From cinema screens and televisions, to smartphones and virtual reality headsets, displays have progressively moved closer to the human eye, featuring smaller sizes and higher resolutions. As display technology advances, a fundamental question arises regarding the ultimate limits of display size and resolution. To provide an immersive and high-fidelity visual experience, the display is designed with a size comparable to the human pupil and an ultra-high pixel density. It establishes a conceptual benchmark inspired by the resolving limits of the retina (Fig. 1a) and defines a practical boundary for display technologies, which we term the retina display.
a, Conceptual illustration of an ultimate virtual reality display. The display is sized to approximate the human pupil and features an ultra-high pixel density serving as a conceptual benchmark inspired by the retina, supporting ultra-fine visual detail. b, Structural diagram of metapixels (subpixels). The metapixels consist of WO3 nanodisks and a reflective layer on a glass substrate. By varying D and W of the nanodisks, the metapixels can selectively reflect RGB colours. Further tuning of T enables the generation of hybrid colours such as CMY. As WO3 is electrochromic it can undergo reversible electrochemical reactions, yielding reflectance modulation of the WO3 nanodisks, enabling an RGB video display.
Assuming an effective display aperture of 8 mm, corresponding to the maximum human pupil diameter under scotopic conditions, and a field of view of 120°, consistent with the functional limits of human vision, achieving the maximum angular resolving capacity of approximately 60 pixels per degree would require a display pixel density of around 23,000 pixels per inch (PPI). This value represents a conceptual benchmark rather than a practical specification, because under typical photopic conditions with a smaller pupil size (~4 mm) the required pixel density would be even higher. In practice, however, as the display is not intended to be positioned directly in the pupil plane (because of safety considerations and to satisfy optical invariants such as étendue), the resolution requirement can be relaxed, because increasing the eye–screen distance effectively enlarges the screen size.
Unfortunately, as pixel sizes continue to shrink in mainstream emissive displays, diminished emitter dimensions lead to reduced brightness, compromised uniformity, increased colour cross-talk and greater fabrication complexity, posing considerable challenges for ultra-high-resolution imaging1,2,3. Currently, commercially available smartphone display pixels are typically around 60 × 60 μm2 (~450 PPI), which is approximately 2,500 times larger than the theoretical size required for the ultimate retina display. Already at this scale, the emitted light becomes difficult for the naked eye to perceive, particularly in bright outdoor environments. Moreover, the smallest published colourful micro-light-emitting-diode (micro-LED) display currently available achieves only a pixel size of 4 × 4 μm2 (excluding the distance between pixels)4, making it challenging to replicate retinal-level resolution across vast fields of view. Furthermore, at such small scales, colour cross-talk and uniformity remain considerable technical hurdles. These limitations expose the large challenges of using conventional emissive display technology to realize the ultimate virtual reality display.
Reflective displays, which rely on ambient light for visibility, do not have luminosity issues, and their optical contrast remains unaffected by pixel size reduction given that reflection is governed by the polarization of materials at the nanoscale. However, existing reflective display (electronic paper or E-paper) technologies are hampered by marked limitations. Reflective liquid crystal displays, for instance, are constrained by the thickness of the liquid crystal layer, whereas electrophoretic displays (such as the Kindle) are restricted by the size of their capsules5,6. So far no commercially available reflective display technology has achieved high resolutions of above 1,000 PPI.
Optical metasurfaces have demonstrated the capability to achieve ultra-high pixel densities of more than 10,000 PPI (~2.5 μm pixel size), with patterned nanomaterials capable of printing images at resolutions of up to around 100,000 dots per inch, approaching the optical diffraction limit7,8,9. However, most modern nanoprinting methods rely on static materials, such as metals or high-refractive-index dielectrics10,11,12,13. When applied to dynamic display systems (for example, meta-organic LED), these materials require modulation through micro-light sources, which are still affected by the inherent limitation of electromagnetic reduced intensity as resolution increases3. Furthermore, they are often affected by lateral light leakage between adjacent colour subpixels, limiting their ability to produce ultra-high-resolution images. In fact, static reflective displays are also influenced by interactions between neighbouring pixels, altering their optical properties at ultra-high pixel densities, making it challenging to use conventional red–green–blue (RGB) subpixel configurations for image display13,14,15.
