Chris Anderson

LOS ALAMOS, NM—Scientists working on a new method to remove photoresist at the Los Alamos National Laboratory (LANL), have developed a process using super critical CO₂ that could significantly alter the need for the rinse and dry activities currently employed in silicon wafer production cleanrooms.

Called SCORR (for Super Critical CO₂ Resist Removal), the technique could be the first to meet the industry's need to move from “wet” cleaning of wafers to a “dry” process.

It also promises to greatly reduce a wafer fab's need for huge amounts of purified water as well as its output of hazardous waste.

A Scanning Electron Microscope image of an aluminum metallized photoresist-coated silicon wafer following SCORR treatment. The photoresist is clearly removed from even the deepest recesses of the aluminum structures and those structures appear to have suffered no ill effects from the process.Click here to enlarge image

This is especially important in the Desert Southwest, where many companies have built their plants “because the land is so cheap,” says Craig Taylor, a research scientist from LANL who worked on the project. “Reducing the amount of water they need and the costs associated with the purification of the water is significant.” An average wafer fab will consume as many as three million gallons of water every day it is in production.

Research began on the project about three years ago, says Taylor. It was funded, in part, by a grant from Hewlett Packard. The SCORR process uses CO₂ at high pressure to bring it to a “supercritical” state where it exists as a very dense gas that behaves like a liquid. In the supercritical state, the CO₂ flows very easily—like a liquid—into some of the very smallest spaces of an integrated circuit, allowing it to whisk away particles and clean the wafer. When the pressure in the system is eased, the CO₂ returns to a gas and leaves the wafer dry.

“Using CO₂ completely removes the wet chemistry involved in cleaning,” says Gumilla Jacobson, a research scientist with LANL. “And that is important as geometries in the industry move below 0.18 micron.”

The implications of this technology are huge for the industry, Taylor says. “The roadmap for the industry says that geometries will be down in the 0.07 range or 70 nanometers by 2005,” he says. “That, in theory, would dictate that in order for the industry to continue its incredible progress it would need to move to a dry process, because today's processes won't be effective at geometries that small.”

The new process is just now being alpha tested by IBM, Jacobson notes. In February, GT Equipment Technologies' subsidiary SC Fluids (Nashua, NH) announced an agreement with ATMI (Danbury, CT) to further develop the supercritical fluid chemistry and manufacturing processes for the technology.

“This joint development agreement matches ATMI's semiconductor materials expertise and experience with SC Fluids' supercritical fluid creativity,” says Gene Banucci, CEO of ATMI. “Together, we believe we can introduce a new, more effective, environmentally safe and inexpensive method of wafer cleaning to the semiconductor industry.''