FROM EDN EUROPE: Thin, flat, and low power: The ideal display is (still) just around the corner

This year will mark a golden anniversary; in December 1953, the Federal Communications Commission approved RCA's proposal for a colour-television service in the United States. It wasn't the first colour-TV service, but it was the first to be based on a single CRT, and RCA developed the shadow-mask tube for it in the remarkably short time of only months. That RCA team would have no difficulty recognising the CRT in almost every TV set produced worldwide today. The basic format of the shadowmask tube has remained essentially unchanged for 50 years, and it remains one of the most pleasing, versatile, and efficient large-display vehicles at our disposal today.

Domestic television is one of the few areas in which CRTs still hold sway, however, and the race is on to replace these CRT TVs with the long-sought, large-screen, flat-panel product. PDPs (plasma-display panels) and large TFT (thin-film-transistor) liquid-crystal panels have so far penetrated only the premium domestic market and look unlikely to achieve the manufacturing cost levels they will need to penetrate volume consumer markets. Meanwhile, the large-format, rear-projection domestic TV is almost entirely a US phenomenon and makes up a very small part of the European market. Technologies such as OLEDs (organic light-emitting diodes), which some hail as the ultimate victors in this race, are still a long way from fulfilling that promise—but they are starting to appear in smaller display formats and offer a new and useful option for human-interface components.

Smaller displays are differentiated more by application but have not for some time seen many fundamental innovations. However, the situation is now changing, and a wide range of consumer-, industrial-, instrumentation-, and portable-product options are coming to the market now or within months. Much depends on the appropriate level of complexity for the display function you need to implement. To begin at the top end of the spectrum, if your application demands full graphics (with or without colour), the major option remains the LCD panel in passive-matrix (for smaller units) or active-matrix form.

One display trend has been to make LCD panels more versatile and readable in a wider range of ambient-lighting conditions; backlighting a standard LCD panel with enough intensity to provide full daylight readability requires considerable power. If you succeed, the power that the light source absorbs will likely consume a large part of your overall power budget.

Most of the major LCD-panel suppliers have been adding transflective full-colour panels to their ranges. As their name implies, these panels work in either transmissive mode (with backlighting) or in reflective mode (using incident ambient light). However, their designers face an overall "photon-budget" problem. For a display to operate in reflective mode, each pixel site must contain a path for ambient light to reach a reflective element lying behind the active layers of the display; the light reflects and returns, through the liquid-crystal shutter and filters of the active pixel itself, to your eye. LCD designers must allocate part of the area of each pixel site to provide that window for incoming light, reducing the percentage of the pixel site that's available for the display itself. Because the display loses efficiency in light transmitted in backlit mode, you must step up the backlighting level to achieve equivalent display brightness. As a consequence, little incident light ever makes it back out to the outside world in reflective mode. (Recall that, for an average TFT LCD panel, the fraction of the light incident on its rear surface that makes it to the outside world is a single-digit percentage. The LCD makes a better shade than a window!)

Transflective displays can look good in readability in reflective mode, but, to achieve this readability, the display requires incident-light levels approaching full sunlight or at least bright daylight. You cannot run displays in reflective mode in most indoor environments and expect the readability to be comparable with backlit mode in low ambient light. However, backlit and incident-light modes are additive (allowing for reflection off the front of the panel), and if your power budget allows the use of continuous backlighting, you can have a pleasing display that is readable in a wide range of ambient-light levels.

Other factors to take into account include the viewing angle. Although designs vary, the reflection of incident light back to the pixel uses some degree of scattering or diffusion rather than simple specular reflection, which would produce a pronounced viewing-angle-pattern shape. Nevertheless, designers will have made some assumptions about the likely positioning of the ambient-light source relative to the viewer's eye, and you should expect some limitation or variation of the acceptable viewing angle as displays change from predominantly backlit to reflective mode. Published light-level figures and viewing-angle polar diagrams are a guide, at best; you will need to power up a sample and view it from all angles while altering the ambient lighting to get a real grasp of its appearance in your application.