In recent years, there has been growing interest in integrating dynamic and static materials to explore tunable nanophotonics systems. Particularly in the field of displays, hybrid nanomaterials—combining tunable conjugated polymers or semiconductors as colour modulators with metallic nanostructures—have demonstrated the ability to modulate the intensity or the reflected colours of subpixels16,17. These technologies significantly enhance the colour gamut, reflectivity and optical contrast of E-paper and enable video display functionality18,19,20,21. However, owing to limitations in structure, materials and fabrication methods, the pixel sizes of these hybrid nanomaterials remain in the range of tens to hundreds of micrometres, making it challenging to achieve ultra-high-resolution displays22,23,24.
Here we propose a conceptually new E-paper technology, termed retina E-paper, capable of achieving ultra-high resolutions exceeding 25,000 PPI (~560 nm), surpassing the theoretical human visual limit of 60 pixels per degree across a 120° field of view on an 8 mm screen. The retina E-paper comprises electrochromic WO3 metapixels, which undergo an insulator-to-metal transition during electrochemical insertion of, for example, alkali ions, allowing electrically dynamic control over optical properties such as refractive index and absorption. This transition enables tunable reflectance and contrast, which is critical for optimizing display performance. The basic colour generation principle builds on the hierarchical structuring of building blocks, Mie scattering and interference between the building blocks. This technology enables the practical construction of ultra-high-resolution displays (for example, exceeding 100 megapixels) within a compact area, paving the way for the ultimate virtual reality display.
The retina E-paper comprises electrochromic WO3 metapixels integrated with a highly reflective substrate (Pt/Al). Its normalized high reflectance (~80%) and optical contrast (~50%) remain unaffected by pixel size reduction, maintaining exceptional visibility even at pixel sizes as small as approximately 400 nm. To minimize interference between adjacent pixels, we carefully optimized the dimensions and spacing of the primary colour metapixels, enabling full-colour displays by precisely mixing RGB subpixels. Furthermore, the substrate (Al/Pt) exhibits excellent conductivity. By reducing the lateral distance between the working and counter electrodes to 500 nm and using short-pulse input signals, we achieved greater than 95% optical contrast modulation of the WO3 nano-pixels within 40 ms, supporting a video display of more than 25 Hz. This refresh rate is more than ten times faster than those of the previously reported fastest WO3-based electrochromic devices25. The short distance between the working and counter electrodes enhances the external electric field driving ion doping, and this nanodisk design significantly reduces the required amount and increases the reaction surface area of WO3 compared with the planar surface. Furthermore, unlike emissive displays that require constant power, retina E-paper features colour memory, consuming energy mainly during pixel switching. Its power usage is around 1.7 mW cm–2 for video and around 0.5 mW cm–2 for static images, significantly less than that of other types of E-reader26.
Finally, to demonstrate the potential of retina E-paper for virtual reality applications, we use cyan, magenta and yellow (CMY) colour metapixels to reconstruct an anaglyph 3D butterfly image. Furthermore, to showcase its full-colour display performance, we reproduced a high-resolution image inspired by the iconic painting The Kiss by Gustav Klimt and dynamically modulated the colours using electrical control. The retina E-paper features a compact surface area of around 1.9 × 1.4 mm2 (about 1/4,000th the size of a standard smartphone display) while achieving an impressive resolution of 4,300 × 700 pixels.
Figure 1b illustrates the fundamental structure of the retina E-paper, composed of electrochromic WO3 metamaterials integrated with a highly reflective (Al/Pt) substrate. In the bright state, WO3 behaves as a dielectric material with a high refractive index ranging from around 2 to 2.4 in the visible spectrum27, enabling the generation of high-resolution colours even at sizes smaller than the incident wavelength11. By precisely tuning the diameter (D) and spacing (W) of the nanodisks, the scattering modes can be adjusted to reflect the primary colours, RGB, forming the subpixels of one pixel of the retina E-paper. As the subpixels are made of meta-material, they are also called metapixels; however, optical interactions between these nanodisk-based subpixels can also affect colour mixing. Further tuning of the subpixel spacing (T) is required to ensure proper additive colour blending for display applications. After patterning the RGB subpixels, the next step is intensity modulation. As WO3 is an electrochromic material, meta-atom absorption can be dynamically modulated by electrochemical reactions under applied voltage, altering the reflectivity of the subpixels. This capability enables the retina E-paper to achieve dynamic colour modulation for display applications. The nanofabrication process is shown in Extended Data Fig. 1.
