Introduction
Liquid crystal display (LCD) technology is the most popular of the display/television technologies currently used in computer monitors. The other technologies in use today include organic light-emitting diode (OLED), plasma, digital light processing (DLP) and cathode ray tube (CRT) systems. However, they cannot match all the advantages of LCDs, which are thin, have high resolution, are low cost, and offer low power consumption and long lifetime. On the negative side, LCD systems do not have high contrast, because liquid crystals cannot completely block light in the off state, and are virtually unusable in the presence of bright ambient light. This last point arises because LCDs transmit only 5–10% of the backlight1 because of the polarizer, which blocks more than 50% of the light. Also, each colour filter transmits only 30% of the remainder of the light, and there are some additional layers that decrease transmission even further. It is desirable to increase the efficiency of backlight transmission to decrease energy consumption and improve visibility in brighter environments. As a solution we present a new display technology—the 'telescopic pixel'. 'Telescopic' indicates that each pixel acts as a miniature telescope consisting of a primary and secondary mirror (Fig. 1). Under applied voltage the primary mirror can change its shape from planar to approximately parabolic.
Figure 1: Side view of one telescopic pixel.
a,b, Telescopic pixel in the off state (a) and in the on state (b). Dimensions of the pixel components: primary mirror radius, 50 
m; secondary mirror radius, 25 m; radius of the hole in the membrane, 20 m; gap between the two electrodes, 6 m; distance between the primary and secondary mirror, 175 m. (See Supplementary Information for the optimization details.).
When the pixel is turned off, the primary mirror is planar. Therefore, both the primary and secondary mirror block the passage of light, reflecting it back to the backlight, so that the pixel appears dark (Fig. 1a). When the pixel is turned on, the shape of the primary mirror is close to parabolic (Fig. 1b). In this state, the mirror focuses light onto the secondary mirror. After reflecting from the secondary mirror, light propagates through the hole in the membrane and the pixel appears bright (Fig. 1b).
Although there are already several display technologies available on the market, the use of the telescopic pixel can provide some important advantages for several applications. The three main existing types of displays are transmissive (backlight passes through the pixel)2, 3, 4, 5, 6, reflective (light reflects from the pixel)7, 8, 9, 10, 11, 12 and emissive (the pixel itself generates light)13, 14, 15. Examples of emissive technologies include OLED and plasma. Organic LEDs are currently used in the fabrication of small screens as the contrast and colour quality of such devices are good, but present yield is low and lifetime is limited, which makes fabrication of large screens very expensive. Plasma technology, in contrast, works well for large screens, but its resolution is limited by the fact that the size of its pixels cannot be significantly scaled down. Therefore, for an application requiring large-screen, high-resolution displays, a telescopic pixel arrangement can offer advantages. The materials from which it is made are stable, and it does not suffer from resolution limitations, because its design is scalable, working well for different sizes of device. In our experiments the pixel size was 100 m, but this can be scaled up or down depending on the desired resolution.
There are two basic types of reflective display technologies. The first uses ambient light7, 8, 9, 10, is very power-efficient, but has low contrast and colour quality. The second uses an expensive microelectromechanical systems (MEMS) chip11, 12 and projection onto a screen. The telescopic pixel operating in reflective mode can be used to produce lower-cost chips for projection televisions (see Supplementary Information). It is simpler than DLP, which gives the telescopic pixel technology a higher fabrication yield, resulting in potentially lower-cost displays.
The most popular example of a display using transmissive technology is LCD, for which the advantages and limitations have already been described. Other technologies are based on different physical phenomena. Electrowetting displays5 do not yet have high contrast and colour quality. UniPixel6 is based on frustrated total internal reflection, a phenomenon that is difficult to optimize for simultaneous high transmission in the on state and low leakage in the off state. In addition, this approach has limitations arising from the non-uniform brightness for bigger screen sizes. Finally, there are displays based on mechanical shutters located on silicon chips2, 3, 4. However, the fabrication process involves etching a hole through a silicon wafer, which at present is expensive and time-consuming. Moreover, the fabrication process for mechanical shutters involves a large number of steps and many masks; this significantly decreases yield and increases overall cost. Finally, the fill factor is low, which decreases overall transmission. The advantages of the telescopic pixel are high transmission efficiency, potentially low cost, as well as relative ease of fabrication and control. One of the current limitations is relatively high operation voltage (120 V); this can be decreased with further design optimization16.
In addition to providing brighter computer displays, telescopic pixel technology could help drive forward the development of low-cost, large-screen displays. The consumer market is demanding affordable, large, high-resolution televisions and displays, but this goal will be hard to achieve without using a modular approach, in which large displays are created by combining smaller building blocks. Currently such modules are being used to produce very expensive reflective modular displays17 of ultrahigh resolution, with the cost being driven up by the large number of expensive MEMS chips that are used for each screen. All attempts to apply such a modular approach to transmissive technologies have resulted in displays with lower light efficiency and contrast than regular LCDs18. The telescopic pixel technology could potentially be used to fabricate cheaper reflective or brighter transmissive modules for very large displays.
The details of the telescopic pixel design and operation are described in the following. The two mirrors are fabricated separately and then stacked together in one pixel. The primary mirror is a 100-m-diameter, 100-nm-thick, suspended circular metal membrane with a hole in the centre. The secondary mirror is a patterned metal film on glass. The change of the primary mirror's shape is accomplished by applying an electrostatic force. When a voltage is applied between the primary mirror (metal membrane) and the transparent electrode (indium tin oxide, ITO), the membrane is pulled towards the transparent electrode (Fig. 1b).
