited size of the original biological sample can be effectively addressed, thus enabling large-area processing of AR surfaces.^54,55

In this study, the Orthetrum triangulare dragonfly wings were chosen as the biological prototype due to their exceptional AR properties across the visible spectrum and at high incidence angles. The photon manipulation enhancement effect of the wing surface's unique grid structure on individual nanocone arrays was systematically analyzed. Subsequently, the nanocone-grid-like (NCGL) hybrid structures were successfully transferred onto polymer materials using bio-template-assisted replica molding technology. The optical performance of the biomimetic NCGL-structured film was analyzed in detail across a range of incidence angles. Besides, a series of environmental and mechanical tests—including ultraviolet irradiation exposure, high-temperature resistance, humidity, sand impact, and pencil hardness tests—were conducted to comprehensively evaluate the practical application potential of the biomimetic NOA61 film. The photoelectric conversion efficiency (PCE) of solar cells coated respectively with the biomimetic NCGL-structured NOA61 film and the smooth film was characterized under AM1.5G simulated sunlight at 25 °C. And the spectral response measurements were conducted to compare the external quantum efficiency (EQE) profiles of the two film structures across different wavelengths. Finally, to evaluate its practical application performance, the PCE of the monocrystalline silicon solar cell integrated with the biomimetic NCGL-structured film was tested under outdoor conditions.

2. Experimental

2.1 Materials

Dragonfly (Orthetrum triangulare) was purchased from Guangzhou Fanyu Art & Craft Trading Co., Ltd. The Sylgard 184 prepolymer solution, consisting of two parts polydimethylsiloxane (PDMS) elastomer and a curing agent, was purchased from Dow Corning Corporation. It is a widely used polymeric silicone elastomer, featuring high transparency, excellent flexibility, chemical stability and biocompatibility, with a refractive index ranging from 1.40 to 1.43. It is commonly employed as an imprint stamp or anti-reflective encapsulation layer for optical devices to effectively protect functional materials from environmental damage,⁵⁶ and was accordingly chosen here as the inverted structural substrate to prepare the biomimetic PDMS film. The UV-curing adhesive NOA61 from the United States was purchased from Chengjian Da Photoelectric Trading House in Bao'an District, Shenzhen. NOA61 is a transparent, colorless liquid photopolymer that, after curing, exhibits excellent transparency (refractive index ∼1.56) and good fluidity—properties that adequately meet the requirements for light transmittance and template wettability in optical components.¹⁶ Accordingly, it was selected here as the positive structural substrate for fabricating the biomimetic NOA61 film. The photovoltaic panels were purchased from Pingyu Xinyang Guangdian Technology Co., Ltd. The solar panel MPPT tester was obtained from Weixunda Electronic Equipment Factory.

2.2 Optical simulation

The simulation of NCGL hybrid structures and nanocone structures was performed using the finite-difference time-domain (FDTD) method, focusing primarily on the effect of ridge structures on the BOAR properties of an isolated nanocone structure. Further evaluation of the optical properties was conducted on the inverted PDMS structure master and the biomimetic NOA61 film. Based on Fourier transform infrared (FTIR) spectroscopic analysis, the primary component of the dragonfly wing was identified as α-chitin, which has a refractive index of approximately 1.56 (Fig. S1).^15,57 The refractive indices of the three materials involved in this study are as follows: dragonfly wing (α-chitin) ≈ 1.56, PDMS = 1.411,⁵⁶ and NOA61 = 1.56.¹⁶ In the direction of light propagation perpendicular to the substrate surface, a PML boundary condition was applied to ensure complete absorption of incident light that is reflected and penetrates the material. Based on the symmetry of the structural array, symmetric/anti-symmetric boundary conditions were applied on four planes perpendicular to the substrate when the plane wave was incident at zero deflection angle. For non-zero deflection angles, the boundary condition was set to Bloch. To enhance the accuracy of the simulation, the grid precision was set to 8, grid size to 2 nm, the simulation area was defined to cover three periods.

2.3 Preparation of the BOAR film with NCGL structures

To comprehensively validate the BOAR performance of the NCGL structures on dragonfly wings, a precise bio-template-assisted replica molding approach was employed to fabricate a PDMS master with negative nanostructures and a biomimetic NOA61 film featuring NCGL structures (Fig. 3a). The experimental procedure began with preparing dragonfly hind wing samples by sectioning them into 3 cm × 2 cm slices using a sterile scalpel, followed by sequential ultrasonic cleaning in deionized water and anhydrous ethanol to eliminate surface contaminants. After drying the cleaned wings at 40 °C for 10 minutes to remove residual liquids, the lower wing surface was firmly affixed to a flat slide using double-sided tape to establish a stable working platform. For negative mold fabrication, a PDMS solution was prepared through thorough mixing of Sylgard 184 polymer elastomer and curing agent in a 10 : 1 weight ratio, with continuous stirring for 5–10 minutes to ensure homogeneity before carefully pouring the mixture over the wing's upper surface. To achieve bubble-free replication, the sample underwent vacuum degassing at −0.6 MPa for 1.5 hours before thermal curing at 60 °C for 2 hours, after which the fully cured PDMS negative mold was delicately separated from the biological template. The final biomimetic NOA61 film was then produced through UV nanoimprint lithography. For this process, approximately 150 μL of NOA61 resin was spin-coated onto the PDMS mold at 1500 rpm for 2 minutes to ensure uniform distribution, followed by 2.5 hours of UV exposure to complete photopolymerization. Upon complete curing, the PDMS mold should be carefully delaminated from the polymerized NOA61 substrate to yield the final BOAR film featuring NCGL hybrid structures. The negative PDMS template remains reusable following ultrasonic cleaning in deionized water, maintaining structural fidelity across multiple replication cycles.

2.4 Characterization

The surface morphology of the wings and biomimetic films was observed using a scanning electron microscope (SEM, Hitachi Regulus 8100) and an atomic force microscope (AFM, Bruker Dimension Icon). The optical properties of both the wing and the biomimetic films were measured using a fiber optic spectrometer in the wavelength range of 350–850 nm and at AOI ranging from 0° to 75°. Fourier transform infrared (FTIR, Thermo Fisher Scientific Nicolet iS20) spectroscopy was used to characterize the functional groups on the surface of dragonfly wings. The UV stability of the biomimetic NOA61 film was tested using a 20 W LED UV curing lamp. The thermal stability of the biomimetic NOA61 film at different temperatures was evaluated using a vacuum drying oven (BPZ-6063). The moisture resistance of the biomimetic NOA61 film was evaluated using a constant temperature and humidity standard curing chamber (ELB-SHBY-40B). The surface morphology of the biomimetic NOA61 film after the sand impact experiment was observed using an ultra-depth-of-field microscope (Easyzoom5). The scratch resistance of the biomimetic NOA61 film was evaluated using a pencil hardness tester (QHQ-A). Current–voltage (I–V) characteristic and external quantum efficiency (EQE) were measured for solar cells coated with the NCGL structured NOA61 film and the smooth film, respectively, using an electrochemical workstation (CHI660D) under AM1.5 standard solar illumination conditions.