Figure 2a demonstrates how the reflective colours of WO3 metapixels vary at a fixed thickness of 110 nm while varying the nanodisk diameter from 220 nm to 320 nm and the spacing from 100 nm to 200 nm. The thickness was chosen to balance Mie-resonance-based colour purity and electrochemical switching speed. Thinner layers may accelerate switching but reduce optical confinement and weaken Mie scattering, thus degrading colour saturation. This geometry range enables the metapixels to cover the entire visible spectrum. However, it is essential to note that not all RGB pixels are suitable for subpixels in retina E-paper. The reflected colour of each subpixel in an ultra-high-resolution E-paper system is influenced not only by its geometry but also by interactions with adjacent pixels (Fig. 2c). Therefore, selecting appropriate RGB pixels and ensuring that their hybrid reflected colours adhere to the principles of additive colour mixing is a critical step for achieving full-colour displays. In Fig. 2a (right), the spectra and corresponding geometries of the selected RGB pixels are presented: R (D = 220 nm, W = 200 nm), G (D = 260 nm, W = 200 nm) and B (D = 260 nm, W = 140 nm). The spectra are normalized to the reflective layer to highlight the WO3 nanodisk colour-tuning by structural changes. Unlike emissive displays, in which visibility diminishes with pixel size reduction, E-paper technology maintains consistent brightness and reflectivity even at ultra-high resolutions. As illustrated in Fig. 2b, the red pixel retains its colour and reflectance in both bright- and dark-field microscope images, even as the size is reduced from 20 μm to 420 nm. A minimum of four nanodisks per pixel is required to preserve the Mie scattering and grating modes of the nanodisks, resulting in minimum pixel sizes of 420 nm for red, 460 nm for green and 400 nm for blue.
a, Metapixel design. Left: tuning D and W of WO3 nanodisks achieves a diverse colour palette. The dashed box highlights selected RGB pixels along with their intermediate regions, which contain CMY pixels. Middle: reflectance spectra of the selected RGB pixels. Right: corresponding D and W values for the chosen RGB pixels. b, Microscopic and structural characterization. Left: bright-field (top) and dark-field (bottom) microscope images of a red pixel with feature sizes of 20 μm, 2 μm and 420 nm, captured under ×100 magnification. Scale bars, 10 μm. Right: SEM images of 2 μm and 420 nm red pixels. Scale bars, 2 μm (top) and 200 nm (bottom). c, Colour mixing by subpixel arrangement. Left and middle: reflective colour varies as a function of T between adjacent RGB subpixels. Right: reflectance spectra of hybrid CMY pixels, corresponding to optimized subpixel spacing. d, High-resolution colour imaging. Top: bright-field microscope (×100) images of hybrid CMY pixels. Scale bars, 1 μm. Bottom: SEM images of the corresponding hybrid pixels. Scale bars, 500 nm.
Once the smallest dimensions for the three primary colour metapixels are determined, the next step is merging them to achieve a full-colour display. In Fig. 2c, the grating modes between adjacent subpixels are influenced by the spacing (T) between subpixels, which changes reflected hybrid colours. According to the principles of additive colour mixing, the overlap between RGB should produce CMY, respectively. As shown in Fig. 2a, the intermediate region (black dashed box) between RGB pixels also contains CMY pixels. As long as the spacing between RGB subpixels is carefully designed, the grating modes of adjacent pixels can produce CMY colours, ensuring compliance with the additive colour principle. For comparison, Extended Data Fig. 2 presents several arbitrarily selected combinations of RGB pixels that fail to reproduce the CMY colours. After carefully selecting RGB pixels and tuning the inter-pixel spacing— 380 nm for red-green, 80 nm for blue–red and 100 nm for green–blue—the desired hybrid colours were successfully generated. The corresponding reflection spectra demonstrate that the reflectance of the hybrid-colour pixels matches that of the single-colour pixels.
