LOS ANGELES The 157-nm lithography development effort is being put on "the back, back burner" by the major scanner companies, researchers said here Wednesday (Jan. 28th) at the Litho Forum, organized by International Sematech to evaluate the prospects for the major forms of emerging lithography.
Though the quality of the calcium fluoride lens material is the major challenge facing 157-nm lithography, the less-well known issue of resist quality also is proving to be a difficult problem.
Karen Turnquest, working at Sematech as an assignee from Advanced Micro Devices, said progress at the 157-resist evaluation program "has slowed down tremendously since last summer." In 2002, the resist test center at Sematech obtained 27 resists for evaluation. In 2003, when the number should have been escalating, it dropped by 40 percent, with few new resists coming in after July, she said.
Turnquest said the 157-nm resists that have been checked thus far have high etch rates, compared with 248-nm resists, as much as 1.3 to 1.8 times higher. Kurt Ronse, in charge of the ongoing 157-nm development program at IMEC in Leuven, Belgium, reported here that the etch rates are nearly double, making it very difficult to control the critical etch steps.
Also, the linewidth roughness of the 157-nm resists is relatively high, and they are "significantly more sensitive to amines," Turnquest said.
Ronse said that for the past four months IMEC, the Interuniversity Microelectronics Center, has been able to work with a 157-nm full-field Micrascan lithography tool from ASML to test resists and other aspects of 157-nm lithography. The rate of progress with 157-nm lithography is roughly the same by 193-nm lithography back in the 1998-1999 time frame.
"There is no reason not to believe," Ronse said, "that if progress continues that an immersion form of 157-nm lithography could be inserted at the 32-nm node in the 2009 time frame."
Last May, Intel Corp. announced that it would not use 157-nm lithography for the 45-nm node. Although initially that resulted in a few other chip companies saying that they would stand by the 157-nm technology, that support evaporated quickly as it became clear that development of immersion 193-nm lithography, if successful, would be technically superior and less costly.
At the Litho Forum, Nikon Precision senior engineer John Wiesner told the 300 participants that Nikon now believes that the "baseline lithography is going to be ArF immersion. The early F2 effort has been largely overhauled by 193 immersion."
"The calcium fluoride material yields are only barely acceptable," Wiesner said. Nikon will continue some minimal 157-nm work, moving its 157-nm program largely to the equipment development consortia such as Selete in Japan, IMEC, and Sematech, Nikon managers said.
"We see 157 as a backup to 193 immersion in case of a 'gotcha' for 193 immersion," Wiesner said. "The definition of a 'gotcha' is something you don't know about until it arises as a 'gotcha,'" he noted.
Kazuhiro Takahashi, a Canon engineer, said Canon's evaluation of the cost of ownership (CoO) for 157-nm lithography indicates that operating costs would be "two or three times larger than for 193 immersion. We estimate the resist itself will be 2.5 times more than 193-nm resist. That means that 157 is a candidate as a backup for EUV starting at the 32-nm node."
Jos Benschop, an ASML senior manager, said 157 faces "grim" commercial prospects, even as progress is made on several technical fronts. ASML is shipping a 157-nm scanner with a 0.85 NA lens, but he said it is "less certain" that ASML would go ahead with plans for a scanner based on a 0.9 NA lens.
Benschop said future development of 157 depends on a "reliable insertion of 157 scanners by the semiconductor industry, a scenario which is not visible to us at this point. It very much depends on industry acceptance of immersion."