But what if your application has a low power budget and cannot afford the backlight or even the few milliamps necessary to continuously drive the LCD panel itself, but you would like the panel to display information all the time? About to come to market is a technology that, in effect, gives an LCD cell a "memory." With this construction, you will be able to write an image to an LCD and remove the power, and the display will retain the image virtually indefinitely. ZBD (Zenithal Bistable Devices) created this technique, which adds a finely ridged grating to the inner glass surface of an LCD cell of STN (super-twist-nematic) construction. (Recall that a passive LCD is multiplexed; the drivers scan the pixels row by row, and persistence effects give a continuous display—hence, the need to move to an active device, or TFT, with per-pixel transistor switches, when pixel count rises.) The presence of the grating "latches" the polarisation state of the liquid crystals and retains it when you remove the power. This ability allows passive STN devices to compete with TFTs, because they retain the state of each pixel from one scan to the next, and they retain the overall image with no power. Despite the fact that the latching is in some sense a mechanical effect—at the molecular level—ZBD says that mechanical shock does not disturb the stored image; it remains until you apply power and rewrite it. Further, variations in the grating structure allow parts of a pixel cell to latch at different applied voltages, so a display can store a grey-scale image of at least seven and possibly as many as 64 levels, the company says.

ZBD's chief financial officer, Richard Scanlon, adds that you can port the process to any LCD line with few manufacturing changes. For this reason, he expects to see the IP (intellectual-property)-licensing deals now occurring to translate into products from companies such as Varitronix before the end of 2003. Companies can manufacture the grating in a process similar to that used in pressing CDs or DVDs, by applying the grating to the glass of a conventional panel. Look for established LCD makers to market these panels as "bistable LCDs."

And what of those small colour panels that the latest generation of cell phones features? Those, too, are becoming available to a more general market. At Sony, LCD product manager Matthew Tapping notes that only two factors really matter for the cell-phone market: how good the display looks in readability and, especially, appearance and its power consumption. Sony is building active panels using low-temperature polysilicon with all the drivers on the glass itself, including, dc/dc converter, row and column drivers, chip selects, and a timing generator. For handheld products, this setup yields a 1.94-in. (49.3-mm) display of 128×160 pixels (portrait format) using a vertical-RGB-stripe mode that measures 37.1×51.5×3.2 mm. Cell phones will first use the unit, which will become available for sampling in June and begin production in August. Driving this display with moving pictures, at 25 frames/sec and using a full palette of 65,000 colours will require about 3 mW, Tapping notes. This requirement is dwarfed by the 150 to 200 mW necessary for the backlight. If a battery drives your handheld application, you may need to reduce the inactivity time-out until the backlight switches off to only a few seconds.

Densitron has just opened a large-scale manufacturing facility in China to produce LCD panels. Its range starts with a 1.9-in., colour-STN module offering 4096 colours and, again, 128×160 pixels, for an active-display area of 33.1×42.4 mm in a module measuring 38.5×51.9 mm (Figure 1). This screen is transflective, but larger screens in the range will offer a choice of transmissive or transflective properties.

Kyocera's offering in the sector begins at 2.5 in. (one-eighth VGA), extends to 10.4 in. and provides you with a symmetrical bezel that centres the active area in the module outline and may assist with your product's mechanical design. (The customary arrangement places the drive electronics offset to one edge.)

Work is also continuing on improving the characteristics of the deposited silicon (on the glass panel of the LCD cell) in which the TFTs are fabricated. Polysilicon is better than amorphous silicon and better still is so-called C-G (continuous-grain) silicon. C-G silicon exhibits electron mobilities several times those of polysilicon and several hundred times those of amorphous silicon. C-G silicon's properties are sufficiently close to those of bulk silicon to allow you to integrate more complex circuitry, such as drivers, on the same substrate. Sharp, for example, calls this concept its System LCD. As the driver devices diffuse into the deposited silicon, the space that packaged or surface-mounted chips occupy becomes smaller, and the assembly can be significantly thinner.

Nevertheless, some industry watchers expect some form of OLED to eventually supersede LCD, and manufacturers have spent years developing these devices. So, you can now buy and use displays based on OLEDs. OLED technology is emissive; layered deposition of organic semiconductor materials, rather than the more familiar metallic semiconductors, form light-emitting structures. Two fundamental technologies, generically known as "small molecule" and "polymer," are under development. Their key proponents are Kodak and CDT (Cambridge Display Technology), respectively. Both enterprises license the IP in their technologies to display manufacturers. After years of development, OLEDs have enough absolute brightness to offer good contrast in bright ambient lighting. They are thin and require no backlighting, and—as the light is emitted from (close to) the surface of the structure—they are efficient, with none of the losses of, for example, an LCD panel in which the light is attenuated as it passes through polarisers and filters. Further efficiency comes from the fact that only the active pixels generate light and consume power; the subtractive structure of an LCD requires full backlighting all the time. OLEDs have a wide viewing angle, and they are fast, active-electronic devices. Full video data rates and operation at low temperatures are well within their switching rates. You can design OLEDs to emit all three primary colours with good purity, and these colours combine to give a good colour gamut.

