Introduction

The external quantum efficiency (EQE) of an organic optoelectronic device is generally EQE = LCE × IQE1,2,3,4,5,6,7,8,9,10. Here, the internal quantum efficiency (IQE) is the probability of the electron–photon conversion in the active regions. The light coupling efficiency (LCE) is the fraction of emitted light that is out-coupled from OLEDs to the air, or incident light that is in-coupled by absorption in the active layer of OSCs. Currently, an IQE of nearly 100% has been realized in OLEDs with phosphorescent materials using both singlet and triplet states4,11,12,13, while efficiency improvement in OSCs is chiefly achieved via optimization of exciton dissociation and charge collection6,10,14. Nevertheless, the LCE for conventional OLEDs is still limited to ~20% due to optical confinement and internal reflection by the high refractive index of organic materials and glass substrate15,16,17,18. Similarly, due to mismatch between optical absorption length and charge transport scale in OSCs, a large fraction of incident light remains wasted for photocurrent generation, limiting EQE and thus power conversion efficiency (PCE)6,9,19,20,21,22,23. Therefore, a substantial improvement in overall device efficiency for organic optoelectronics can be readily achieved via optimal light manipulation in the photo-active regions1,2,3,6,7,24.

Various photonic structures have been implemented in the appropriate interfaces for guiding and retaining light within the active regions of OLEDs and OSCs, such as the use of low-index dielectric grid2,28, high refractive index substrate1, subwavelength photonic crystals25,26,27, microlens arrays29,30,31, optical spacers32,33, optical cavities34,35, wrinkles7, spontaneously formed buckles3, or metallic nanostructures-induced plasmonic effect3,23,27,36,37,38,39,40. Often, those reported methods for light manipulation are hindered for practical adoption due to undesirable spectral distortion or shifting, response at specific light wavelengths or complicated fabrication process.

It is well-known that biological systems in nature possess amazing ability in manipulation of light propagation via unique nanostructured surfaces. Particularly, biomimetic moth's eye nanostructure (MEN) with a two-dimensional sub-wavelength structure can exhibit unique optical properties in suppressing the Fresnel reflections at an interface, yielding high transmittance over a broad spectral range and wide angles of incidence41,42,43,44,45,46. Here we introduce a universal method for efficient light manipulation in OLEDs and OSCs by integrating a dual-side MEN through nanoimprint lithography, as shown in Fig. 1a. The present MEN has the great advantage of facile manufacturing compatibility for large-area and low-cost electronic devices and is especially suitable for light manipulation with broad spectral response and directional randomness. The OLED device constructed with the MENs shows a remarkable improvement in LCE of more than 2 times. The maximum EQE and current efficiency of a MEN-patterned green phosphorescent OLED are increased to 119.7% and 366 cd A−1 at 1,000 cd m−2 with no spectral distortion or any change in device electrical properties. Similarly, the PCE of MEN-patterned OSCs can be enhanced to 9.33% with a 20% increase in short-circuit current density (JSC) without any detrimental changes on open circuit voltages (VOC) and fill factor (FF). We anticipate our findings will stimulate the development of new device structures of much improved efficiency for commercial applications.

Figure 1
figure 1

Device structure and atomic force microscope (AFM) images of moth's eye nanostructure (MEN).

(a), Schematic diagram of OLED and OSC integrated with dual-side MENs. Internal MEN is first patterned on a PEDOT:PSS or ZnO layer on ITO coated glass substrate for OLED or OSC, respectively. The organic active layers and metal electrode are then deposited on the corrugated PEDOT:PSS or ZnO layer. Diagram is not to scale. (b), View of internal MEN transferred on PEDOT:PSS layer on ITO surface for OLEDs, showing a period of ~200 nm, a groove depth of ~80 nm, a diameter of ~100 nm and duty cycle of 0.5. (c), View of internal MEN patterned on ZnO layer on ITO surface for OSCs, showing a period of ~400 nm, a groove depth of 80 nm, a diameter of ~200 nm and duty cycle of 0.5. (d), View of external MEN patterned on UV-resin on glass surface, showing a period of ~200 nm, a groove depth of 180 nm, a diameter of ~120 nm and duty cycle of 0.6.

