X-ray Lithography Is Just an NGL Dream
Aaron Hand, Managing Editor -- Semiconductor International, 6/1/2002
It has come to my attention that Semiconductor International
does not cover nearly enough in the way of X-ray lithography. As the editor responsible for writing about lithography issues in the semiconductor industry, I admit that we could use more about X-ray's merits and particular niche applications. On the other hand, talk of X-ray lithography's ultimate triumph as a mainstream semiconductor printing technique is fruitless.
X-ray lithography has a long history as a leading potential candidate in the debate on next-generation lithography (NGL) techniques. But it has consistently scored low in surveys of NGL candidates for years, with industry players choosing instead to pursue extreme ultraviolet (EUV) lithography and, to some extent, e-beam techniques.
At the Emerging Lithographic Technologies conference of SPIE Microlithography, the number of papers presented on EUV lithography jumped from 49 last year to 69 this year. Meanwhile, the number of X-ray lithography papers presented fell from eight to four. In a recent survey of lithographers conducted by Semiconductor International and Reed Research, 3% of respondents chose X-ray lithography as the most promising NGL technique (see "NGL: Forever Next-Generation? "). The question there is why the response rate was so high.
Yes, we can continue to let the debate rage on, considering and reconsidering the merits of a technology that ultimately didn't make the cut. In this magazine's pages a year ago, we published a piece from a professor at the Massachusetts Institute of Technology that detailed advances in Japan's proximity X-ray lithography program, and why Japan should continue to pursue this technology (see "Japan Could Dominate Industry With X-ray Lithography ," Semiconductor International, February 2001). But even Japan has largely given up on X-ray today. Continuing to assert that X-ray has a shot at the big times is beating the proverbial dead horse.
Whether X-ray's demise had anything to do with political motivations, most people in the industry say that it is just not technologically realistic. X-ray lithography could be used today for 100 nm lithography, according to Phil Ware, senior fellow, lithography, at Canon U.S.A. Inc. (Irving, Texas), but it's just not needed. And extending this 1× technology to work at 50 nm or below seems a daunting task.
Aside from the common synchrotron that few companies can afford, masks have been a huge problem, according to Peter Silverman, an Intel fellow and director of lithography capital equipment development at Intel Corp. (Santa Clara, Calif.). The masks simply are not producible, he said.
John Cossins, EUV product manager at ASML (Tempe, Ariz.), agreed, adding that his company rejected the adoption of any 1× mask technology because of difficulties with manufacturing the masks.
X-ray lithography may prove useful for MEMS or other niche devices, Silverman said, but definitely not for making semiconductors. That seems to be the general consensus. "X-ray may actually find a niche application somewhere, but for mainstream chip applications it appears to be dead as a rock," Canon's Ware said.
For now, EUV lithography is overwhelmingly the NGL frontrunner. How did it get there? It certainly didn't hurt that its name was wisely changed from soft X-ray to avoid negative connotations.
