Die-Sized Displays Enable New Applications

Haviland Wright, Mark Handschy, Diplaytech, Inc. Longmont, CO -- Semiconductor International, 9/1/1998

Microdisplays provide a low-cost, small cost form-factor solution for many applications. Some microdisplay applications are familiar: television, computer monitors and instrumentation displays (Fig. 1). Others are more cutting edge: wearable computer displays, personal video viewers and heads-up endoscopic surgical displays (Fig. 2). For each application, design trade-offs will determine the microdisplay technology used. Each technology offers a different roadmap for future performance and cost development. All share the common characteristic that microdisplays are imaging components whose performance depends on overall system design.

Microdisplay types

There are many ways to differentiate between microdisplays, including power requirements, achievable resolution at a given display size and the ability to produce brightness and color in particular system configurations. From a system design perspective, there are three basic kinds of microdisplays: reflective, transmissive and emissive.

Reflective microdisplays are made with integrated circuits (ICs) as backplanes. In some cases the IC is used to control liquid crystal (LC) switching. In other cases the IC is used to control micro-mechanical structures on the display surface. In all cases, reflective microdisplays have the advantage of utilizing most of the display surface without losing surface area to control circuits. Reflective displays depend on an external light source for display brightness and color.

Development efforts in reflective microdisplays have been in-tense. The first IC-based reflective display product was the Digital Micromirror Device (DMD) introduced by Texas Instruments (TI, Dallas, Texas) in 1994. Each pixel is a tiny tilting mirror that uses time-division grayscale techniques to weight the colors produced in three-panel color-channel systems or in single-panel sequential-color systems.

Of the many liquid crystal on silicon (LCOS) display technologies, Displaytech's ferroelectric liquid crystal displays (FLCD) are the closest to TI's DMD. Its displays are entirely digital, like DMD, and it uses binary ferroelectric liquid crystals much in the way that TI uses its binary micromechanical mirrors.

Other LCOS displays utilize nematic liquid crystals to produce analog grayscale. These materials are commonly available, making LCOS devices the focus of intense development competition. Both single-panel and three-panel systems have been demonstrated.

Transmissive microdisplays modulate light passing through the display. In some cases, color is produced by filters in the displays. In others, color-switchable light sources provide both color and brightness. In all cases, transmissive microdisplays lose some optically useful surface area to control circuitry. Kopin Corp. (Taunton, Mass.) is pursuing an approach that utilizes a transparent-substrate silicon-on-insulator (SOI) integrated circuit that can be processed in a regular IC fab with proprietary silicon lift-off processing.

Emissive microdisplays produce their own light. One approach used by Planar Systems (Beaverton, Ore.), applies an electroluminescent phosphor to the surface of an IC. An alternating voltage is applied to a transparent electrode on the face of the phosphor opposite the IC. High-voltage circuitry making up an array of pixels on the IC determines whether alternating current is passed through the phosphor or blocked at any particular pixel location.

Another approach being pursued by Micron Display Corp. (Boise, Idaho), uses field-emitting devices (FEDs) made on the upper surface of an IC. The FEDs act as cold electron-emitting cathodes that excite a phosphor separated from the IC surface by a thin vacuum. The FEDs are made by micro-machining sharp points, or Spindt tips, out of a refractory metal. Otherwise the device functions much like a cathode ray tube (CRT). Each pixel contains many tips, so that variations in the responses of individual tips do not cause pixel-to-pixel variations.

Fig. 2. New microdisplay technologies enable a wide variety of personal display applications.

Silicon-based technologies

In conventional active-matrix LCDs, a thin film of liquid crystal is sandwiched between two transparent plates. An array of thin film transistors allows the display to be transmissive. However, circuitry remains opaque. The fraction of the pixel open to light transmission, called the aperture ratio, often drops as low as 30% in high resolution miniature LCDs. Color is produced in one of two ways. The substrate opposite the active matrix may carry an array of red, green and blue filter dots, allowing display of full-color images on a single panel using color triads. Alternately, three panels may be used, each illuminated with one of the three primary light colors and their images combined to produce a full-color image.

By allowing the display to be reflective instead of transmissive, a fully conventional IC can be substituted for the TFT array. Aluminum pads on the upper surface of the IC serve as both pixel mirrors and liquid crystal drive electrodes. In an IC process with multiple metalization layers, the pixel mirrors overlay the pixel circuitry, allowing aperture ratios to be much higher than in transmissive displays.

Numerous companies, including IBM (Hopewell Junction, N.Y.), JVC (Yokohama, Japan), Colorado Micro Display, S-Vision (Santa Clara, Calif.), Spatialight (Novato, Calif.) and others currently are developing reflective microdisplays using the same nematic liquid crystals used in conventional transmissive displays. Displaytech is developing reflective microdisplays using ferroelectric liquid crystals (FLCs).