Beyond those features, however, OLEDs offer the possibility of large, low-cost displays on flexible substrates. Manufacturers can deposit the organic materials by printing processes; CDT and other companies have demonstrated the feasibility of ink-jet printing of full-colour displays. As with other matrix-based displays, you can structure OLEDs as passive-matrix devices, which are adequate for lower resolution and display size, or active-matrix devices for higher resolution and panel size. And, as with other active-matrix devices, you can build the switching transistors in any of the available deposited-silicon types. The ultimate combination of the possibilities that OLEDs hold might be a 1m or larger, full-colour graphics display that you produce in volume by ink-jet printing the active materials onto a flexible substrate, perhaps thin enough to call a film or foil, on a continuous, reel-to-reel production facility (see sidebar "Toward the 50-in. TV panel").

Needless to say, that dream is some ways off. The multifaceted problems of handling the organic chemistry that OLEDs involve beset their development. Although researchers have produced all three primary colours, blue has been more difficult to create than red or green and still lags behind those colours in ultimate lifetime, constraining progress toward full-colour displays. As the emitters age, a display built in early materials would suffer a colour shift due to differential aging effects. Absolute lifetimes (defined as the time to the –3-dB light-output level) now extend to acceptable values for commercial devices in all three colours. CDT quotes 11,000 hours for blue emitters versus 40,000 hours for green and 60,000 for red (Figure 2). Therefore, many of today's first product offerings are monochrome, although with a choice of colour. However, Kodak has announced the AM550L full-colour OLED module, which it uses as a digital-camera-back display on its Easyshare LS633 camera. The QVGA (521×218-pixel) module measures 5.5 cm diagonally, and you can evaluate it with the AMEV1-100 kit, which includes a controller, software, and interfaces to standard video-driver boards. Kodak produces the display through a joint venture with Sanyo (SK Display Corp) and, also with Sanyo, has demonstrated 5.5- and 15-in.-diagonal, active-matrix, full-colour panels.

Contact with the atmosphere and with water vapour in particular rapidly degrades OLED chemistry, so sealing the deposited chemistry is critical. For this and other reasons, first offerings have been on glass substrates; plastic and flexibles will follow. Hence, partnerships have arisen, such as the one recently announced between the displays division of Samsung and Vitex Systems. Samsung, like virtually all of the other large display manufacturers, has taken OLED-technology licences, and Vitex Systems specialises in thin-film barrier coatings. Samsung aims to produce thinner, glass-based displays by eliminating the need for a second glass sheet acting as a sealing plate.

One of the first of CDT-technology-based OLED displays to appear on a product is on a shaver, the Norelco Spectra. Norelco is the US branding for the product line sold in Europe as Philishave. A similar monochrome alphanumeric display will soon appear on an MP3 player from Delta Optoelectronics in Taiwan. CDT's vice president of business development, Stewart Hough, says that he expects progress in OLED to track that of LCDs, progressing rapidly through monochrome displays to colour and then scaling upward in size, with polymer gaining ground to catch up on small-molecule technology's initial lead.

Typical of the first-generation displays beginning to emerge is the product currently shipping from DuPont Displays, a passive-matrix monochrome device of 128×64 pixels with 2.1-in. diagonal measurement and a yellow/green emission (Figure 3). DuPont has announced a 160×160-pixel version of the same technology and a QVGA colour active-matrix display that measures 4 in. diagonally. DuPont, developing polymer OLED technology from CDT, exemplifies the network of intercompany cooperation that is attempting to drive OLED forward; it has agreements, for instance, with UDC (Universal Display Corp), covering aspects of small-molecule and polymer chemistry. In particular, this agreement focuses on the solution chemistry that will be necessary to convert the active compounds into ink-jet-printable form. DuPont also has an agreement with SEL (Semiconductor Energy Laboratory) covering the migration of polysilicon TFTs to the OLED platform. Jutta Rasp, director of DuPont Displays in Europe, says that the prices of OLED products will be similar to those for comparable-complexity LCD panels, complete with backlights.