Results

MEN fabrication

Figure 1a depicts a schematic diagram of the device structure constructed with dual-side MEN, which was transferred by an imprinting technique using perfluoropolyether (PFPE) mold containing the MEN with compressive stress on the surface of poly(3,4-ethylene dioxythiophene):polystyrene sulfonic acid (PEDOT:PSS), ZnO, or UV-curable resin films supported on ITO glass substrates (see the Methods and Supplementary Figs. S1–S5 for the fabrication processes of PFPE molds and devices). Both the sub-wavelength structures and a continuously tapered morphology can be observed on the surface of PFPE molds, which effectively mimic the natural structures of a moth's eyes.

Figure 1b–d present the atomic force micrographs (AFM) of MENs patterned on a PEDOT:PSS layer (Fig. 1b), a ZnO layer (Fig. 1c), as well as UV-curable resin film (Fig. 1d), showing the produced MENs is smaller than visible wavelength. Tapping-mode AFM studies of these layers show that the surface root-mean-square (RMS) roughness of the PEDOT:PSS layer was changed from 1.03 ± 0.01 nm to 5.28 ± 0.05 nm upon the imprinting, while that of the ZnO layer varied from 0.98 ± 0.02 nm to 6.43 ± 0.07 nm. Note that the MEN diameter of imprinted ZnO or PEDOT:PSS layer is smaller than the feature size of the PFPE mold (Supplementary Figs. S2–4), which is attributed to shrinkage of ZnO or PEDOT:PSS volume upon the solvent removal during the imprinting procedure. In addition, the surface morphologies of the deposited organic layers confirm the conformal coating of each functional layer on the patterned PEDOT:PSS or ZnO layer, although deposition of each successive layer causes progressive smoothing of the MEN structuring (Supplementary Fig. S6).

To evaluate the capability of dual-side MEN on light coupling in organic optoelectronics, we constructed OLED and OSC devices using representative materials in a conventional architecture on dual-side MEN-patterned substrates as well as flat substrates (See Methods for materials composition and device structures used for OLEDs and OSCs). The device with dual-side MEN is fabricated as shown in Fig. 1a. The conventional device is on a flat substrate without MEN. For device with internal MEN, a MEN is only transferred to the PEDOT:PSS or ZnO layer on top of ITO surface before organic layer deposition. For device with external MEN, the MEN is only transferred to UV-curable resin on the glass surface. To ensure consistent results, each series of all four samples for OLEDs or OSCs were simultaneously fabricated in the same batch.

OLED fabrication with MEN

The performance characteristics of green phosphoresent OLEDs including a multi-photon emission (MPE) structure of single, double or triple electroluminescence (EL) units47 on four different substrates are compared and plotted in Fig. 2. It is shown that the electrical properties of devices patterned with MEN are almost identical to that on a flat substrate (Fig. 2a), while the luminance is dramatically increased compared to that without MEN (Fig. 2b–d). The EQE (Fig. 2b–d) and current efficiency (Supplementary Fig. S7) for these devices at a luminance of 1,000 cd m−2 and the EQE enhancement ratios relative to that of a conventional device are summarized in Table 1. For OLED combining internal MEN and single EL unit, the EQE is 69% higher than that of the conventional device constructed on flat substrate, which corresponds to an increase in the amount of confined light from the organic layers and ITO electrode into the glass substrate2. The external MEN promotes the out-coupling of light originally trapped in glass due to total internal reflection, as illustrated by the 77% enhancement in EQE. As illustrated in Table 1, the dual-side MEN enables a further improvement of efficiency by a factor of 2.46 times that of the conventional device without MEN, due to enhanced light extraction from organic layers, ITO as well as glass. The resulting OLED combining dual-side MEN with a MPE structure of triple EL unit yields a maximum EQE of 119.7% and current efficiency of 366 cd A−1 at 1,000 cd m−2, which is among the highest values ever reported for OLEDs in the literature47,48,49.