3. Results and discussion

3.1 Biological characteristics and analysis

As insects living in well-lit environments, the wings of Orthetrum triangulare dragonfly exhibit excellent AR properties. Under strong light, the word ‘university’ is clearly visible behind the wings (Fig. 1a). The optical properties of the wings’ surfaces were tested using a fiber optic spectrometer. As shown in Fig. 1b, the average reflectance of the dragonfly's forewings in the natural light wavelength range is only about 1.5%, while the hindwings display even better optical characteristics, with an average reflectance of approximately 0.7%. Fig. 1c shows that within the 350–850 nm wavelength range, the transmittance of the hindwings decreases by only about 7% as the angle increases from 0° to 75°, demonstrating excellent angle-insensitive performance. The exceptional broadband, omnidirectional AR performance of the dragonfly's wings is closely related to their structural features. Fig. 1d shows that these wings have a distinctive web-veined structure, with primary veins forming quadrilateral networks and secondary veins predominantly pentagonal and hexagonal. This design provides dragonfly wings with exceptional toughness, allowing them to maintain structural integrity during high-speed flapping while directing more light onto the transparent wing membrane. Additionally, the surface of the wing membrane is covered with numerous NCGL hybrid structures (Fig. 1e), which effectively reduce Fresnel reflection on the wing surface. These nanotapers have an average height of approximately 268.5 ± 39.4 nm, an average top diameter of around 44.2 ± 5.7 nm, an average base diameter of about 93.4 ± 25.3 nm, and an average period of approximately 99.2 ± 20.4 nm.

Fig. 1 Structure and exceptional optical properties of the wings of Orthetrum triangulare dragonfly. (a) The biological sample is placed on a piece of paper with 'University' written on it. (b) Reflectivity of the forewing and hindwing surfaces. (c) Transmittance spectra of dragonfly wings at various angles. (d) Mesh venation. (e–g) SEM images of the wings observed from the side and top views. (h) AFM image of the wings. (i) Profiles of different regions of the wings. (j) Schematic illustrating the mechanism of BOAR characteristics enabled by nanostructures on dragonfly wing surfaces.

Two nanotaper structures are closely connected by a ridge structure, which has a peak of about 160 nm, a trough of approximately 108 nm, and an average thickness of 21.2 ± 2.7 nm. Four nanotapers and four nanoridge structures together form NCGL hybrid units arranged in a quadrilateral pattern (Fig. 1f and g). The NCGL hybrid structures were characterized by using AFM (Fig. 1h and i), and three scans revealed that the height of the structures follows a normal distribution. Extensive research has established that nanostructures with normally distributed height dimensions consistently demonstrate optimized omnidirectional AR performance, where the synergistic integration of heterogeneous nanostructural configurations further enhances their photonic control capabilities.⁵⁸ This mechanistic insight explains the exceptional omnidirectional AR properties observed in dragonfly wings, whose biological nanostructures inherently combine these two critical design principles through evolutionary optimization.

As is well established, when light irradiates the surface of dragonfly wings, both macro- and micro-scale light-regulatory mechanisms coexist. As illustrated in Fig. 1j, incident light first undergoes diffuse reflection at the dark veins of the wing, causing a small portion of light to scatter back into the air along disordered paths, while the remaining light is concentrated toward the central transparent regions by these veins. This pre-conditioned light then interacts with the NCGL hybrid nanostructures, initiating the microscopic regulation phase. Under vertical incidence conditions, incident photons undergo multiple interfacial reflections between adjacent nanocones, with eventual coupling into the basal layer through the conical morphology's waveguiding properties. Concurrently, upon illumination of the grating regions within the nanocone array, an intrinsic light-guiding effect efficiently channels optical energy toward the basal interfaces of neighboring conical structures through directional photon coupling. The unique advantage of the grating configuration becomes fully pronounced under oblique illumination. Under such conditions, it effectively intercepts photons escaping through gaps between neighboring nanocones, thereby enhancing the omnidirectional AR performance of the standalone conical architecture. Furthermore, the grating structure increases the fill factor of the overall system. To validate this photonic mechanism, we constructed two models: an isolated nanocone array and an integrated NCGL hybrid array. Both models share identical conical units (periodicity L), while the latter incorporates an additional nanograting structure (height h₂) (Fig. 2a). By defining the center of the conical base as the origin of the Cartesian coordinate system, the nanocone unit can be mathematically represented as a surface of revolution generated by rotating a predefined curve about the z-axis:⁵⁹

Z = h[1 − (R/r_d)²] (1)

Fig. 2 Analysis of the optical performance of the isolated nanocone array and the NCGL hybrid array. (a) NCGL cell. (b) Nanocone cell and equivalent nanocone cell. (c) Equivalent refractive index curves of two cells. (d) Transmission and reflection of the isolated nanocone array within the 350–850 nm wavelength range. (e) Transmission and reflection for incident angles from 0° to 75°. (f) Electric field strength of the isolated nanocone array. (g) Transmission and reflection of the NCGL hybrid structure array within 350–850 nm wavelength range. (h) Transmission and reflection for incident angles from 0° to 75°. (i) Electric field strength the NCGL hybrid structure array.

In the formulation, where h denotes the height of the unit nanocone and r_d represents its base radius, the volume of the unit nanocone (A₁) can be calculated. The grating unit, modeled as a truncated pyramid with trapezoidal cross-section of thickness B, yields a volume A₂. Since the NCGL hybrid structure integrates one nanocone unit with four surrounding grating units, its total volume (A₃) is derived as A₃ = A₁ + 4A₂. Following eqn (1), the NCGL unit can be equivalently described as a nanocone with base radius r_e and height h, generated by rotating its defining contour about the z-axis. As illustrated in Fig. 2b, the substrate surface is tessellated into hexagonal cells with center-to-center spacing D, each circumscribed by a circle of radius r_e. The fill factors (f₁ and f₂) of both the isolated nanocone array and NCGL array models can be independently calculated using this geometric framework:

Owing to the continuous contour of the nanostructures, the protrusions on the surface of dragonfly wings can be treated as a superposition of multiple thin films with continuously varying refractive indices. Consequently, the Bruggeman effective medium approximation (EMA) is applicable for addressing this phenomenon:⁶⁰