Finally, Fig. 2d presents microscope and scanning electron microscope (SEM) images of the merged pixels producing CMY colours. Under high magnification (×100), the arrangement of alternating subpixels along the x axis to form hybrid colours is clearly visible. Notably, the Mie scattering mode of individual nanodisks is mainly determined by their size, whereas the grating mode of the subpixel arrays in the x direction governs the generation of reflective mixed colours.
As an electrochromic material, WO3 exhibits an electrically tunable refractive index (n) and extinction coefficient (k) across the visible spectrum (400–700 nm). In the insulator (colour) state, the refractive index varies from around 2.38 to 2.14, with k-values of <0.01. In the metal (black) state, n decreases from around 2.25 to 1.95, whereas k increases significantly (k > 0.4; Extended Data Fig. 3). This transition occurs on electrochemical reduction of WO3, when, for example, alkali ions or a proton, M+, are introduced into the material, altering its electronic structure. The reaction can be expressed as:
$${{\rm{WO}}}_{3}+x{{\rm{M}}}^{+}+x{{\rm{e}}}^{-}\to {{\rm{M}}}_{x}{{\rm{WO}}}_{3}$$
where M+ represents the doping ion (for example, Li+), x is the number of ions inserted into the WO3 structure (0 ≤ x ≤ 1) and e− represents an electron. When WO3 is reduced, electrons are introduced into the conduction band comprising the W 5d bands. The injected electrons lead to the formation of localized polarons, which result in a transition from insulating (semiconducting) behaviour to increased conductivity or metallic-like behaviour as x increases and the polarons overlap. As a result, the refractive index decreases and the extinction coefficient increases, contributing to the optical contrast between the insulating (clear) and metallic (dark) states. Furthermore, the nanodisk structure of the WO3 metapixels, with its high refractive index (n), enables the concentration of incident light in the nanodisk. This effect further amplifies the light–matter interaction and absorption efficiency for stronger optical contrast.
Figure 3a presents the experimental set-up to electrochemically control the colour states of WO3 metapixels. The electrolyte consists of 1 M LiClO4 in acetonitrile, and the RGB pixel size is 350 μm. Metallic electrodes (Pt/Al) were used to minimize potential drops. Notably, we designed a lateral electrode configuration with a narrow 500 nm gap between the working and counter electrodes, enhancing the local electric field and significantly improving the switching speed28,29.
a, Experimental set-up for electrochemical characterization. Left: cross-sectional schematic diagram illustrating the set-up used to characterize the optical properties of WO3 metapixels under electrochemical control. Right: microscope images of the characterized sample show a 500 nm gap between the working and counter electrodes, which enhances the local electric field for improved switching performance. Scale bar, 100 μm. b, Optical response of electrochemically controlled RGB metapixels. Left: reflectance spectra of the selected RGB metapixels in their electrochemically modulated on (sold lines) and off (dashed lines) states, demonstrating dynamic tunability. Right: numerical simulation of the reflectance modulation of red metapixels at a 650 nm incident wavelength, validating that most of the electric field is concentrated in the WO3 nanodisks. c, Switching speed characterization. By applying a short pulse voltage signal, 95% normalized optical contrast is achieved within 40 ms, demonstrating video-rate display applications. d, Electrical properties corresponding to optical modulation. Left: applied voltage pulses of ±4 V with a 40 ms duration produce corresponding current responses, with 95% of the current change occurring within 30 ms, indicating fast ion transport in the WO3 nanodisk structures. Right: relationship (turning off) between optical contrast change, response time and charge consumption. Owing to the colour memory effect of retina E-paper, video playback typically requires only minor updates, with intensity variations typically around 20% (equivalent to about 30 greyscale levels), thereby reducing the energy and time needed for dynamic content display.