Perhaps the most radical approach to joining light-modulation functionality with electronic functionality is the DMD pioneered by TI. It has micromachined tiltable mirrors on top of an array of CMOS pixel circuits. Each mirror can be switched by signals from the underlying circuitry between two tilt states, which, by reflecting incident light in different directions, can produce high contrast images. The response time of the mirrors is measured in microseconds, enabling production of gray-scale through "digital" time-division techniques.

Applications for microdisplays

The simplicity and flexibility of designing microdisplays into products, combined with other characteristics like fast switching speeds, small size and sequential color, provide equipment manufacturers with a high-quality, low-cost versatile display product. Since microdisplays are generally too small to be used for direct view applications, their images must be magnified or projected for presentation to users.

Projection applications include rear projection displays, in which the projection system performs the function that is provided most often by CRTs today. Front projection systems include the familiar multimedia and personal projectors that produce images on separate screens. Magnified displays are generally used in systems where the display operates close to the user's eye.

Projection applications

The worldwide market for all types of projection devices is significantly larger and more mature than the magnified personal display marketplace. In 1997, 1.56 million projectors were sold for a total of $6.4 billion, averaging $4103 per unit (Table 1). Sales are forecasted to increase to 3.8 million units and $11.2 billion in 2004,¹ for an average price per unit of $2947.

The projection market is presently dominated by rear CRT projection systems, but major growth is occurring in the front- and rear-LCD and chip-based projector market. The business segment, particularly in the portable and ultra-portable projector area, is the fastest growing segment in this market. Stanford Resources estimates that this part of the business segment will grow to 1 million units, or 26% of total projector unit sales, by 2004.²

Wide screen televisions and monitors are one market area that microdisplays can serve. A microdisplay chip can create a projected image that is brighter and sharper than today's complex flat panel displays and at a fraction of the cost. Another application is home theater. Microdisplays can also be used for multimedia projectors. Many microdisplay chips can withstand the extreme temperatures common in high-intensity projection systems and reduce the manufacturer's product cost, weight and complexity.

Personal microdisplays

There has been renewed interest in personal displays recently. Largely, this is a result of advances in microdisplay technology, an increase in bandwidth and data transmission capabilities, the decrease in the size of electronic products and the integration of multi-functional applications into mobile products. Though there are many applications, the most promising examples are the most obvious (Table 2).

"Smart" telephones are digital telephones that incorporate at least one other form of information beyond voice, such as e-mail, faxes or internet use. The digital telephone market in the United States is still relatively new, and the infrastructure is not complete. However, given the popularity of digital "smart" telephones in Europe, with estimated revenues of $604 million in 1998, compared to $269 million in the United States,² it is forecasted that the U.S. market will grow significantly over the next five years.³

Digital cameras have been available for a number of years, though mainly concentrated at the high-end of the
photography market. In the last 12-18 months, this market has grown rapidly. Digital cameras have moved down the price curve to incorporate sub-$400 offerings. Microdisplays consume less power than other digital camera viewfinders, allowing them to extend battery life substantially, which currently is a market-limiting problem.

Digital camcorders comprise the most mature market segment in terms of using microdisplays. Most camcorders today have a view- finder feature. The viewfinder displays are generally 2-5 in. in both monochrome and color. They are direct-view displays based primarily on TFT-LCDs.

Handheld PCs (HPCs) are increasing in sophistication as new operating systems such as Windows CE and PalmPilot Grafitti evolve and reach implementation. Some systems support e-mail retrieval, fax capability and web browsing. The Windows CE operating system now supports color. Sharp (Osaka, Japan), which uses a proprietary operating system, introduced the first color PDA in Japan in 1997. Hewlett Packard (HP, Palo Alto, Calif.) introduced its first color HPC in January.

Medical applications primarily involve instruments like endoscopes. This segment is less price-sensitive than others, but it is not particularly large. Color displays with resolutions above SVGA are required.

Laptop computers represent the largest potential segment for microdisplays. For any microdisplay to penetrate this market, it must be capable of delivering at least the same clarity, resolution and color depth that users have come to expect from their desktop systems.

Wearable computers combine the functionality of laptop computers with the portability required of the new mobile products. Many of these products have neither a full-size display screen nor a keyboard. Rather, they use voice recognition software for data input and a monocular miniature magnified display for data output. The monocular display allows individuals to use the wearable computer as a reference resource with hands-free operation and still work within their particular functions.

Video-to-go is another market on the horizon. A microdisplay panel can be placed in a handheld unit with the appropriate optics to make a comfortable, easy-on-the-eyes portable video player. It is possible for the image to have enough clarity and brightness for many to prefer it to a traditional TV screen.

Miniature display technology

Though the silicon IC industry has existed for over 30 years, only recently has the possibility of making integrated microdisplays emerged. Work on microdisplays began in the early 1980s, but is developing more rapidly now due to important changes in both the business model and technology used by the silicon industry. First, today's silicon foundry model allows greater access to custom IC services by separating the requirements for display technology ownership from those of fab ownership. Second, the push for smaller line widths has resulted in widespread adoption of planarization techniques, which also provides a high quality optical surface on the IC.

