Mason - Here is SAL's (Suss Advanced Lithography) response to your note to me:
Dear Ms. Smallfelt:
This answer may seem unnecessarily long, but there are so many problems with Mr. Barge's letter we wanted to try to correct some misconceptions.
Light (photon) sources have been used in a process called photolithography since the beginning of computer chip production in a tool called an "exposure system." Early systems used mercury vapor lamps, then mercury/xenon was used for shorter wavelengths. More recent designs have used laser sources. Over time, the wavelength of light used has shifted through 436 nm (G-line), 365 nm (I-line) and 248 nm (DUV) which is used today. All these wavelengths are still in current use depending on the technology needed to build the device of interest. The wavelength to be used in the next generation of devices is still under debate, but the strongest contenders are 1 nm (x-ray), 193 nm, and 13 nm (EUV or long x-ray).
The wavelength of light determines to a large extent what minimum feature sizes can be printed. In order to produce patterns with 0.13 micron and below, wavelengths of 0.8 nm to 1.3 nm have been proposed. These wavelengths belong to the part of the spectrum called X-ray. X-ray covers the range roughly from 5 x 10-9 through 5 x 10-11 meters, the extreme ends of the range depending on who is doing the defining. X-ray overlaps UV on the one end, and gamma radiation on the other end of its range. X-ray is itself divided into hard X-ray and soft X-ray (You can reference any standard elementary physics text. If one chooses to ignore those standards one can call it whatever they like).
If we discuss the tools necessary to print patterns on integrated circuits we find by looking into history that the first tools used were called contact printers ( e.g., the early Cobilt exposure systems) because they exposed an entire wafer with the mask in contact with the wafer surface during each exposure event. The next tool design (e.g., the Canon PLA 500 series) placed the mask in very close proximity to the wafer, but not in "hard" contact. These could also do contact printing and were essentially modified contact printers.
Perkin-Elmer (now SVGL) pioneered the scanning projection exposure tools (their Micralign series). To improve overlay for smaller feature sizes new tools were introduced in the mid 70's called steppers (GCA was a pioneer in this field) from "step and repeat." The word stepper refers to the fact that after exposing one field the systems steps to the next field and repeats the process and so on! These early steppers were 1:1 projection systems, later tools used reduction rations of 5:1,4:1 3:1 or 10:1, depending on the manufacturer.
The more recent exposure tools are called "step and scan" systems. These tools were introduced in the 1980's by Karl Suss in their original x-ray stepper systems, and later by Perkin-Elmer (SVGL) in the Micrascan series. Both Nikon and Canon are said to be working on step and scan systems. The step and scan systems expose one field by scanning the wafer and mask through a light path and then stepping to the next field. These can be viewed as a combination of the scanning and the stepping exposure systems although that is overly simplistic. Step and scan systems have generically been called steppers or scanners, either term is partially correct but potentially confusing.
By any definition, the SAL exposure system IS a stepper and is designed to be able use X-rays of various wavelengths. As far as the SAL equipment being a stepper, I think the definition used by the writer is a bit narrow. The older SAL systems were 1:1 step and scan systems (the first step and scan system in the industry, by the way). The current SAL systems should properly be called "wafer step/beamline scan" in that they do not scan the wafer and mask, but scans the x-ray beamline instead. Mirrors, commonly used in today's x-ray steppers, cut off the hard x-ray light on purpose! However, very early x-ray steppers also used the "hard" part of the x-ray spectrum and systems to do that can be built today if the need arises.
I have not seen a definition that says an expose system has to be a reduction stepper to be called a stepper. The industry uses the definition that a stepper is a system that exposes a single field at a time in a stepwise manner, e.g. one after the other instead of all at once. There are in fact two 1:1 stepper manufacturers in the US: SAL (x-ray) and Ultratech (optical).
All steppers use a mask, it's just that the position of the mask to wafer is relatively large in most systems as compared to SAL, and those systems must have intervening mirrors and/or lenses to get the image to the wafer.
SAL's 14 systems have been primarily used in R&D, but have also been used for limited production purposes. 1 nm X-ray and e-beam are the only technologies "beyond 248 nm" that have actually been used to make working devices.
SAL is currently building its fourth generation system. Other vendors in the x-ray stepper field have built a single generation. SAL is building its 15th system, others have built a maximum of 3 each, and 8 total (all vendors combined). SAL has sold every stepper it has built. Others have given their steppers away at no or low cost. The first stepper in the SAL lineage was delivered in 1984.
JMAR has not built a single x-ray stepper. They do have the old Hampshire Instruments steppers in their stock, but these can hardly be called usable by today's exacting lithography standards.
And a couple of more points: 1) Steppers "expose" a photoresist film, they are not used to "etch" anything. Etch is a process downstream of the expose process. 2) 50 Angstrom (5 nm) is not used in any current exposure system.
"---ahead of JMAR in time but behind in money."-- We don't understand this statement so do not care to comment.
Ted Bettes
VP Marketing
Klaus Simon
Applications Scientist
Bob Selzer
VP Technology |