Figure 3b illustrates the measured normalized reflectance modulation of RGB metapixels in the on and off states. As both Mie scattering and grating modes are influenced by variations in the refractive index of the surrounding environment, the optimized nanodisk dimensions for the RGB metapixels are: R (D = 300 nm, W = 140 nm), G (D = 260 nm, W = 100 nm) and B (D = 260 nm, W = 40 nm). A clear distinction is observed when comparing the metapixel reflectance in air with that in electrolytes. Specifically, the reflectance in electrolytes is notably higher, whereas scattering in the red-light region is significantly suppressed. This effect arises because Mie scattering depends on the relative refractive index \(m=\frac{{n}_{{\rm{particle}}}}{{n}_{{\rm{env}}}}\), where nparticle and nenv are the refractive index of the WO3 nanodisks and the surrounding environment, respectively. When nparticle is close to nenv, the scattering is markedly reduced, making the material seem more transparent and desaturated in colour. Therefore, the colour contrast of WO3 metapixels in electrolytes is less saturated than in air; however, despite the ultra-thin WO3 nanodisks (~100 nm), the optical contrast remains at ~50%, significantly outperforming most planar WO3 electrochromic devices with the same thickness30. This enhancement is due to the high refractive index of WO3, concentrating the electric field of incident light in the nanodisks. After switching to the dark state, most of the incident light is absorbed, as confirmed by numerical simulations of red subpixel on and off states (Fig. 3b, right). The on–off switching of RGB pixels enables high-contrast modulation, which is essential for reflective display applications with full colour coverage.
Benefiting from the strong local electric field between the closely spaced working and counter electrodes, as well as the ultra-thin amorphous WO3 nanodisks, yielding fast ion insertion31, the electrochemically tunable metapixels achieve an exceptionally fast switching time of only 40 ms to reach more than 95% of the total optical contrast modulation. Supplementary Video 1 demonstrates the rapid on–off switching of the red metapixel, confirming that metapixels support video display. Figure 3c shows ten cycles of reflectance variation in ±4 V pulse input signals. The normalized optical contrast is calculated by \(\frac{R-{R}_{\min }}{{R}_{\max }-{R}_{\min }}\), where R, Rmin and Rmax are the real-time, minimum and maximum reflectance of the WO3 metapixels, respectively; this normalization clearly illustrates the change in optical contrast. The measured optical contrast is also shown in Extended Data Fig. 4. Notably, the operating voltage is comparable to the solid-state two-electrode WO3 electrochromic systems32. However, due to the significantly enhanced switching speed (>65-times faster), precise pulsed voltage control—rather than a constant bias—can be used to minimize energy consumption and mitigate side effects.
Retina E-paper offers ultra-low energy consumption due to its inherent colour memory effect. Energy is primarily consumed when the pixel intensity is actively changed. As shown in Extended Data Fig. 5, pixels in the on state can retain more than 90% of their reflectance for more than 150 s without any power input. This indicates that for relatively static pixels, a brief, low-energy signal applied every several tens of seconds is sufficient to maintain the displayed content. Figure 3d presents the electrical characteristics associated with colour modulation in retina E-paper. Figure 3d (left) shows the current response to a ±4 V voltage pulse, indicating that alkali ions can rapidly intercalate into the WO3 nanodisk layer at video rates, thereby modulating pixel colour states. Figure 3d (right) illustrates the relationship between optical contrast change, response time and charge consumption. Importantly, during video playback, only about 10% of pixels typically undergo state changes20, and among those, the average greyscale shift is around 30 levels33. This means that, on average, only approximately 20% normalized optical contrast is required for video display. Under such conditions, retina E-paper achieves an average switching time of around 5 ms (200 fps) and an energy consumption of about 1.7 mW cm–2. For static images, the energy use drops further to around 0.5 mW cm–2. This energy consumption is significantly lower than that of conventional electrophoretic displays and electrowetting displays26.