Based on this, the effective refractive indices (n_e1 for the isolated nanocone array and n_e2 for the integrated NCGL hybrid array) of their respective air-composite media systems can be mathematically expressed as:⁶¹

n_e1 = [f₁n^2/3₁ + (1 − f₁)n₀]^3/2 (5)

n_e2 = [f₂n^2/3₁ + (1 − f₂)n₀]^3/2 (6)

The refractive indices are defined as n₀ = 1.00 for air and n₁ = 1.56 for chitosan,^15,57 with their volume fractions satisfying f₁ = 0.9f₂. As detailed in SI, the computational procedures for these parameters are systematically documented. As shown in Fig. 2c, as the filling fraction increases, the NCGL graded structure achieves a continuous and smooth transition in effective refractive index from 1.00 to 1.56, exhibiting a uniform trend without abrupt changes in slope. In contrast, the isolated nanocone array reaches a maximum effective refractive index of only approximately 1.50, leading to a significant refractive index discontinuity at the substrate interface. This discrepancy primarily arises from the introduction of grid units in the NCGL structure, which enhances the spatial packing density and geometric coupling of the conical units. Consequently, the overall filling fraction is increased, and a more optimized gradient path for the effective refractive index is enabled through material distribution. From the perspective of effective medium theory, these results confirm that the incorporation of the grid structure significantly improves the continuity of the refractive index profile, thereby effectively enhancing the overall optical control capability of the structure. Compared to the isolated nanocone structure, this behavior confirmed quantitatively the superior light-modulating capability of the NCGL structure, enabling the precise and predictable design of gradient refractive index profiles (SI).

To investigate the influence of ridge structures on the optical characteristics of isolated nanocone arrays and AR performance between two models under wide-angle light incidence, the reflectivity, transmittance, and electric field intensity were analyzed using the FDTD method. A plane wave was incident perpendicular to the array direction as the light source. For the broadband AR simulation, the incident light wavelength ranged from 350 nm to 850 nm, and R and T monitors were used to measure reflectivity and transmittance. In the omnidirectional AR simulation, the incident light wavelength was fixed at 580 nm, and the angle θ was gradually increased to 75°, with Bloch boundary conditions applied around the array structure. Additionally, an isolated nanocone structure array was created with a height of 270 nm, a base diameter of 92 nm, a top diameter of 44 nm, and a period of 100 nm. A NCGL hybrid structure array with ridge features was also constructed, where the ridge peak is 160 nm, the valley is 108 nm, and the ridge thickness is 20 nm. Fig. 2(d–f) and (g–i) show the BOAR performance and electric field intensity for the isolated nanocone structure array and the NCGL hybrid structure array, respectively. As shown in Fig. 2d and g, both models exhibit excellent AR performance across the natural light wavelength range. The average reflectivity of the nanocone array was approximately 2.7%, varying from 2% to 3.3% in the 400–630 nm range, and stabilized around 3% thereafter. The reflectivity of the NCGL hybrid array was generally about 1.4% lower than that of the nano-cone array, with a particularly low reflectivity of only 0.7% at 580 nm. As demonstrated in Fig. 2e and h, both models exhibit stable AR performance across incident angles of 0°–60°, with the NCGL hybrid array achieving approximately 0.86% lower reflectivity compared to the isolated nanocone array. Both arrays reached their lowest reflectivity at around 20°. However, when the angle of incidence (AOI) was between 60° and 70°, the photon manipulation capability of both models sharply decreases, with reflectivity increasing by about 10%. Nevertheless, the AR performance of the NCGL hybrid array remained superior to that of the isolated nanocone array.

To investigate the AR mechanism, specifically the interaction between incident light and structures, the electric fields of two array designs were simulated (θ = 0°, λ = 580 nm). A plane wave was chosen as the incident source from the far field to facilitate analysis and measurement. FDTD software's monitoring module was then used to observe the electric field intensity within both arrays. The incident light experiences multiple interfacial reflections among the conical structures, ultimately becoming confined at the base of the nanocone array, resulting in significantly enhanced electric field intensity within this region. Remarkably, in the NCGL hybrid array, incident photons experience multiple interfacial reflections coupled with ridge-mediated optical guidance, inducing a photonic redistribution where normally converging base-focused light becomes spatially redistributed within the structure. This phenomenon generates an inverted electromagnetic field gradient, with bottom-region intensities moderately attenuated relative to top-region counterparts (Fig. 2f and i). Overall, the incorporation of ridge structures significantly enhances the BOAR performance of nanocone arrays through strategic modification of photon propagation pathways. By redirecting light–matter interaction dynamics, this architecture enables markedly improved optical transmission within the array. The design establishes an advanced solution for optimizing AR surfaces in photothermal conversion systems and photoelectric absorption devices.

3.2 Analysis of the optical properties of the biomimetic BOAR film

Fig. S2 displays the inverse-structured PDMS master and the biomimetic BOAR NOA61 film, respectively. Under strong light, both films exhibit excellent transparency, allowing the flowers behind them to be clearly visible. The blurring in the structured areas was due to the unavoidable replication of the supporting wing vein structure during the molding process. Although the material transitioned from black to transparent, the macrostructure caused significant light reflection near the mesh pattern, resulting in a frosted glass-like effect. Currently, there is no method to fully eliminate the mesh support structure without damaging the wing membrane area. In Fig. S3, to test the flexibility of the PDMS film, it is manually bent to nearly 180° and immediately returns to its original shape once the external force is removed. This flexibility enabled the PDMS master to be used for curved imprinting, showing great potential for applications on the surfaces of curved optical devices. Additionally, Fig. 3b and c provide a comparative visualization of light reflection behavior between unstructured and nanostructured surfaces under controlled illumination conditions. Using a 650 nm laser source at a fixed 30° incidence angle, the unstructured substrate produced a distinct diffraction-limited spot on the observation plane, whereas the bioinspired structured surface completely suppressed visible scattering, demonstrating its superior photon management capability. This ensures that more light passes through the slide, minimizing light energy loss. The SEM image of the inverse-structured PDMS film is shown in Fig. S4.

Fig. 3 The fabrication process of the BOAR film and the structural characteristics and optical properties of the experimentally prepared AR films. (a) Schematic of the fabrication process for the biomimetic BOAR film. This process included the acquisition of the biological template and the creation of the BOAR film. Illustration: structural changes during the secondary molding process, with markers alongside beside the image. (b) Schematically illustrates the laser beam reflection pathway on an unpatterned substrate, whereas (c) demonstrates the modified optical trajectory resulting from the NCGL hybrid structured surface. (d) SEM image of the BOAR film. (e) Transmittance spectrum of the structured film compared to the unstructured film within the 350–850 nm wavelength range. (f) Presents the angular-dependent transmittance spectra (0°–75° incidence) of the inverse-structured PDMS film, while (g) displays the corresponding BOAR performance of the biomimetic NOA61 film across identical angular conditions.

