Business: Silicon carbide products
Number of Employees: 270

No longer just the grit on sandpaper, silicon carbide (SiC) is getting a new image. This versatile material is more heat-tolerant than silicon, nearly as hard as diamond, and able to light up a wall of color on a tall building.

The new image is courtesy of Cree Research, Inc., which used co-funding from NIST's Advanced Technology Program (ATP) to develop new processes for making SiC wafers that are larger and have fewer defects than was previously possible. Since then, SiC has begun appearing in all kinds of applications, from car dashboard lighting to "instant replay" boards in stadiums to experimental high-definition television (HDTV) and blue lasers.

For Cree, riding the success of a rising technology is a thrill. "I know how it must have felt to be in on the ground floor of the space program," Neal Hunter, president and chief executive officer, told stockholders recently. For the nation, the improved SiC wafer quality and yields and corresponding cost reductions achieved through the ATP project have:

• Provided a means to increase substantially the efficiency of electric vehicles and other powered systems, the storage capacity of digital video disks (DVDs) and other optical devices, and the power of communications and radar systems.

• Helped keep a U.S. company at the forefront of highly competitive technology areas, including light-emitting diodes (LEDs), next-generation semiconductors, and blue lasers.

• Enabled the formation of two new companies serving prospective new markets for full-color LED displays (offering more than 16 million colors) and diamond-like jewelry.

The growing enthusiasm for Cree's technology is reflected by the company's $42.5 million in revenues in 1998, a 50 percent increase over the previous year. Much of the growth was in blue LEDs, which Cree has made from SiC wafers for nearly a decade but have been greatly improved, partly as a result of the ATP project. In addition, sales of the raw SiC wafers grew by 42 percent as electronics and semiconductor companies began using them in product development.

The ATP played a pivotal role in Cree's success because it accelerated by 18 months the development of the substrate for all the company's products and the emerging SiC industry. "Everything we make we build on SiC wafers, and ATP invested in the basic technology," says Calvin Carter, co-founder and director of materials technology.

The challenge was to realize the inherent potential of SiC, a largely synthetic material with many useful properties, such as extreme hardness (hence its use in sandpaper) and radiation resistance. Of great interest commercially is its tolerance for high temperatures, which makes it superior to conventional silicon crystals for high-temperature electronics and, because of the high-energy wavelengths involved, blue LEDs.

Cree was established in 1987 by researchers at North Carolina State University who had developed a process for making SiC crystals and wafers. Cree began manufacturing 1-inch wafers and quickly became the leading producer of blue LEDs (which are very small). But yields were limited by defects in the material, and the wafers were too small to be used for high-power applications and integrated circuits.

In the two-year ATP project, which ended in 1994, Cree developed new processes for growing SiC crystals and making them into wafers. The proprietary crystal-growth process involves heating a solid SiC source to convert it into a vapor, which condenses on a cooler "seed." The resulting crystal then is sliced into wafers, which are layered with thin films of SiC or other materials. In this process, for which a patent has been sought, Cree achieved uniformity in layer thickness and increased control of doping (the addition of impurities to achieve the desired properties).

These advances enabled Cree to double wafer size to 2 inches and reduce the number of defects per wafer from 400 to fewer than 180 per cm², leading to the fabrication of larger and more complex devices, higher device yields, and reduced costs. The knowledge gained in the ATP project has led to further improvements; some wafers, for example, now have as few as 15 defects, Carter says.

Among the more important results of the ATP project was a reduction in LED price from 46 cents to 18 cents each. Materials advances unrelated to ATP also enabled 50- to 100-fold increases in brightness, but cost is key. "We could have made them 100 times brighter, but if they weren't cheap, they wouldn't sell," Carter says. "Cost is almost everything."

The low cost is key, agrees Rick Waltonsmith, product marketing manager for intelligent displays in the Optoeletronics Division of Siemens Microelectronics in California. Siemens buys blue LEDs for applications such as the dashboard lighting in cars. Apart from their low cost, the LEDs are "much better than any other technology out there for large screens," says Waltonsmith, who predicts a burgeoning market for the lightweight LEDs for "instant replay" boards in sports stadiums and signs in other venues.

To capture the display market, Cree formed a wholly owned subsidiary, Real Color Displays, Inc., which is selling 3-inch-thick modules that combine red, green, and blue LEDs to display full-motion video in more than 16 million colors. The company recently received an order for 800 modules, which will be used in a new sports arena.

Cree also has begun selling 2-inch research-grade SiC wafers. Asea Brown Boveri AB (ABB) recently ordered $2.4 million worth of SiC wafers to develop power semiconductors. These semiconductors, which carry over 1 ampere of current, are expected to enhance by up to 20 percent the efficiency of electric vehicles, electric power switching systems, and other products.

The success of the ATP project led to Cree's ongoing efforts, funded in part by the Defense Advanced Research Projects Agency, to make larger wafers and related products. In 1997, Cree demonstrated a 3-inch SiC wafer and a functional blue laser. Because it operates at a shorter wavelength than do conventional lasers, a blue laser could enable four- to eight-fold increases in the storage capacity of DVDs and improve many other commercial and military products. The blue laser would not be commercially viable in the absence of the SiC cost reductions enabled by ATP, Carter says.

In another potentially lucrative area, Cree is the sole supplier for C3, Inc., a small North Carolina start-up that plans to use near-colorless SiC crystals to make a high-quality diamond substitute. In a market survey conducted for C3, 93 percent of jewelry store personnel mistook SiC for diamond. This application is expected to spur further improvements in SiC wafers.

In a tantalizing hint of future applications, the SiC wafers were used to make a transmitter for a broadcast demonstration of HDTV in 1996. The resulting system was described as more compact, reliable, and safer than traditional tube-based technology and able to deliver three times the power of silicon at half the cost.

April 1998