To further validate the display performance of retina E-paper, we fabricated metapixels to reproduce an anaglyph 3D butterfly and a high-resolution image inspired by The Kiss. The anaglyph butterfly demonstrates the feasibility of stereoscopic image rendering for virtual reality applications, whereas the reproduction of The Kiss—featuring intricate geometries and a wide colour gamut—highlights the suitability of retina E-paper for ultra-high-resolution, full-colour image display. As our substrate is a highly reflective material analogous to a white canvas, we used the CMY subpixels and used subtractive colour mixing to render the image. Importantly, the patterned image only demonstrates the display capability of WO3 metapixels. For the display application, a thin-film transistor (TFT) array should be used to independently control the reflectance of each pixel, whereas the background should be set to black. The image rendering should follow the additive colour principle using RGB subpixels (Fig. 2).
Figure 4a (left) illustrates the nanodisk diameters and periodicities for the CMY metapixels alongside their corresponding reflection spectra, which show similar reflectance as RGB pixels. Figure 4a (right) presents the merged hybrid colour metapixels, in which the spacing between subpixels are B (T1 = 100 nm), R (T2 = 60 nm) and G (T3 = 60 nm). Notably, except for similar reflectance to RGB metapixels, the CMY system also ensures that the intermediate regions contain RGB pixels (Extended Data Fig. 6). The optimized nanodisk dimensions for CMY pixels are C (D = 260 nm, W = 160 nm), M (D = 240 nm, W = 100 nm) and Y (D = 180 nm, W = 180 nm).
a, Optical properties of CMY metapixels. Left: the reflectance spectra of the selected CMY metapixels demonstrate the spectral response. Right: photographs of the RGB pixels with optimized adjacent subpixel spacing for improved hybrid colour and display fidelity. b, Anaglyph 3D demonstration on retina E-paper. Left: the anaglyph stereo butterfly (Anaglyph 3D) original image (OI) decomposed into magenta (M) and cyan–yellow (CY) channels (top), and the corresponding reconstructed retina E-paper images (RI) for each eye (bottom). Scale bars, 200 μm. Middle: microscope images of individual M and CY channel pixels, illustrating submicrometre pattern fidelity. Scale bars, 2 μm. Right: full-colour anaglyph butterfly image: anaglyph 3D butterfly original image (top) and simulated retina E-paper reconstruction (bottom) demonstrating high-resolution 3D depth rendering. Scale bars, 200 μm. Original butterfly image licensed from Adobe. c, High-resolution display of The Kiss on retina E-paper versus iPhone 15. Photographs comparing the display of The Kiss on an iPhone 15 and retina E-paper. The surface area of the retina E-paper is ~1/4,000 times smaller than the iPhone 15. SEM and microscope images confirm that the displayed colours are generated by precisely arranged CMY subpixels. Scale bars, 2 μm (top and bottom left) 200 μm (right). Image of The Kiss reproduced with permission from Kingston Frameworks. d, Electrochemical display of The Kiss by retina E-paper. Photographs showing the display of The Kiss on retina E-paper in the on (left) and off (right) states, demonstrating reversible colour modulation when electrochemically tuned.
Figure 4b illustrates an ultra-high-resolution anaglyph 3D display, achieved by encoding stereo image pairs (anaglyph 3D original image) into complementary colour channels—M for the left eye and CY for the right eye—and reconstructing them using the retina E-paper. This demonstration serves as a proof of concept for binocular disparity rendering, a fundamental mechanism underlying stereoscopic vision in virtual reality systems. As the retina E-paper does not connect to TFT arrays to individually adjust each subpixel, it achieves colour rendering solely through the precise geometric design of M–CY metapixels (Extended Data Fig. 7). By precisely reconstructing left and right eye images with submicrometre-resolution metapixels, the device successfully generates a full-colour 3D image (anaglyph 3D; Fig. 4b) through passive, compact optical configurations. This demonstrates an anaglyph 3D display with resolution exceeding 35,000 PPI (M) and 30,000 PPI (CY). Furthermore, this demonstration highlights the versatility of the retina E-paper platform, which can operate not only in full-colour mode but also in monochrome or dual-primary channel formats, broadening its applicability across a range of advanced display technologies.
