One start-up is trying to commercialize a new "step-and-flash" technique developed at the University of Texas. Another new firm has just introduced a tool that can be used to repair photomasks as well as build structures molecule by molecule at resolutions smaller than 15 nm. A third, more-settled supplier uses hot embossing to imprint ultrafine structures on substrates measuring up to 200 mm. And yet another fledgling firm working in the photolithography realm offers maskless technology developed at the University of South Florida.

Molecular Imprints of Austin, TX, is set to ship a beta tool incorporating the company's step-and-flash imprint lithography. Called the Imprio 100, the system will be installed at a Motorola plant in Tempe, AZ, says Ron Voisin, the company's vice president of business development.

The technology now operates at a throughput of five to six wafers per hour. The company hopes to extend throughput to 30 wafers per hour. The first-generation tool operates in the 500-nm range with a layer-to-layer alignment of 50 to 100 nm required for advanced chips. In general, imprint lithography has a resolution capability of less than 100 nm, Molecular Imprints points out.

Given the exorbitant costs of its optical counterpart, imprint lithographic solutions such as step-and-flash present themselves as attractive alternatives. At the moment the technology could supplement well-entrenched optical lithography, not supplant it, agrees Norman Schumaker, CEO and president.

Standard optical lithography has "many applications where it is very effectively and very economically used. But, as you're aware, the cost of lithography as you go to smaller and smaller dimensions is growing by leaps and bounds, faster than people's ability to pay for it."

WRITE STUFF: NanoInk's Dip Pen Nanolithography uses scanning probe microscopy to build nanoscale structures. IMAGE COURTESY OF NANOINK

Schumaker insists that one of the main advantages of step-and-flash "is it gives the researchers and also small manufacturers the capability of going to very-high-resolution features without having to invest tens of millions of dollars. And they can do this by using simple electron-beam writing tools to make templates."

In its commercialization efforts, the start-up is targeting clients and applications "where they have reasonably small chip sizes so that defect density doesn't become a controlling factor," says Schumaker. Molecular Imprints is conducting defect studies. During a tour of the start-up's cleanroom facilities, Voisin points to two KLA-Tencor analytical tools.

"Defect analysis is a critical area for imprinting, and everybody wants to know about it, but imprinting has never actually been deployed as yet in a volume manufacturing application where the truth of defectivity could be revealed. So, we have some equipment here that is allowing us to do some baselining.

Voisin notes the "Catch-22" problem facing the company: no defectivity data based on volume production is available because the step-and-flash tool isn't in a volume-production environment. Nevertheless, he expects the defectivity rates "to be entirely competitive" based on his previous experience working at Ultratech Stepper.

"Let's put it this way," Voisin says, "from about 1.5 µm down to about 0.75 µm, Ultratech Stepper played right at the very leading edge, and I was an early player in that. One of the things that we had to do was get through the whole defectivity issue, and it was all raised in very much the same fashion, and for all of pretty much the exact same reasons.

"Quite frankly, in that environment we established an entirely competitive defectivity performance, and the fact is, for all the very same reasons, I'm expecting to do it again. It's just that the feature sizes, the relevant defects, are going to be very much smaller."

Schumaker points out the obvious: that Molecular Imprints is not going after memory chip manufacturing, for instance. Instead, the company will target applications in microfluidics, nanofluidics, gallium arsenide ICs, thin-film heads, and medical devices. Another promising area is nanoelectromechanical systems (NEMS). Unlike MEMS, the NEMS feature sizes are small enough to make Molecular "look like a very attractive alternative because all of the [other] alternatives wind up being very, very expensive," says Voisin.

The Imprio 100 carries a price tag of approximately $2 million. An R&D version, the Imprio 50, costs in the range of $500,000 to $550,000. It's designed for use in academic settings, Voisin says.

Another start-up, NanoInk, has introduced a nanolithography tool based on proprietary technology developed at Northwestern University. Called the Nscriptor, the system uses NanoInk's Dip Pen Nanolithography (DPN) process to build structures with almost any molecule at resolutions smaller than 15 nm and spacing as close as 5 nm. DPN was developed by Chad Mirkin, a professor in the department of chemistry and institute for nanotechnology at Northwestern.