Table 1 Efficiency comparison for OLEDs with different MENs, measured at 1000 cd m−2
Figure 2
figure 2

Performance characteristics of OLEDs.

(a–d), Current density-voltage (J-V) characteristics (a) and EQE versus luminance (cd m−2) (b–d) of representative OLEDs without MEN (filled symbols) and with dual-side MEN (open symbols), which combine a MPE structure with single (b, squares), double (c, circles) and triple (d, triangles) EL units. Filled and open stars in (b–d) represent the devices only using internal and external MEN, respectively. (e), (f), Normalized EL spectra (e) and angular dependence of light intensities (f) for OLEDs without MEN (filled symbols) and with dual-side MEN (open symbols), measured at J = 1 mA cm−2. The dashed line in (f) represents a guide to the ideal Lambertian emission pattern.

To investigate the influence of MEN on light extraction, the normalized EL spectra of devices without and with MEN were measured and shown in Fig. 2e. It is shown that the device with MEN exhibits wavelength-independent enhancement in efficiency, confirming no grating effects with the quasi-periodic sub-wavelength structure. Additionally, unified EL spectra of OLEDs with MEN were obtained, which are independent of the viewing angle, in contrast to the conventional OLED showing a variation of the emission spectra (Supplementary Fig. S8). The angular dependences of light intensity for OLEDs without and with MEN are displayed in Fig. 2f. The emission of the conventional device is close to a Lambertian pattern, whereas the integration of dual-side MEN in OLED causes a stronger side emission, which is mainly due to redirected emission over all azimuthal directions. Similar enhancement was also observed for blue and red OLEDs without spectral distortion (Supplementary Fig. S9), indicating the excellent light manipulation capability of MENs for use in full-color and white OLEDs using only one single patterned structure.

OSC fabrication with MEN

Figure 3a plots the J-V characteristics of OSCs constructed on four different substrates under 100 mW cm−2 air mass 1.5 global (AM 1.5 G) illumination, where model polymer constituents of thieno[3,4-b]thiophene/benzodithiophene (PTB7) and [6,6]-phenyl C71-butyric acid methyl ester (PC71BM) were used to create the donor-acceptor bulk-heterojunction photo-active layer in the OSCs. The detailed parameters deduced from the J-V characteristics are summarized in Table 2. Regardless of the variation of MEN-patterned substrates, these four OPVs yield almost constant VOC of 0.73 V and FF of ~60%, which are in good agreement with the reported values in the literature10. However, the use of internal, external and dual-side MENs causes large and substantial variations in JSC compared with that of a conventional cell on a flat substrate. Specifically, after incorporation of dual-side MEN the JSC is highly enhanced from 17.59 mA cm−2 to 21.15 mA cm−2, corresponding to an enhancement ratio of as high as 20%. Consequently, a maximum PCE value of 9.33% is obtained, which is among the highest values ever reported to date for OSCs with single photo-active layer10. Since the blended PTB7:PC71BM active layers in these four OSCs were prepared during the same batch processing with similar degree of crystallinity, the performance enhancement, especially the increase in JSC, is ascribable to enhanced light harvesting within the active layer of PTB7:PC71BM due to improved light trapping in the active layer after integration of MENs.

Table 2 Photovoltaic parameters for OSCs with different MENs, measured under 100 mW cm−2 AM 1.5 G illumination. The average values obtained from at least 24 cells are given in brackets
Figure 3
figure 3

Photovoltaic performance of OSCs.

(a), J-V characteristics of OSCs constructed with a structure of ITO/ZnO/PTB7:PC71BM/MoO3/Al without MEN (squares) and with internal (diamond), external (triangles) and dual-side (circles) MEN, respectively, under 100 mW cm−2 AM 1.5 G illumination. (b), The EQE spectra of the corresponding devices and enhancement ratio as a function of emission wavelength, obtained by dividing the spectra of the devices with MEN by that without MEN.