Fig. 3d clearly resolves the fine nanostructures on the NOA61 film surface, featuring precisely oriented inclined nanocones with periodic nano-ridges interspersed between them, which conclusively demonstrates the successful fabrication of a NCGL hybrid structure. Subsequent optical characterization revealed that the bioinspired PDMS film achieves a remarkable 7.1% increase in transmittance across the visible spectrum (350–850 nm) relative to its unstructured counterpart, as quantified in Fig. S5. Simultaneously, the biomimetic BOAR NOA61 film exhibits a consistent 3.7% broadband transmittance enhancement, with performance metrics detailed in Fig. 3e. Notably, the biomimetic NOA61 film demonstrates less than 2% transmittance enhancement in the 350–400 nm range compared to the non-structured film, which we attribute to the material's inherent UV sensitivity. During photopolymerization, the NOA61 adhesive's strong absorption of 350–380 nm wavelengths increased surface reflectivity, thereby diminishing the nanostructures’ AR efficacy in this spectral region. On the other hand, across the 400–850 nm visible to near-infrared range, the structured film's transmittance progressively increased from 93.8% to 96.3%, with the performance gap between structured and non-structured films systematically widening with wavelength. This wavelength-dependent behavior clearly demonstrates the enhanced photon management capability of the NCGL hybrid architecture within natural sunlight spectra. Collectively, both the NCGL hybrid structure and its inverse replica demonstrate significant broadband transmittance improvement, confirming their versatile light-harvesting potential.

To evaluate the omnidirectional AR performance of both the NCGL hybrid structures and their inverse counterparts across the 350–850 nm spectral range, the AOI was incrementally varied from 0° to 75°. As revealed in Fig. 3f, the inverse-structured PDMS film exhibits a characteristic angular dependence, with transmittance progressively decreasing from 97.8% at normal incidence (0°) to 79.7% at 75° AOI. Notably, the film maintains excellent transmittance (>90%) within the 0°–30° range, demonstrating robust wide-angle performance. Even at 45° incidence, the transmittance remains comparable to that of the non-structured film, with only marginal reduction, highlighting the effectiveness of the bioinspired design in preserving optical efficiency across moderate angles. Furthermore, with increasing incidence angles, the mesh support structure induces pronounced light scattering upon illumination, significantly attenuating the total transmitted light intensity through the film. As shown in Fig. 3g, the biomimetic NOA61 film exhibits similar angular dependence, with its transmittance advantage diminishing beyond 45° incidence and ultimately underperforming the non-structured film at extreme angles (∼78% at 75°). This performance limitation likely stems from scattering losses induced by the support structure, suggesting that optimized fabrication approaches eliminating such constraints could potentially enhance the film's omnidirectional transmittance characteristics.

3.3 FDTD simulations of the biomimetic BOAR film

To evaluate the intrinsic optical performance independent of the mesh support structure, we conducted FDTD simulations of both inverse-structured PDMS films and biomimetic BOAR NOA61 films. The simulation geometry, illustrated in Fig. 4a, employs a plane wave source (350–850 nm wavelength range) propagating normal to the array surface, with angular sweeps performed from 0° to 75° incidence to comprehensively characterize the structures’ angular-dependent response. Optical monitors were positioned to quantify the transmittance spectra and electric field intensity for both structures. The simulation domain employed perfectly matched layer (PML) boundary conditions along the light propagation axis to eliminate back reflections, while periodic Bloch boundary conditions were implemented parallel to the array plane to accurately represent the infinite periodic nature of the nanostructures.

Fig. 4 FDTD simulation comparing the optical performance and electric field distribution of the two models. (a) Setup of the FDTD simulation model. (b) Transmittance and reflectance spectra of the NCGL hybrid structure and its inversed structure under natural light, obtained from the FDTD simulation. (c) Transmittance spectra of both structures as a function of incident angle. (d–f) Electric field intensity distribution of the inversed NCGL hybrid structure at angles of 15°, 45°, and 75° (from left to right). (g–i) Electric field intensity distribution of the NCGL hybrid structure at angles of 15°, 45°, and 75° (from left to right).

As demonstrated in Fig. 4b, when isolated from vein-induced optical interference, both nanostructured films achieve exceptional optical performance, exhibiting peak transmittance of 98.9% with correspondingly minimal reflectance. Fig. 4c further reveals their wide-angle characteristics under idealized conditions: the films maintain remarkably stable transmittance (∼98.6%) across 0°–60° incidence, with divergence emerging at higher angles. Between 60°–75°, the inverse-structured PDMS film shows moderate attenuation (8.5% reduction), while the biomimetic BOAR NOA61 film experiences greater transmittance loss (12.2%). This observed divergence likely stems from the inherent material refractive index contrast (n_PDMS = 1.411, n_NOA61 = 1.56), which fundamentally alters their photon interaction dynamics. Furthermore, our analysis suggests that the hole-type morphology exhibits superior angular tolerance compared to conical architectures, potentially due to enhanced phase matching at oblique incidences.

To elucidate the underlying optical mechanisms, we simulated the electric field distributions of both film types across varying incidence angles as shown in Fig. 4d–i. As the angle increases, the maximum electric field intensity inside the inversed NCGL hybrid structures decrease from 2.05 to 1.12 (Fig. 4d–f). Fig. 4f specifically reveals that at 75° incidence, the structure's asymmetric light–matter interaction becomes dominant: while the right facet of the inverted cone serves as the primary photon capture surface, the left facet demonstrates substantially reduced light manipulation capacity. This geometrically induced asymmetry directly correlates with diminished overall light trapping efficiency, ultimately causing the observed significant transmittance reduction at extreme angles. Fig. 4g–i show the angular-dependent electric field distribution of the biomimetic BOAR NOA61 film, revealing greater sensitivity to incidence angle variation compared to the inverse-structured PDMS film. Quantitative analysis shows a 1.718 reduction in electric field intensity as the incident angle increases from 15° to 75°. This suggests that top-down fabricated nanostructures exhibit superior light-coupling efficiency at extreme angles. Additionally, as the angle increases, the ridge-like structures sustain significant transmittance enhancement by continuously redirecting photons into the substrate. However, the NCGL hybrid structure's asymmetric angular response becomes dominant at 75° incidence, where preferential light interaction with the left-side nanostructures leads to progressive transmittance degradation. The decrease in transmittance of the biomimetic NOA61 film with increasing incident angle primarily stems from the synergistic effect of angle-induced effective structural asymmetry and material refractive index difference. Under normal incidence, the NCGL nanostructure enables symmetric scattering and efficient redistribution of photons, combined with graded refractive index matching at the NOA61-air interface, which effectively suppresses Fresnel reflection and results in high transmittance. However, this balanced state is disrupted at oblique angles. On one hand, anisotropic constraints from nanostructural geometric dimensions (e.g., nanocone height and grid thickness) cause non-uniform photon path deviations at the cone-grid interface under oblique incidence, thus inducing effective structural asymmetry. This asymmetry leads to partial photon localization through multiple reflections or deviation from the transmission direction, reducing transmitted energy. On the other hand, the significant refractive index difference between NOA61 (n ≈ 1.56) and air (n ≈ 1.00) becomes more pronounced with increasing angle. The actual incident angle at the interface increases, enhancing Fresnel reflection, while the altered effective interaction ratio between photons and the two media (air/NOA61) in the NCGL structure deviates the effective refractive index profile from the optimal matching state. This further disrupts propagation symmetry and aggravates energy attenuation in the transmission direction. Collectively, the combined action of the effective structural asymmetry and amplified material refractive index difference under large angles leads to the observed gradual decline in transmittance.