These changes have resulted in the development of a diverse range of microdisplay technologies, with a wide variety of silicon requirements and operating characteristics. Microdisplays share some of the usual cost vs. size tradeoffs faced by electronic-only ICs but also face some tradeoffs unique to their optical function.

Achievable resolution is determined by the feasible pixel pitch. Maximum panel size for most silicon-based microdisplay panels is set by stepper field limitations on the backplane, while minimum pixel pitch is determined by pixel complexity. Almost 3500 lines can be fit within a 20320 mm stepper field with a pixel pitch of 5 mm.

The technologies having more complex pixels are capable of resolutions serving today's mainstream markets. Examples include FEDs, which rely on redundant structures, active matrix electroluminescence (AMEL) that relys on high voltages and micromechanical mechanisms such as DMD. Technologies with simpler pixels, such as low-voltage LCD devices, enable display resolutions approaching that of the printed page.

When maximum resolution is not the goal, panels can be made smaller than a full stepper field, in which case pixel complexity determines backplane cost through the usual yield-based mechanism. A smaller pixel pitch increases the number of devices per wafer, decreasing backplane cost. Also, assuming a constant defect density, a smaller die size increases die yield. Clearly, a substantial advantage is obtained by reducing pixel pitch.

Standard vs. nonstandard CMOS

The technologies described above all succeed in integrating microdisplays onto silicon ICs. The requirements they place on IC backplane processing though, range from nearly standard to quite exotic.

For the reflective technologies described above it is important that the upper surface of the IC be very flat and highly reflective. Flatness requirements for liquid crystal technologies are measured in fractions of a micron. Fortunately, available chemical-mechanical planarization (CMP) processes are easily capable of meeting this requirement. Meeting the reflectivity requirements usually involves omitting antireflection coatings from the last-metal layer. To squeeze out the last few percentage points of light return, many developers are using thinner final metal layers along with reflectivity enhancing dielectric passivation coatings. Micromechanical devices like DMD require extra mask steps with specialized micromachining.

The silicon backplanes required for the various display technologies differ perhaps most significantly in the drive voltages they require. Some, such as AMEL, need several tens of volts, requiring process variants such as lightly doped drains. LCOS devices vary significantly among themselves depending on which LC operating mode is chosen. Many of the technologies require pixel drive voltages in the 9-12 V range, some operating at 5V and below, and FLC devices even operate below 3.3 V.

Color production

Microdisplays produce color in a variety of ways. One of the oldest techniques groups three sub-pixels into a color triad to make a full-color pixel. A red, green or blue color filter is placed over each sub-pixel. This technique is the same one used in color televisions and LCD camcorder viewfinders. This color-production technique requires sophisticated processing and patterning capability and is less favored by most developers of silicon-based microdisplays.

High-output projection displays cannot afford to waste any light. In this case, the color-production technique of choice utilizes three separate microdisplay panels. The white-light illumination is split into three separate beams of red, green and blue light by dichroic mirrors. Each of these beams illuminates one microdisplay. The separate red, green and blue microdisplay images are then optically combined to project a full-color image on the screen.

The most recent color-production technique is field-sequential color. Red, green and blue light sources are reflected off the display in sequence. Only one panel is required, but it must be capable of presenting grayscale images at three times the normal video frame rate. The color switching can be accomplished with an electro-optic color filter, with a rotating color filter wheel or with intrinsically colored light sources such as LEDs. The sequential-color technique simplifies the functionality required of the microdisplay, placing the color generating means elsewhere in the display system, often at a substantial cost advantage.

Conclusion

Although still an emerging technology, microdisplays could trigger a fundamental shift in the way products are designed and produced for some of the industry's largest consumer markets. Certainly, any significant displacement of CRTs on a price/performance basis in television or computer monitors would be such a shift. Microdisplays also have the potential to serve other markets.

Small, light-weight, low-power, low-cost, fully capable displays have important product implications for products that exploit broadly available information for entertainment, business and communication. Seen in this light, microdisplays have the potential to be revolutionary in the same sense that the mouse and graphical user interfaces have been. By challenging assumptions about how the industry views and interacts with visual information, microdisplays will open up new possibilities for products and services that do not exist yet.

References

1. Electronic Display World, November 1997.
2. IDC, June 3, 1997.
3. Electronic Display World, November 1997.

	Haviland Wright joined the Displaytech Board of Directors in 1994 and became CEO in 1995. He received his doctorate and masters in business administration from the Wharton School at the University of Pennsylvania and has held faculty positions at the University of Colorado at Boulder and the University of Denver. Phone: 303-772-2191 Fax: 303-772-2193 Email: hwright@displaytech.com
	Mark Handschy is a founder of Displaytech, has directed its research since its inception and was named president in 1993. He received his doctorate in physics from the University of Colorado. Handschy has developed a number of FLC light modulators and led development of the company's LightCaster miniature displays. Phone: 303-772-2191 Fax: 303-772-2193 Email: mark@displaytech.com