The effort initially 20 companies from across the chip spectrum, including EDA vendors, ASIC suppliers, makers of e-beam tools, photomask suppliers and others. The group is lead by Direct2Silicon Inc. (D2S) and guided by a formal steering committee that includes Advantest, CEA/Leti, e-Shuttle, Fujitsu Microelectronics and Vistec.

The semiconductor industry has long been interested in applications of direct-write e-beam technology, which removes costly photomask sets from the equation and enables patterning of circuitry directly on a wafer. As photomask costs have skyrocketed, many companies have tried to develop various direct-write technologies. E-beam, prized for its tight linewidth control, has never demonstrated the run time necessary to make it practical.

The spiraling cost of photomasks has been blamed largely for a decrease in overall chip design starts that has been occurring for several years. Firms are hesitant to invest the money—sometimes more than $1 million at leading process nodes—to secure a mask set for a prototype device without knowing how a chip will perform or, more importantly, how it will sell. As a result, critics say, chip vendors are increasingly focusing on a smaller number of devices for big market opportunities, stifling innovation. Reducing mask costs for chips would increase design starts and reduce time-to-market for a wide range of chips, according to the eBeam initiative.

Representatives of the eBeam Initiative, speaking at a launch event here Tuesday, say the work will focus on developing a specific design approach that can reduce write times. The goal of the project is to make it possible to write a wafer on a single e-beam tool in one hour. Currently, it typically takes about 24 hours, they said.

Aki Fujimura, chairman and CEO of D2S, said reducing write time by decreasing the number of e-beam "shots" needed to write a pattern to between three and five from currently 10 to 25 would be the key breakthrough needed to make e-beam direct-write more practical.