To verify the effect of JSC enhancement, the EQE spectra of the corresponding devices are measured and presented in Fig. 3b. The enhancement ratios of EQE spectra relative to that of the conventional cell are also plotted in Fig. 3b to distinguish which part contributed to the enhancement. For the cells only using external MEN, a uniform increase of the spectral response is clearly observed across the wavelength range of 350 nm to 800 nm, which is in agreement with the fact that the MEN can function as an effective anti-reflective layer44. The use of internal and dual-side MENs can cause a further EQE enhancement and the averaged photo-response with dual-side MEN is ~20% enhanced as compared to that on a flat substrate in the wavelength range from 350 nm to 800 nm.

Optical characterization of OLEDs and OSCs with MEN

Figure 4 compares the transmittance of ITO-glass substrates without and with MENs used in OLEDs and OSCs, respectively. The measurements were conducted with the propagation of incident light towards the glass side for OLEDs and towards the ITO side for OSCs, respectively. The broadband optical nature of different MENs patterned on substrates used in OLEDs and OSCs can be observed in Fig. 4. The incorporation of dual-side MEN sustains a high value on transmittance over a wide spectral range of 350–800 nm, which is 7% higher than that of the flat ITO glass. It is also noted that the use of only internal or external MEN leads to an intermediate increase in transmittance compared to that of flat ITO-glass, showing a good match between the transmittance of MEN patterned ITO glass substrate and the experimentally obtained light coupling enhancement in OLEDs and OSCs. The enhancement of optical transmittance over the entire spectrum is attributed to the contributions from the broad distribution of periodicity of the MEN pattern and the gradient refractive index profiles of MENs at the interface41,42,43,44.

Figure 4
figure 4

Optical properties of OLED and OSC with MEN.

(a), (b), Transmittance of the incident light from ITO side to PEDOT:PSS-coated OLED substrates (a) and from glass side to ZnO-coated OSC substrates (b) patterned without MEN (squares) and with internal (diamond), external (triangles) and dual-side (circles) MEN. Inset depicts the schematic of the measurement procedure.

To determine the amount of light scattering present in MENs, specular and diffuse transmission measurements for ITO-glass substrates without and with MENs used in OLEDs and OSCs were performed (Supplementary Fig. S10). The transmittance haze (haze = diffuse transmission/total transmission) shows the percentage of total light transmitted that is scattered. The average haze for ITO glass substrates with dual-side MENs used in OLEDs and OSCs are 4.04% and 2.92%, respectively, over the visible spectrum. For comparison, a PEDOT:PSS- or ZnO-coated ITO glass substrate without MEN has negligible haze of ~0.4%. It indicates that scattering is present and significantly contributes to the total transmission.

Simulation of OLEDs and OSCs with MEN

To gain further insight into the optical manipulation of light in OLEDs and OSCs using dual-side MEN, we perform the optical modeling calculations of light propagation using the finite-difference-time-domain (FDTD) method (see Methods for the detailed procedure of the optical modeling). Compared to the light that initially undergoes in a conventional OLED (Fig. 5a), dual-side MEN effectively increases the out-coupling probability of the emitted light as observed clearly from the intensity field distribution with propagation not only inside the device but also at the glass/air interface (Fig. 5b). This improvement of light extraction in OLEDs is chiefly attributed to the diffraction grating of the emission layer, corrugated metal electrode and the gradual transition in refractive index of the UV-curable resin-patterned MEN on the surface of glass substrate. These factors enable the emitted light towards the surface normal with higher out-coupling efficiency44.

Figure 5
figure 5

Simulation of OLED and OSC with MEN.

(a), (b), Normalized cross-section intensity field distributions with propagation of OLEDs without (a) and with (b) MEN-patterned PEDOT:PSS layer using Finite Difference Time Domain (FDTD) method (Rsoft FullWave). (c), (d), Normalized cross-section magnetic field profiles of TM polarized light at 720 nm for OSC without (c) and with (d) MEN-patterned ZnO layer.