Figure 4c presents a reconstructed full-colour image of The Kiss, directly comparing the retina E-paper with a commercial mobile-phone display (iPhone 15) in terms of both physical dimensions and display resolution. Whereas the iPhone screen measures 147.6 mm × 71.6 mm, the retina E-paper is only 1.9 mm × 1.4 mm, amounting to merely ~1/4,000th the area of the smartphone display. Despite this minuscule size, the retina E-paper achieves a resolution of 4,300 × 700, similar to the resolution of the smartphone display (2,556 × 1,179). Due to the inherent challenges of accurately controlling the reflectance of subpixels, as well as the narrower colour gamut compared with emissive displays, the perceived colour saturation of the retina E-paper is lower than that of the iPhone 15; however, this is the first demonstration of full-colour imaging achieved by three primary colour metapixels at such an ultra-high resolution. With an average pixel size of only around 560 nm, the display reaches an unprecedented >25,000 PPI, surpassing the resolution requirements for ultimate virtual reality displays. High-magnification microscope (×100) and SEM images (Fig. 4b, right) further confirm the well-ordered CMY metapixel arrangement and the vibrant colour rendering, validating the ultra-high resolution of the retina E-paper.
To evaluate its electrically tunable colour performance, we reproduced The Kiss using CMY metapixels in an electrolyte environment. To maintain the presence of RGB pixels in the intermediate regions, the dimensions of the CMY metapixels were further optimized: C (D = 280 nm, W = 20 nm), M (D = 220 nm, W = 80 nm) and Y (D = 300 nm, W = 80 nm). The merged RGB subpixel spacing was adjusted accordingly: B (T1 = 40 nm), R (T2 = 300 nm) and G (T3 = 60 nm) (Extended Data Fig. 8). Extended Data Fig. 9a presents the corresponding reflection spectra of CMY pixels in their colour and dark states. Figure 4d showcases the photos of The Kiss under colour and dark states in the electrolyte. Owing to the weaker Mie scattering of WO3 nanodisks in an electrolyte environment compared with air, the displayed colours seem less saturated, with a noticeable reduction in extinction in the red region, resulting in an overall red-shifted colour. However, the system exhibits a distinct reflectance modulation between on and off states, highlighting its potential for dynamic display applications. Extended Data Fig. 9b compares the colour gamut coverage of commercial emissive displays, the retina E-paper in both air and electrolyte and the commercial colour electrophoretic display. Although the colour gamut of retina E-paper remains narrower than emissive displays, its performance in both air and electrolyte significantly surpasses a commercial colour E-reader34.
Currently, more than 80% of the information people perceive is through visual signals35. With the development of Internet-of-Things-based technology and increasing information transfer speeds, the demand for next-generation visual display technologies keeps growing. Retina E-paper not only reaches the theoretical resolution limit of human vision but also offers exceptional visibility. It enables full-colour video display while maintaining high reflectivity and optical contrast, which is promising for realizing ultimate virtual reality displays.
Unlike conventional emissive displays, retina E-paper devices require front-illumination to enable image visibility. Extended Data Fig. 10 illustrates two distinct optical architectures that accommodate this requirement: one compatible with conventional virtual reality headsets, and the other tailored to state-of-the-art, waveguide-based augmented reality–virtual reality lenses36. In fact, retina E-paper also holds significant potential for augmented reality applications, as it can leverage ambient light as the illumination source. This inherent compatibility with the environment enables natural visual integration, reducing reliance on light engines. Furthermore, as the primary illumination is provided by ambient light, its low power consumption enables substantial downsizing of the battery and even opens up the possibility of fully self-powered displays when combined with solar cells (with a typical output of ~15 mW cm–2).
Despite its advantages, retina E-paper requires further optimization in colour gamut, refresh rate, operational stability and lifetime. Lowering the operating voltage and exploring alternative electrolytes represent promising engineering routes to extend device durability and reduce energy consumption. Moreover, its ultra-high resolution also necessitates the development of ultra-high-resolution TFT arrays for independent pixel control, which will enable fully addressable, large-area displays and is therefore a critical direction for future research and technological development. Looking ahead, we anticipate significant advancements in this field and firmly believe that the evolution of the retina E-paper will ultimately influence everyone.