The Nscriptor uses scanning probe microscopy and specially developed chemistry to deposit the molecules onto surfaces made of silicon, gold, glass, or other materials, according to the two-year-old firm. The technique works with almost any molecule or ink. No resists are required, and the tool is capable of precise alignment.

NanoInk is using the technology to explore a potential solution for a problem in one of the most costly aspects of lithography, says Jezz Leckenby, director of sales and marketing. "Nscriptor is the first hardware product, but part of the developmental cycle is taking the software...and the chemistry from the Dip Pen to [pursue] specific nanomanufacturing-type opportunities. One area where we saw an opportunity was in photomasks."

The company has been developing chemistries that can be placed on masks "to replace those 'nanodivots' in the chrome masks," Leckenby says. NanoInk is able to make features at the scale of next-generation masks and the 130- and 90-nm technology nodes, he asserts. In addition, the repair chemistries are available for binary and phase shift masks, and the firm is actively developing chemistries to directly etch features onto masks and wafers.

"NanoInk's mission in life is to bring the DPN process to a whole variety of industries," Leckenby boasts. The company recently bought a redundant MEMS plant in Campbell, CA, to make probe systems. Acknowledging that photolithography is "deeply entrenched" in the semiconductor industry, the executive sees DPN in small-volume high-value-added production settings such as ASIC manufacturing.

"What we're starting to see is companies in the semiconductor industry and in biotechnology and life sciences bringing the technologies together," Leckenby says. "That's where the chemistry comes in. We understand chemistry." The combination of the Dip Pen's precise placement capability and chemistry expertise would enable positioning of the growth of single-wall carbon nanotubes. Researchers at Duke University are working on just such a project, he says.

The EV Group's nanoimprinting tool, the EVG520HE, is a hot embossing system capable of nanoimprinting on substrates measuring up to 200 mm. "Our system is able to apply force in a very uniform manner," says Chad Brubaker, a process engineer in the Phoenix office of the Austria-based company. The temperature-controlled tool operates down to feature sizes between 100 and 50 nm.

One of the primary benefits is the system's ability "to do small- and large-scale features on the same substrate," Brubaker says. The biggest challenge is separating the polymer from the stamp. "Obviously, after a highly uniform impression you're going to have a lot of surface area in contact. Trying to separate that is not exactly a trivial matter." Brubaker says that nearly 15 to 20% of product inquiries concern the tool. "Another application for this technology is microfluidics because it can allow direct patterning of biocompatible or bioMEMS materials," he adds.

Intelligent Micro Patterning is another start-up with academic origins. Based in St. Petersburg, FL, the company was launched in July 2001 to market the SF-100, an exposure system that eliminates the need for photomasks. Jay Sasserath—a co-owner with David Fries, a researcher at the University of South Florida—notes that removing masks from the lithographic equation provides two immediately obvious benefits. One, it reduces processing costs, and, two, minimizes the design time required for new devices.

"The biggest bang for the buck is when you can turn a design around almost instantly," brags Sasserath, who says interest in the tool has been good. "We're doing a lot of samples, primarily with R&D customers in universities as well as industry." Applications include MEMS, multichip modules, optoelectronic devices, and bioMEMS.

The tool works with surfaces other than silicon and gallium arsenide wafers, including "nonstandard" substrates such as metals, plastics, and ceramics. At the moment the system accommodates devices with 5-µm geometries and 10-µm pitches. "Particles are not a big concern unless they're boulders in the 100-µm range," Sasserath quips. As for the future capabilities, he asserts, "We can take the thing submicron. It's not going to happen tomorrow, though, I'll say that."

Intelligent Micro Patterning also makes a wet bench that it couples with the SF-100 for electroplating and wet etching applications. The start-up recently sold a system to a Spanish technical university that is setting up a MEMS production line, Sasserath says.

Despite the short inroads made by these firms and the high costs attached to photolithography, optical technology will continue to dominate the fab landscape pending some major breaking point. Chad Brubaker recalled a joke he read recently in a magazine article. The comment was also heard at this year's SPIE microlithography conference. The jape: "The end of optical lithography is always seven years away."