Yet, JMAR Technologies Inc. last year took a big gamble and acquired Semiconductor Advanced Lithography (SAL), a Burlington, Vt.-based supplier of X-ray steppers. The San Diego-based company is pushing its X-ray steppers for the production of compound semiconductors, and in the future, for next-generation lithography (NGL) applications.

JMAR, which is betting the farm on the technology, faces some major hurdles to compete against conventional optical tools. To get a handle on the company's strategy, SBN recently interviewed Kenneth Mason, vice president of sales and marketing for the JMAR/SAL NanoLithography division in Burlington.

SBN: JMAR is aggressively pushing X-ray lithography. How can you make this technology succeed when others failed?

Mason: The X-ray lithography of the eighties, and the effort at IBM you refer to, was synchrotron-source based. The very high cost involved in starting the technology effectively locked out many potential users. The cost of entry was about $25-to-$35 million eight to ten years ago. Interestingly, in those early years, there were some very advance devices produced by X-ray. IBM produced 400-megahetz Power PC chips with X-ray lithography in 1997. Four-gigabit DRAM chips were produced by Mitsubishi in 2000.

JMAR's source technology incorporates two significant breakthroughs. First, our collimator gain is greater than ten. Second, we have a path to scalability through 'stacking' the laser sources. This collimated laser-plasma, source-based technology...maintains the robustness of X-ray lithography at much lower costs than those being projected for advanced optical and soft X-ray/extreme ultraviolet (EUV) technologies.

SBN: One of the downfalls of X-ray lithography is high mask costs. What are the mask costs for X-ray and how are you going to deal with that problem?

Mason: High mask costs are a characteristic of all advanced lithography techniques. The cost of an X-ray lithography mask is comparable to a DUV mask. EUV masks may be several times higher. To answer your question more directly in dollars, an X-ray mask can cost $15,000-to-$45,000, depending on minimum feature and pattern density--certainly similar in costs to those prevailing for OPC and phase-shift masks today.

SBN: Didn't JMAR recently sign a deal with IBM to produce X-ray masks? When will IBM deliver these masks and what will that bring to the party?

Mason: IBM's mask shop brings a tremendous amount of mask building expertise to the party. Clearly, this group leads any U.S. producer of X-ray masks and stands among the best in the world. IBM will deliver masks against the contract from April of 2002 to March of 2003.

SBN:: My guess is that some X-ray lithography is being used to produce devices for niche military applications. Will X-ray ever make it to mainstream fabs?

Mason: Yes, today the MMICs being fabricated at BAE Systems are for military applications. However, we see a clear path for X-ray lithography into mainstream usage. In its present form, X-ray lithography is rated at an output of 5 wafers per hours. Our message today, therefore, is that X-ray makes sense in lower volume environments.

Today, the compound semiconductor market produces devices, which require sub-150-nm exposures. Typically, they employ direct-write e-beam. We have a strong message in this high-resolution/low-volume segment. Our output exceeds an e-beam writer by several times. And we offer ten times the depth of focus compared to a DUV Stepper. This combination of higher throughput and greater depth of focus is a message that makes sense to the compound semiconductor manufacturer.

SBN: What advantages does X-ray have over e-beam in this area?

Mason: X-ray has a distinct throughput advantage over direct-write e-beam, commonly four to six times faster. Direct-write e-beam, while maskless, is very throughput constrained. Still worse as minimum linewidth shrinks, the direct-write e-beam throughput becomes even slower.

SBN: What are the origins of JMAR's X-ray stepper program?

Mason: The company has its roots in the Karl Suss program of the late 1980s. Karl Suss spun off SAL, then known as Semiconductor Advanced Lithography, in 1994. JMAR Technologies acquired SAL in August of 2001. The JMAR/SAL marriage is a good one. SAL brings market, applications and stepper experience; JMAR brings the laser plasma expertise.

SBN:: What's the status of your X-ray stepper program?

Mason:: We have produced sixteen steppers. There is one NanoPulsar System installed. NanoPulsar is our name for the integrated stepper and stand-alone X-ray source. This tool is at BAE Systems in Nashua, N.H., and is used to produce high-speed MMIC devices. Its source was produced by Science Research Laboratories as a proof of concept of X-ray lithography with a point source.

SBN: In the long term, JMAR is targeting next-generation lithography. How can you succeed against the likes of EUV?

Mason:: Yes, the task is huge in a certain context. We believe the fundamental characteristics of our system will allow it to be competitive. Those fundamentals are simplicity: no complex optical column. Effectiveness. Our lithographic technique allows a great depth of focus, something that is working against optical techniques. Collimated laser plasma lithography (CPL) printing also displays greater process latitude than DUV printing, resulting in manufacturing economies.

Finally the EUV solution will be expensive and will effectively lock out to all but the largest silicon manufacturers. We have a disruptive technology, which is obtainable for fraction of the cost of a EUV tool.