The opening in the membrane is smaller than the secondary mirror to minimize diffraction leakage through the pixel. The first prototype's contrast ratio was 20:1, mainly due to the use of non-collimated back light. This was a limitation of the current prototype, not of the technology. This is supported by simulations (see Supplementary Information, Fig. S5), which show that a ratio of at least 800:1 is possible. The software used for the simulations was a finite-difference time-domain program (OptiFDTD, Optiwave).
Each pixel's output light was projected on a closely located screen. Light diverges after leaving the pixel (Fig. 1b), so for a screen placed at a distance of 0.35 mm, pixel overlap leads to the pixels appearing to cover the whole screen (see Supplementary Information). To maximize the viewing angle, a diffuser similar to that used for LCD fabrication should be used.
A simple calculation can justify the high efficiency of the telescopic pixel backlight transmission. The total light transmitted by the display is a product of the pixel fill factor and the transmission of a single pixel. The circular pixels can be stacked in a two-dimensional (2D) array, one next to the other. Therefore, the maximum fill factor (fraction of the area covered by pixels) is 
/4 (78%). Figure 2 shows such a 2D array of telescopic pixels, where this high fill factor is demonstrated experimentally. This image is a top-view photograph of the telescopic pixel array taken with a microscope. (See Supplementary Information for a video demonstrating telescopic pixels in action, recorded with a videocamera and microscope on a working telescopic pixel prototype.)
Figure 2: Two-dimensional array of telescopic pixels (looking from the primary mirror side).
Pixels are placed next to each other so that the maximum possible fill factor of 78% is achieved. The picture is taken using a digital camera mounted on a microscope.
Full size image (0 KB)The maximum transmission of a single pixel in the on state can be derived from the fact that the secondary mirror has a diameter equal to half that of the primary mirror and blocks 25% of the backlight. Thus, 75% of the backlight will reach the primary mirror. Simulations indicate that 95% of the light from the primary mirror can reach the pixel's output. In the experiment it was measured to be 61%, which can be further optimized.
The total amount of backlight that can be transmitted by a telescopic pixel display based on the experiment is /4 
0.75 61% 
36%, and simulations show that up to 56% is possible. The current experimental value is 3.5–7 times greater than that of LCDs1, and therefore for the same backlight intensity, the telescopic pixel is 3.5–7 times brighter.
The experimental results are summarized in Fig. 3 and Table 1, and show the important pixel properties. (See Supplementary Information for details of the experimental measurements and simulations.) Figure 3a illustrates the measured response of the pixel array to a square wave, and the experimental transfer function of the pixel is shown in Fig. 3b. Figure 3a shows the rise and fall times to be 0.625 ms and 0.61 ms, respectively, which gives a pixel response time of less than 1.5 ms. This means that the technology is not only very fast but may also display colours using a sequential colour red–green–blue backlight (three colour LEDs—red, green and blue—working sequentially, as opposed to the three separate red, green and blue pixels of the LCD arrangement). High operation speed will also help to avoid the artefacts of slower displays that are apparent while playing video and computer games. Moreover, the experimental pixel transfer function shows that the light intensity can smoothly change from zero to one. This means that the grey-scale and colour shades can be realized by varying the intensity in a single timeframe as opposed to the case for binary pixels (for example, in DLP systems), which have to make use of several cycles for every colour value. Thus, telescopic pixel technology can potentially achieve both fast operation and high-quality colour performance.
Figure 3: Pixel performance.
a, Pixel response to an applied square-wave voltage. b, Transfer function of the pixel response.
Full size image (0 KB)
This pixel design also solves a well-known problem of transmissive imagers, in which the free aperture of each pixel is limited by the part of the pixel occupied by opaque circuitry. Using the telescope design, the circuitry can be placed under the primary mirror membrane so that it does not block the light, thus enabling a very high fill factor. Finally, the fabrication process should be inexpensive because of the low-cost materials that can be used and the small number of fabrication steps that are required.
To summarize, the telescopic pixel technology could potentially lead to brighter, higher contrast and lower-power monitors and ultimately enable very large displays at low cost. These are currently hot research topics in the display research community.
Methods
We used standard MEMS techniques to fabricate the prototype of the two-dimensional telescopic pixel array. However, for mass production the fabrication may be accomplished on current LCD production lines, as the processes are compatible with LCD fabrication. The most important device components are the primary and secondary mirrors. The primary mirror is a suspended metal membrane. Similar structures are used in various MEMS devices—sensors and deformable mirrors with adjustable focus16, 19. The fabrication techniques are already well established and mostly based on membrane release accomplished through dry etching (Fig. 4a–d). Fabrication of the secondary mirror was achieved by means of regular patterning of metal on a glass surface (Fig. 4e,f), which is a simple step. Primary and secondary mirrors were first fabricated separately, then aligned and stacked together, separated by a thin plastic spacer (Fig. 4g). (See Supplementary Information for the design, optimization20 and fabrication details, and the video for a demonstration of the operation of the prototype pixel array.)
Figure 4: Design flow (not to scale).
a–d, Stages of primary mirror fabrication: ITO coated glass (a), spin-coating of polyimide (b), aluminium sputtering and patterning (c), and dry etch of polyimide (d). e–f, Stages of secondary mirror fabrication: metal sputtering (e), photolithography followed by etching of the metal (f), and final device alignment (g).
Full size image (0 KB)For mass production, the ITO layer could be substituted by a patterned aluminium layer. The whole pixel could then be made using only three aluminium layers, one layer of polyimide, and one plastic spacer (five relatively low-cost layers in total). This would suggest that high-yield and low-cost fabrication is a possibility.
To test the device we applied a voltage between the transparent electrode and the membranes (Fig. 1). The electrodes were connected to a square-wave signal generator and amplifier with a peak-to-peak voltage of up to 120 V. Light illuminated the device from the bottom, and the pixels were observed from the top through a microscope. An Oriel photomultiplier and amplifier were used to measure the light intensity of the pixel array.