Figure 5c,d display the cross-section field profiles of transverse magnetic (TM) polarized light at 720 nm in OSCs without and with MEN. As shown in Fig. 5c for the device constructed on a flat substrate, the power flux is localized around the PTB7:PC71BM active layer and quickly decreases towards the Al electrode. On the contrary, the modeled results (Fig. 5d) show that the incorporation of internal MEN patterned on ZnO layer can significantly influence the optical field distribution of incident light within the active layer. The simulation results provide theoretical support to the photonic enhancement of light trapping and absorption in the active layer as well as the PCE enhancement, which is induced not only by a geometric effect on light scattering but also other related causes (e.g., surface plasmonic resonance) from nanostructured active layer and corrugated metal electrode (see the detailed theoretical analysis in Supplementary Fig. S11)36.

Discussion

As demonstrated here, integrating a dual-side MEN with broadband quasi-omnidirectional properties provides a universal method of light manipulation for use in the performance improvement of organic optoelectronic devices of various material systems. An out-coupling enhancement by a factor of over 2 times that using a conventional flat glass substrate is realized in OLEDs, resulting in significant improvement of EQE to 119.7% and current efficiency to 366 cd A−1 at a luminance of 1,000 cd m−2. The 20% enhancement of light harvesting yields a highly efficient OSC with an efficiency of 9.33%. It has been found that both the morphology and the dimensions of the nanostructure in biomimetics are critical to produce the desired functionality50. The light coupling enhancement in OLEDs and OSCs with MEN is the combined result of both the two-dimensional sub-wavelength structures and the continuously tapered morphology on the patterned surface with a superior gradient refractive index profile at the interface. The light is therefore manipulated in all azimuthal directions over the entire emission wavelength range. The optical simulations provide an understanding of optical manipulation of light out-coupling and in-coupling process in OLEDs and OSCs. Note also that the method developed here brings about an invaluable advantage, which enables the processing compatibility with the high-throughput large-area roll-to-flat and roll-to-roll manufacturing techniques in future mass production of low-cost organic optoelectronic devices.

Methods

Nanoimprint lithography of mens and characterization

The MENs were fabricated by nanoimprint lithography. PFPE molds containing different MEN were used to perform the imprinting on both sides of ITO-patterned glass substrate with a sheet resistance of 20 Ω per square. For the fabrication of OLED devices, the ITO-patterned glass substrate was spin-coated by the aqueous PEDOT:PSS solution. For the case of OSCs, the sol-gel-derived ZnO layer was formed on ITO glass substrate through spin-coating of the ZnO precursor solution, which was prepared by dissolving 0.5 M zinc acetate dehydrate (Zn(CH3COO)2·2H2O, Alfa Aeser, 99.5%, 220 mg) in the solvent consisting of ethanolamine (NH2CH2CH2OH, Amethyst Chenmicals, 99.5%, 0.061 mL) and 2-methoxyethanol (CH3OCH2CH2OH, Alfa Aeser, 99.5%, 2 mL) under vigorous stirring for 12 h for the hydrolysis reaction at ambient conditions. Prior to the molding, the originally spin-coated PEDOT:PSS and ZnO layers were annealed in ambient air at 140°C for 10 min and 150°C for 5 min, respectively. Then, the PFPE mold was put in comformal contact with a 100 nm-thick PEDOT:PSS or ZnO layer and pressed under a constant pressure of 1.5 bar for 3 min at 140°C. For ensuring the complete formation of a uniform nanohoneycomb morphology on the PEDOT:PSS or ZnO layer, the substrates were annealed for further 9 min at the same temperature after the mold was peeled off. The organic layer and metal electrode were subsequently deposited on the substrate through thermal evaporation or spin-coating to complete the device. Finally, the second UV-assisted imprinting of MEN on the glass surface, which was coated with UV-curable resin, was repeated with a negative PFPE mold after encapsulating the devices with a cap glass to prevent the damage to the formed nanostructures under pressure. The UV resin drop-casted on the glass surface was embossed with a negative PFPE mold under a constant pressure of 1.5 bar for 10 s with a UV illumination at light power intensity of 500 mJ cm−2 at a wavelength of 395 nm. The imprinting recipe was optimized by a systematic study of the imprinting time and pressure.