3.4 Analysis of environmental stability and mechanical properties of biomimetic BOAR film

To systematically evaluate the optical performance and environmental adaptability of the biomimetic NOA61 film, a series of tests were conducted, including ultraviolet (UV) irradiation, high-temperature resistance, humidity tolerance, sand impact, and hardness tests. Fig. 5a and b show the transmittance curves of the film after 144 hours of UV exposure under normal incidence and 30° incidence, respectively. The results indicate that the average transmittance decreased from an initial 96% to approximately 94% after 144 hours at normal incidence, and from 87% to about 84% at 30° incidence. Similar trends were observed at 60° and 75° incidence angles (Fig. S6a and S6b). This phenomenon is primarily attributed to the photo-aging effect of the NOA61 material under prolonged UV exposure, while the slight 2% performance degradation indicates the excellent UV stability of the biomimetic NOA61 film. Fig. 5c and Fig. S6c–e present the transmittance curves of the biomimetic NOA61 film measured after being maintained for 20 minutes at different temperatures (25 °C–105 °C) in a vacuum drying oven.

Fig. 5 Characterization of environmental stability and mechanical properties of the bioinspired NOA61 film. Transmittance curves of the film after 144 hours of ultraviolet (UV) irradiation under (a) normal incidence and (b) 30° incidence. (c) Transmittance curves of the film under normal incidence within the temperature range of 25–105 °C. (d) Transmittance variation curves of the film at multiple incidence angles in an environment with 99% relative humidity and 20 °C. Transmittance curves of the film after the sand impact test under (e) normal incidence and (f) 30° incidence. (g) Damage morphology of the NCGL structure on the film surface after 50 sand impacts. (h) Transmittance variation curve of the film under normal incidence after the pencil hardness test. (i) Micro-damage morphology of the NCGL structure at the scratch induced by the 3H pencil.

As shown in Fig. 5c, under normal incidence, the average transmittance remains stable at approximately 96% despite increasing temperature, indicating no significant impact from thermal variation. However, as shown in Fig. S6c, at a 30° incidence angle, the transmittance decreases from 89% (at 25 °C) to around 87% (at 105 °C), with similar declining trends observed at 60° and 75° incidence angles (Fig. S6d and S6e). The angular dependence of the optical performance after thermal exposure across the 25 °C–105 °C range can be attributed to the synergistic effect of thermal stress and optical path changes. Under normal incidence, the shortest optical path and symmetric interaction with the NCGL structure make the graded refractive index design insensitive to thermally induced micro-deformations, thereby maintaining transmittance stability. In contrast, under oblique incidence, the increased effective optical path through the material, combined with minor localized nano-structural distortions induced by thermal stress, disrupts the initially optimized phase-matching conditions. This enhances interfacial scattering and disrupts diffraction behavior, leading to a slight reduction in average transmittance. Nevertheless, even after high-temperature exposure, the film retains an average transmittance above 73% at a high incidence angle of 75°, demonstrating that temperature variations do not severely affect its transmittance and confirming its excellent thermal stability. Furthermore, to evaluate the moisture resistance of the film, the biomimetic NOA61 film was placed in a constant-humidity chamber (99% relative humidity, 20 °C) for 36 hours, after which its transmittance was measured (Fig. S6f). As shown in Fig. 5d, a comparison of the test results before and after humidity exposure reveals no significant changes in the average transmittance of the film across different incidence angles, with only minor fluctuations being observed, thus confirming that the film retains excellent optical stability even under high-humidity conditions. Fig. S6g illustrates the schematic diagram of the sand impact test setup for the biomimetic NOA61 film, in which the sample was fixed on a glass slide tilted at 45° and subjected to sand impacts from a height of 50 cm through a funnel and a 30 cm long glass tube, with each impact using 8.5 g of sand and transmittance recorded every 5 impacts. Fig. 5e and f illustrate the variation of the film's transmittance during the sand impact process. Under normal incidence, the transmittance gradually decreased with increasing impact cycles but remained around 65% after 50 impacts. At 30° incidence, a similar decreasing trend was observed, with transmittance reaching approximately 45% after 50 impacts. Comparable patterns were also noted at 60° and 75° incidence angles (Fig. S6h and S6i). Fig. S7a and S7b indicate that this phenomenon is primarily attributed to the attached fine sand particles and the induced surface micro-damage, while the SEM images after 50 impacts further clearly reveal significant degradation of the surface NCGL nanostructure (Fig. S7c and 5g), which directly accounts for the observed reduction in film transmittance. Fig. S7d illustrates the experimental procedure for evaluating the pencil hardness of the biomimetic NOA61 film, in which the film was placed horizontally and subjected to scratching by pencils with increasing hardness (2H, 3H, 4H) under a constant 750 g load at a 45° angle (Fig. S7e). Fig. 5h displays the transmittance curves of the film surface after the pencil hardness tests. The results demonstrate that after the 2H pencil test, the transmittance at normal incidence remained approximately 96% compared with that of the initial untested film. No significant changes in transmittance were observed at 30° and 60° incidence angles, while a decrease of about 2.5% occurred at 75° incidence. Notably, no visible scratches were observed on the film surface under these testing conditions. After the 3H pencil test, the normal incidence transmittance decreased by approximately 3%. Although no damage was visible to the naked eye, SEM images revealed damage to the NCGL structure along the scratch path (Fig. S7f and 5i). Transmittance decreased across all incidence angles, yet remained around 70% at 75° incidence (Fig. S7g–i). In contrast, visible scratches and significant transmittance reduction occurred when using the 4H pencil. Based on the above results, the biomimetic NOA61 film is determined to have a pencil hardness grade of 3H, demonstrating a favorable level of scratch resistance.

