They Said it Couldn't Be Done: Applied Materials Etches Copper

Conferences devoted to advanced interconnect schemes usually explore only one possible path for copper integration: deposition of copper into dielectric trenches, followed by chemical mechanical polishing to remove the excess copper. This "damascene" scheme rests on the assumption that a production worthy copper etch process does not exist. Most companies would avoid a messy, difficult-to-control process like CMP if they could.

The assumption is reasonable. Common etch gases like fluorine, chlorine, and oxygen react with copper, but the reaction products are non-volatile. As Yan Ye, an Applied Materials senior manager for metal etch, explained in a presentation Monday at the IEEE International Interconnect Technology conference in Burlingame, CA, the etch products "just sit there" on the wafer. Approaches that remove the etch products also attack the sidewalls of copper lines. While aluminum oxide passivation prevents sidewall attack during aluminum etch, the corresponding copper oxides do not. In fact, Ye presented micrographs in which the sidewall remained intact, but the underlying copper lines were hollowed out and spongy.

Nonetheless, Ye claimed that Applied Materials has developed an etch process that can repeatably cut 0.25 micron lines and spaces with aspect ratios greater than 3:1. The etch used a standard Applied Materials DPS Centura chamber with a special high temperature ceramic electrostatic chuck. Ye declined to discuss the process chemistry, citing patent concerns, but said the chemistry "won't be a surprise" to other etch experts.

The process uses several thousand angstroms of SiO2 as a hard mask, and causes significant erosion of the mask layer. Ye claimed that etch rates between 5,000 and 10,000 angstroms/minute are possible, depending on the geometry and barrier layers.

Without post-etch treatment, the lines sustained severe corrosion within 30 hours after etching. With treatment, the specifics of which Ye also declined to discuss, no corrosion was observed after two weeks.

Can copper etch be a production worthy technology? Ye emphasized that further work will be required to answer that question. Still, today's presentation is sure to raise the hopes of CMP critics throughout the industry.

By Katherine Derbyshire