Other companies that have announced or are shipping OLED products include Philips and Densitron, which offers a dot-matrix part of 64×128 pixels, plus a 132×176-pixel, 2-in.-diagonal, full-colour display that offers 40 to 50-cd/m² brightness. Densitron has also built a 250-cd/m², single-digit, 47-mm-high display with OLED technology for use in public-information displays.

OLEDs are current-driven devices that require dedicated driver ICs to achieve the correct performance in lifetime, brightness, and grey scale. One of the first sources of such drivers is Clare Micronix (Figure 4); offerings are also available from TDK, STMicroelectronics, and Philips Semiconductors.

Operating on a somewhat different structure is a small colour panel from TDK; TDK can supply multicolour but not full-colour OLED devices, using a composite OLED structure to produce white light, which it then filters to yield the colours (Figure 5). You can have 16 colours (two base colours with four-level grey-scale drive) or 64 colours (three base colours). TDK has for some time been supplying monochrome OLED displays in small dot-matrix panels to the automotive market and will develop the technology toward full colour. The current product offers an active area of 23.5×79 mm with 240×64 dots.

Meanwhile, a consortium of companies cooperating with the objective of establishing flat-panel OLED displays in Germany has formed DFF, the Flat-Panel Display Forum (www.displayforum.de). The forum has 79 members, 13 of which are moving toward a pilot manufacturing facility.

One further OLED technology deserves mention. In 2002, CDT acquired the IP of Opsys, another Cambridge (United Kingdom) start-up that had developed the "dendritic" molecule. One of the problems of small-molecule technology is the tendency of adjacent structures with similar electronic bandgap characteristics to absorb light as readily as their neighbours emit it, limiting overall efficiency. Dendritic technology builds the so-called snowflake molecule, in which other organic structures surround the active, light-emitting "core," in effect, giving the light-emitter elbowroom as well as fine-tuning other properties of the structure. Dendritic technology potentially offers high electrical/light-conversion efficiency and low manufacturing cost, with only one active layer to deposit or print. But it is less developed than polymer or small-molecule technology, which has the longest history.

At a lower level of complexity of displayed information, a vast market exists for products that present a limited amount of data. Chris Barnardo, marketing director at Pelikon, quotes estimates that the segmented, nonmultiplexed-display market is worth some $4 billion worldwide. Occupying a niche in displays for domestic appliances and consumer products that require only segmented displays without multiplexed drive, Pelikon is exploiting electroluminescent-emission technology to produce a custom product called InterfaceDisplay. Inorganic electroluminescent materials produce a limited range of colours, mainly in the blue and the green parts of the spectrum. They are moderately efficient, at around 10 to 15 lm/W but are limited in absolute brightness achievable to 50 to 200 cd. You might associate electroluminescence with backlighting, but Pelikon has developed a technique for printing discrete areas or segments and addressing them on flexible substrates. Because the display is deposited on a flexible film, you can overlay it on conventional membrane switch assemblies. The film, complete with emissive layer, can accommodate sufficient travel to operate collapsing metal-dome switch mechanisms without impacting the performance of the display film.

Using this technology, you can produce a display with fixed segments or areas of illumination as small as 1×1 mm or as large as required in a flexible substrate less than 0.5 mm thick. The display is not usable outdoors, because its absolute brightness is too low, but it is readable in all indoor ambient-light conditions, and you can dim it to provide a low brightness level for use in darkened conditions. In addition to its switch facility, you can arrange to illuminate only the legends over switch options that are active at any time, guiding a user through a device's operation.

The technology sandwiches the light-emitting material between two planar electrodes; the front-face one is transparent and excited by an ac drive voltage of 100 to 200V at 200 Hz to 1 kHz. As you might expect from the construction, the electrical load is mainly capacitive. You can couple separate lit segments that need to be on at the same time and treat them as a single segment. Pelikon has developed a 32-channel ASIC to accompany the printed-pattern technology to generate the driving voltage from supplies as low as one battery cell. The ASIC includes a microcontroller core, and Pelikon can package the entire offering as a customised option with a tailored display pattern and software. You might use such a display on, for example, a remote control, an audio unit, or a set-top box, as a space-efficient alternative to a vacuum-fluorescent display.

Author Information

You can reach Editor Graham Prophet at +44 118 935 1650, fax +44 118 935 1670, e-mail gprophet@reedbusiness.com.