3.5 Application of the biomimetic BOAR film on solar cells

To evaluate the impact of the biomimetic NCGL-structured NOA61 film on photovoltaic performance, current–voltage (I–V) characteristics of solar cells coated with this structured film and those with a smooth NOA61 film were measured under AM1.5 standard solar illumination. The I–V tests showed that compared to the smooth NOA61 film, the device integrated with the biomimetic NCGL structure exhibited significant improvements in both open-circuit voltage (V_oc) and short-circuit current (J_sc) (Fig. 6a). This elevated the output power from 0.01453 W to 0.01985 W, corresponding to a relative enhancement of 36.6%. This performance improvement is primarily attributed to the wide-angle and broadband AR properties of the NCGL nanostructure, which effectively enhances the capture and total transmission of incident light. More optical energy is thus able to penetrate the active layer of the solar cell, ultimately yielding higher photoelectric output. Besides, to evaluate the optical performance of the biomimetic NCGL-structured NOA61 film, we comparatively measured the EQE of solar cells coated with this structured film versus the smooth NOA61 film across different wavelengths. Fig. 6b shows the EQE spectra and corresponding integrated current curves for both the biomimetic NCGL-structured NOA61 film and the smooth NOA61 film. The results demonstrate that over the 350–900 nm spectral range, the biomimetic structured NOA61 film exhibits significantly higher EQE values than the smooth film, with the calculated average EQE improved by 2.9%. Concurrently, the integrated current density derived from the EQE spectra increased by 2.6%, confirming that the NCGL structure effectively enhances the device's PCE across the entire working spectrum by improving light transmission capability. To evaluate the practical application performance of the biomimetic NOA61 film, we developed a variable-angle photovoltaic characterization device. Using this device, we tested the PCE of the monocrystalline silicon solar cell integrated with this film under outdoor conditions (Fig. S8a). Identical-area samples of both unstructured and structured films were mounted on monocrystalline silicon solar panels with an adjustable-angle goniometric stage. Experimental measurements were conducted during solar noon (12:00–13:00 local time) when the solar altitude angle reached ∼25°. The initial test was conducted at a panel tilt angle of θ = 65° to achieve normal light incidence. Subsequently, systematic characterization of the electrical performance was carried out at different deflection angles to evaluate the device's angular adaptability. Specifically, the I–V characteristics (Fig. 6c, d and Fig. S8c–e) and V–P characteristics (Fig. 6e, f and Fig. S8f–h) were measured under tilt angles of θ = 40° (with a deflection of 25°), θ = 20° (with a deflection of 45°), θ = 0° (with a deflection of 65°), and θ = –10° (with a deflection of 75°), respectively. Fig. S9 characterizes the angular-dependent PCE of photovoltaic modules under varying illumination conditions. The experimental results demonstrate that varying the incidence angle (θ) from 65° to 20° produces largely unchanged in short-circuit current while generating a consistent 0.2 V enhancement in open-circuit voltage, attributable to improved charge carrier collection efficiency. Under optimal normal incidence conditions (θ = 65°), the biomimetic structured film achieves a notable PCE of 3.84%, representing a 12.3% relative improvement over the unstructured reference sample (3.42%, Fig. 6c), thereby conclusively validating the enhanced light-harvesting capability of the engineered nanostructures. As the angle between the light and the photovoltaic panel increases to 25° (θ = 40°), the structured film-coated photovoltaic panel demonstrates a 4.9% relative enhancement in PCE (Fig. 6d). This improvement becomes more pronounced at 45° deflection (θ = 20°), where the PCE increases by 20.7% relative to the unstructured reference (Fig. S8c). Tests under specific large incident angles (65° and 75° relative to the film) revealed the following results. As shown in Fig. S8d, at 65° incidence (corresponding to θ = 0°), the biomimetic NCGL-structured film achieved a PCE of 12.04%—markedly exceeding the 9.31% of the smooth film. Even under weak-light conditions at an increased incident angle of 75° (θ = −10°), the biomimetic structure still maintained a noticeable relative advantage in efficiency (Fig. S8e). This result highlight the discrepancy between laboratory measurements and comprehensive outdoor performance evaluations. The outdoor environment, characterized by more intense global irradiation and multi-directional diffuse light, collectively contributes to higher photoelectric conversion efficiency. In contrast, laboratory tests typically use highly collimated light spots with concentrated energy flux. Consequently, over a wide range of incidence angles, the NOA61 film with the NCGL structure exhibits superior transmittance for diffuse light—the dominant light component under real outdoor conditions. These findings thus underscore the significant potential of the NCGL structure for photovoltaic applications, particularly stemming from its capacity to sustain efficiency enhancement across a broad range of solar altitude angle variations. This observed performance enhancement arises from fundamental differences in light–matter interactions at the material interface. Conventional unstructured films incur significant Fresnel losses due to the abrupt refractive index transition between air and the substrate, which causes substantial light reflection back into the air. In contrast, the biomimetic structured film generates a gradual refractive index gradient that enables efficient photon transfer from air to the substrate. Fig. S8b illustrates the sophisticated light management mechanism of the nanostructured surface, where incident photons undergo multiple internal reflections and refractive events both at the structure interfaces and within the nanocavities. This intricate optical interplay creates a pronounced light-trapping effect, effectively confining photon energy within the nanostructured layer and photovoltaic active region, thereby significantly enhancing the overall PCE. The demonstrated capability to manipulate photon propagation paths with such precision highlights the remarkable potential of this architecture for advanced optoelectronic and photothermal energy conversion systems, particularly in applications requiring efficient operation under variable angle.

Fig. 6 Photovoltaic conversion performance of solar panels coated with biomimetic BOAR film. (a) Output power–voltage (P–V) characteristics of the biomimetic NCGL-structured NOA61 film and the smooth NOA61 film under AM1.5 standard solar illumination. (b) EQE spectra and integrated current curves of the biomimetic NOA61 film and the smooth NOA61 film. I–V characteristics of silicon solar cells coated with BOAR films and unstructured films under an illumination angle of (c) normal incidence and (d) 25°. P–V curves of silicon solar cells with BOAR films and unstructured films at an illumination angle of (e) normal incidence and (f) 25°.