Optical transmittance measurements were carried out using an UV/vis/near-IR spectrometer (Perkin Elmer Lambda 750). Surface morphology of MENs was characterized with AFM (Veeco Multimode V) in the tapping mode. The refractive index of the MEN-patterned films was determined by the alpha-SE™ Spectroscopic Ellipsometer with the angle of incidence at 70°.

OLED fabrication and measurement

For the OLED fabrication, organic layers and metal cathode were thermally deposited in high vacuum chamber with a base pressure of 2 × 10−6 Torr, in which the deposition rate and film thickness were monitored by a quartz crystal oscillator. The OLEDs were composed of a MPE structure of one to three consecutive green phosphorescent EL units, which were electrically connected in series by a bilayer intermediate connector of a 10 nm-thick Mg-doped 4,7-diphenyl-1,10-phenanthroline (Mg:BPhen) and a 5 nm-thick MoO3. The EL unit consisted of a 40 nm-thick hole-transporting layer of 4,4-bis[N-(1-naphthyl)-N-phenyl-amino]biphenyl (NPB), a 20 nm-thick green emitting layer of 4′-bis(carbazol-9-yl)biphenyl (CBP) doped with bis(2-phenylpyridine)(acetylacetonate)iridium(III) [Ir(ppy)2(acac), 8 wt%] and a 40 nm-thick electron-transporting/exciton blocking layer of 2,2′,2″-(1,3,5-benzinetriyl)-Tris(1-phenyl-1-H-benzimidazole) (TPBi). This was followed by 0.5 nm-thick LiF and 100 nm-thick Al layers as a bilayer cathode, which was deposited through a shadow mask.

The current density-luminance-voltage (J-L-V) characteristics and EL spectra of OLEDs were measured simultaneously in the forward direction using a programmable source meter (Keithley 2400) and a luminance meter/spectrometer (PhotoResearch PR655) in air ambience. The angular emission patterns were measured by integrating the EL spectra of the devices according to emission angles.

OSC fabrication and measurement

The OSCs were constructed with a structure of ITO/ZnO (100 nm)/PTB7:PC71BM (1:1.5 by weight, 100 nm)/MoO3 (10 nm)/Al (100 nm). PTB7 and PC71BM were purchased from 1-material Chemscitech and Lumitec, respectively and used as received. After the preparation of ZnO layer without or with MEN, the substrates were transferred into nitrogen-filled glove box for the active layer coating. The PTB7:PC71BM active blend layer was prepared by spin-coating the solution of PTB7:PC71BM in chlorobenzene/1,8-diiodoctane (97:3% by volume) mixed solvent (concentration, 25 mg mL−1) at 1000 rpm for 2 min on top of the ZnO layer and dried in vacuum, forming a nominal thickness of ~100 nm. To complete device fabrication, a 10 nm MoO3 layer and a 100 nm Al layer were subsequently evaporated to form a top anode. A shadow mask was used to define the Al anode, determining the device area of 0.1 cm2.

The photovoltaic characteristics of OSCs were examined in air at room temperature using a programmable Keithley 2612 source measurement unit under a 1 sun, AM 1.5 G spectrum from a solar simulator (Oriel model 91160; 100 mW m−2). The EQE measurement was conducted with a photo-modulation spectroscopic setup (Newport monochromator). The illumination intensity of solar simulator was calibrated using a standard Si reference cell with known spectral response.

Optical modeling and simulation

The optical properties for OLEDs and OSCs were theoretically determined by 3 dimensional modeling with commercial RSoft 8.1 (RSoft Design Group, Inc), in which hexagonal closely packed nanostructures with linearly tapered profile were used for simplicity instead of randomly distributed geometry. For simulation of light out-coupling in OLEDs, the FDTD method (RSoft Fullwave) was performed to investigate the light transmission characteristics. For characterizing light absorption in OSCs, the rigorous coupled wave analysis (RCWA) (RSoft DiffractMOD) was employed to solve Maxwell's equations for electromagnetic modes present in each layer after introduction of MEN. A plane-wave source was directed along the z-axis to calculate their transmittance properties. The extinction coefficients and complex refractive indices of n and k of all layers used in the simulation were experimentally obtained by spectroscopic ellipsometer measurement.