4. Conclusion

In summary, the Orthetrum triangulare hind wings demonstrated exceptional AR properties, exhibiting merely 0.7% average reflectivity with less than 10% optical degradation across full solar spectrum (350–850 nm) and wide incidence angles (0°–75°). Theoretical analysis revealed that the parabolic conical configuration derived from the NCGL hybrid structures on wing-inspired surfaces substantially enhances the substrate's photon manipulation performance through a graded refractive index profile generation at the air–substrate interface. FDTD analysis further verified that synergistic coupling between nanocone and grid structural elements enhances the BOAR capability by approximately 1% compared to isolated nanocone arrays. By means of bio-template-assisted replica molding, this biological structure design was successfully replicated in NOA61 polymer films. At normal incidence, the biomimetic NCGL structured film attains a visible-light transmittance of around 96.3%. Notably, its transmittance remains as high as ∼78% even under the extreme condition of 75° incidence. FDTD simulations excluding wing venation structures revealed peak transmittance of 98.6% across the natural light spectrum, with approximately 86.3% retention at 75° incidence angle, demonstrating significantly enhanced performance over conventional designs. Electric field analysis confirmed the grid component's critical role in enhancing photon capture and redirection into conical structures, highlighting promising applications in photothermal systems and optoelectronic devices. After undergoing a series of environmental and mechanical tests—including UV irradiation, high-temperature resistance, humidity tolerance, sand impact, and pencil hardness evaluation—the biomimetic NOA61 film exhibits excellent environmental stability and mechanical properties, demonstrating considerable potential for practical applications. A comprehensive performance evaluation showed that integrating the BOAR film yielded a 36.6% relative improvement in the PCE of solar panels. It also enhanced the EQE and current density by an average of 2.9% and 2.6%, respectively. Moreover, outdoor multi-angle tests verified that the NCGL structure effectively boosts the practical photoelectric conversion efficiency of solar cells. The combination of exceptional wide-angle optical performance, scalable manufacturing via bio-templating, and significant efficiency gains positions this biomimetic architecture as a highly viable solution for next-generation, large-area AR coatings targeting solar energy harvesting and optical sensor applications.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Characterization of the chemical composition of biomaterials; theoretical derivation and computation of optical models for biological structures; structural characterization of biomimetic thin films along with comprehensive performance evaluations, including its optical properties, ultraviolet irradiation resistance, high-temperature resistance, humidity tolerance, sand erosion resistance, and pencil scratch resistance; as well as analysis of the photoelectric conversion performance of the biomimetic thin films under varying angles of incidence. See DOI: https://doi.org/10.1039/d5nr03946k.

Acknowledgements

This work was supported by the National Outstanding Youth Science Fund Project of National Natural Science Foundation of China (No. 52222509), the Foundation for Innovative Research Groups of the National Natural Science Foundation of China (No. 52021003), National Natural Science Foundation of China (No. 52475300, No. 52305303), the Natural Science Foundation of Shandong Province (ZR2024ME185), and the Joint Fund of the Natural Science Foundation of Liaoning Province (No. 2023-BSBA-262).

References

1. L. Zhang, J. Xu, Z. Hu, P. Wang, J. Shang, J. Zhou and L. Ren, ACS Appl. Mater. Interfaces, 2024, 16, 38690–38701.
2. J. Wu, J. Tu, S. Yu, H. Wu, Y. Xie, Y. Yang, Z. Xv and Q. Zhang, Colloids Surf., A, 2023, 667, 131424.
3. X. Zou, G. Zhou, Y. Wan, B. Li, B. Jiang, H. Yan and F. Wang, Chem. Eng. J., 2023, 473, 145234.
4. L. Sun, J. Guo, B. Li, G. Yang, X. Zu, X. Xiang and L. Chen, Appl. Surf. Sci., 2025, 684, 161954.
5. B. Hao, Y. Luo, W. Chan, L. Cai, S. Lyu and Z. Luo, Chem. Eng. J., 2024, 496, 153750.
6. Z. Zhang, H. Zhang, H. Zhang, Y. Shao and J. Zhu, Chem. Eng. J., 2024, 490, 151434.
7. Y. Zhan, C. Li, Z. Che, H. C. Shum, X. Hu and H. Li, Energy Environ. Sci., 2023, 16, 4135–4163.
8. X. Wang, W. Wang, J. Liu, J. Qi, Y. He, Y. Wang, W. Hu, Y. Cheng, K. Chen, Y. Hu, A. Mei and H. Han, Adv. Funct. Mater., 2022, 32, 2203872.
9. J. Zhang, X. Li, M. Zhong, Z. Zhang, M. Jia, J. Li, X. Gao, L. Chen, Q. Li, W. Zhang and D. Xu, Small, 2022, 18, 2201716.
10. S. M. Almenabawy, R. Prinja and N. P. Kherani, Sol. Energy, 2023, 256, 88–95.
11. S. Liu, Y. Qian, Y. Lin, L. Sun, Y. Zhu and D. Li, Sol. Energy Mater. Sol. Cells, 2024, 266, 112679.
12. R. Xu, T. Morimoto and J. Takahara, Nanophotonics, 2023, 12, 2461–2469.
13. H. Lee, S. Chae, A. Yi, V. Devaraj, J. W. Oh, I. Cho and H. Kim, Chem. Eng. J., 2024, 502, 157155.
14. Y. Feng, Z. Huang, X. Zhang and T. Qiu, Sol. Energy, 2024, 284, 113088.
15. K. Yoshida, L. Takahashi, A. Takashima, Y. Fujii and I. Nishio, Langmuir, 2020, 36, 4207–4213.
16. Z. Wang, H. Ding, D. Liu, C. Xu, B. Li, S. Niu, J. Li, L. Liu, J. Zhao, J. Zhang, Z. Mu, Z. Han and L. Ren, ACS Appl. Mater. Interfaces, 2021, 13, 23103–23112.
17. W. K. Kuo, J. J. Hsu, C. K. Nien and H. H. Yu, ACS Appl. Mater. Interfaces, 2016, 8, 32021–32030.
18. A. Jacobo-Martin, J. J. Hernández, E. Solano, M. A. Monclús, J. C. Martínez, D. F. Fernandes, P. Pedraz, J. M. Molina-Aldareguia, T. Kubart and I. Rodríguez, Appl. Surf. Sci., 2022, 585, 152653.
19. B. Lee and M. Y. Jeong, Appl. Surf. Sci., 2025, 708, 163801.
20. D. Chowdhury, S. Mondal, M. Secchi, M. C. Giordano, L. Vanzetti, M. Barozzi, M. Bersani, D. Giubertoni and F. B. D. Mongeot, Nanotechnology, 2022, 33, 305304.
21. M. C. Giordano, F. B. D. Mongeot, D. Chowdhury, S. A. Mohamed, G. Manzato, B. Siri, R. Chittofrati, M. Hussein, M. F. O. Hameed, S. S. A. Obayya, P. Stadler, M. C. Scharber and G. D. Valle, ACS Appl. Nano Mater., 2023, 6, 6230–6240.
22. Z. Han, Z. Wang, B. Li, X. Feng, Z. Jiao, J. Zhang, J. Zhao, S. Niu and L. Ren, ACS Appl. Mater. Interfaces, 2019, 11, 17019–17027.
23. Z. Wang, B. Li, X. Feng, Z. Jiao, J. Zhang, S. Niu, Z. Han and L. Ren, J. Bionic Eng., 2020, 17, 34–44.
24. F. Liu, Y. Sun, Z. Wang, B. Li, S. Niu, J. Zhang, Z. Han and L. Ren, ACS Appl. Mater. Interfaces, 2024, 16, 63049–63058.
25. L. Wang, Z. Li, S. Shen and T. S. Wong, Proc. Natl. Acad. Sci. USA, 2024, 121, e2312700121.
26. M. S. Shih, H. Y. Chen, P. C. Li and H. Yang, Appl. Surf. Sci., 2020, 532, 147397.
27. P. C. Li, H. Y. Chen, K. T. Chiang and H. Yang, J. Colloid Interface Sci., 2021, 599, 119–129.
28. H. Ding, B. Li, Z. Wang, S. Niu, Z. Han and L. Ren, Matter, 2022, 5, 2990–3008.
29. A. L. Davis, H. F. Nijhout and S. Johnsen, Nat. Commun., 2020, 11, 1294.
30. A. F. Pomerantz, R. H. Siddique, E. I. Cash, Y. Kishi, C. Pinna, K. Hammar, D. Gomez, M. Elias and N. H. Patel, J. Exp. Biol., 2021, 224, jeb237917.
31. H. Ding, D. Liu, B. Li, W. Ze, S. Niu, C. Xu, Z. Han and L. Ren, ACS Appl. Mater. Interfaces, 2021, 13, 19450–19459.
32. J. Hasan, A. Roy, K. Chatterjee and P. K. D. V. Yarlagadda, ACS Biomater. Sci. Eng., 2019, 5, 3139–3160.
33. C. Wang, X. Wang, J. Yu and B. Ding, ACS Nano, 2023, 17, 10888–10897.
34. R. Y. Chen, C. J. Lai, Y. J. Chen, M. X. Wu and H. Yang, J. Colloid Interface Sci., 2022, 610, 246–257.
35. C. Y. Chien, Y. H. Chen, R. Y. Chen and H. Yang, ACS Appl. Nano Mater., 2020, 3, 789–796.
36. Y. Ding, L. Liu, C. Wang, C. Li, N. Lin, S. Niu, Z. Han and J. Duan, ACS Appl. Mater. Interfaces, 2023, 15, 30985–30997.
37. Y. H. Chen, H. Y. Chen, C. J. Lai, J. H. Hsu, K. Y. A. Lin and H. Yang, Langmuir, 2021, 37, 9490–9503.
38. B. D. Du, Z. Xu, L. Wang, Y. Guo, S. Liu, T. Yu, C. Wang, F. Wang and H. Wang, Sol. Energy, 2021, 224, 10–17.
39. M. Ko, H. S. Choi, S. H. Baek and C. H. Cho, Nanoscale Adv., 2022, 4, 1074–1079.
40. X. Li, Y. Li, I. Muzammil and M. Lei, Plasma Processes Polym., 2020, 17, 2000050.
41. H. Y. Tseng, Y. H. Chen, R. Y. Chen and H. Yang, ACS Appl. Mater. Interfaces, 2020, 12, 10883–10892.
42. J. X. Zheng, K. S. Tian, J. Y. Qi, M. R. Guo and X. Q. Liu, Opt. Laser Technol., 2023, 167, 109793.
43. L. Gao, Q. Zhang and M. Gu, Int. J. Extreme Manuf., 2025, 7, 022010.
44. F. Ye, C. He, T. Wu, S. Chen, Z. Su, X. Zhang, X. Cai and G. Liang, Adv. Funct. Mater., 2024, 34, 2402762.
45. Y. Zhao, N. Huo, S. Ye and W. E. Tenhaeff, J. Mater. Chem. C, 2023, 11, 4005–4016.
46. J. Li, G. Chai and X. Wang, Int. J. Extreme Manuf., 2023, 5, 032003.
47. J. Qu, H. Jia, W. Wang, Y. Wang and S. Zhu, Silicon, 2023, 15, 4959–4966.
48. Z. Han, J. Yuan, P. Tian, B. Liu, J. Tan and Q. Zhang, J. Mater. Res., 2023, 38, 3316–3323.
49. A. Papadopoulos, E. Skoulas, A. Mimidis, G. Perrakis, G. Kenanakis, G. D. Tsibidis and E. Stratakis, Adv. Mater., 2019, 31, 1901123.
50. N. Nedelcu, D. Webb, N. Ackroyd, E. Scott and F. C. de Santana, J. Electron. Mater., 2024, 53, 3692–3701.
51. J. X. Zheng, K. S. Tian, J. Y. Qi, M. R. Guo and X. Q. Liu, Opt. Laser Technol., 2023, 167, 109793.
52. D. Bokov, A. Jalil, S. Chupradit, W. Suksatan, M. Ansari, I. H. Shewael, G. H. Valiev and E. Kianfar, Adv. Mater. Sci. Eng., 2021, 2021, 5102014.
53. D. Yoo, S. Garud, C. T. Trinh, D. Amkreutz and C. Becker, IEEE J. Photovolt, 2022, 12, 97–106.
54. S. Mattaparthi and C. S. Sharma, J. Bionic Eng., 2019, 16, 400–409.
55. A. Jacobo-Martín, M. Rueda, J. J. Hernández, I. Navarro-Baena, M. A. Monclús, J. M. Molina-Aldareguia and I. Rodríguez, Sci. Rep., 2021, 11, 2419.
56. Z. Han, Z. Jiao, S. Niu and L. Ren, Prog. Mater. Sci., 2019, 103, 1–68.
57. S. Yoshioka and S. Kinoshita, Phys. Rev. E:Stat., Nonlinear, Soft Matter Phys., 2011, 83, 051917.
58. R. H. Siddique, G. Gomard and H. Hölscher, Nat. Commun., 2015, 6, 6909.
59. S. Chattopadhyay, Y. F. Huang, Y. J. Jen, A. Ganguly, K. H. Chen and L. C. Chen, Mater. Sci. Eng., R, 2010, 69, 1–35.
60. A. Gombert, W. Glaubitt, K. Rose, J. Dreibholz, B. Bläsi, A. Heinzel, D. Sporn, W. Döll and V. Wittwer, Thin Solid Films, 1999, 351, 73–78.
61. D. A. G. Bruggeman, Ann. Phys., 1935, 416, 636–664.

---