His first foray into the computer chip market wasn't quite the success he'd hoped for. So now Walter Finkelstein is banking on a new strategy that takes him out of production and into design to put him in the black.
After his former company ran out of money in 1997, Finkelstein regrouped, shifted his focus and formed NanoFab Inc., a fledgling Columbia company that makes a high-tech machines to aid chip design.
Finkelstein, 59, has taken out loans and used his savings and those of his wife, Kay, to pay for his latest efforts. He has also brought on one investor who has a 10 percent stake. He hopes the company will be profitable by year's end. "I'm betting my life savings on it," he said.
Like surgeons who use lasers to correct vision or remove wrinkles, engineers can use NanoFab's focused ion beam machines, which emit a ray 1,800 times thinner than a human hair, to change a chip's characteristics.
It's like a baker trying to find the perfect cookie recipe. Imagine a machine that, instead of making batch after batch, could change the ingredients in a single cookie after it's been made. Once the best combination was found, the baker would simply use the recipe to mass-produce the new cookies.
"What my device does is change the characteristics of the memory chip or analog chip or digital chip so it improves, either by speed, increased density or whatever you want to do with the device to make your process a lot better," Finkelstein said.
His technology could benefit the semiconductor industry, because chips are produced in batches of about 150 on round discs called wafers, and all the chips on a wafer have the same characteristics. Each wafer, slightly larger than a compact disc, can cost more than $300,000 to produce. With a focused ion beam -- which adds or subtracts material from an individual chip to give it more speed, memory or power -- an engineer can change one chip at a time without having to produce a new wafer.
"It will save the semiconductor industry, or anybody who's doing this, a small fortune in both time and dollars, so they can improve their process and get [their products] out faster," said Finkelstein, who has sold the machines to several universities.
John Melngailis, a professor of electrical and computer engineering at the University of Maryland, College Park, has been using NanoFab's machine for research in conjunction with the National Security Agency. The machine allows designers to "make interesting, unique devices with unique performance," Melngailis said.
The NSA also bought a machine from NanoFab, but officials at the super secret agency won't discuss what they're doing with it, other than to say they are studying chip modification.
The machine, with a price tag of $1.5 million to $2 million, hasn't garnered any commercial sales. But Finkelstein is aiming for $12 million to $13 million in sales by 2003.
Though NanoFab's machines are designed to work with today's computers, that could change, Finkelstein said. Someday, they could have a ray so small it could be used for "nanotechnology" -- building tiny computers atom by atom.
As President Clinton mentioned in his State of the Union address in January, when he asked Congress for $3 billion in increased science and technology funding, nanotechnology could be on the verge of exploding.
"Soon," said Clinton, "researchers will bring us devices that can translate foreign languages as fast as you can talk, materials 10 times stronger than steel at a fraction of the weight, and -- this is unbelievable to me -- molecular computers the size of a teardrop with the power of today's fastest supercomputers."
Nanotechnology "will become as socially transforming as the development of running water, electricity, antibiotics and microelectronics," the president's National Science & Technology Council said in a January report.
Researchers, the report said, believe that they could use miniature computer chips in things such as pigment-changing paint that would allow fighter planes to take on a chameleon characteristic to become virtually invisible. They envision bricks made with tiny circuits that could sense weather changes and respond by altering their inner structures to be more or less permeable to air and humidity.
Scientists are also working on a microchip that would stimulate damaged spinal chords and allow paralyzed people to walk.
What all that means for NanoFab is unclear. For one thing, its machines are not capable of working on such a minute scale. For another, the 12-person company must first contend with the competition it faces in its focused ion beam business plan.
For example, FEI Co., a huge company in Oregon, recently purchased another ion beam machine maker, Micrion Corp. FEI has sales of hundreds of millions of dollars and has cornered the market.
But NanoFab, promoting a David-and-Goliath image, says its machines are more advanced and will emerge as the tool of choice in the billion-dollar focused ion beam machine market.
It isn't the first time Finkelstein has tried to convince people he has a winning technology.
In 1993, he shifted his attention from his defense-industry consulting firm FAI Inc. to Advanced Lithography Group, a nonprofit consortium that he formed with universities, private companies and defense agencies to promote a new way of making computer chips.
But even with $15.5 million in federal grants and $900,000 from Maryland's Sunny Day fund, the group ran out of money in 1997, and its promotion of ion projection lithography stagnated.
Finkelstein, who has a degree in mechanical engineering and sees himself more as an entrepreneur than an inventor, then turned back to FAI Inc. He purchased the intellectual property rights of Microbeam Inc., a focused ion beam machine manufacturer near Los Angeles that had gone out of business. He changed his company's name to NanoFab in 1998.
Finkelstein said NanoFab machines have an edge over FEI's, because his use a variety of elements in their beams -- which makes them able to complete more functions -- while his competitor uses only the element gallium. Finkelstein said his tools also have a higher voltage and are more powerful.
Ronald Opel, an analyst at H. C. Wainwright & Co. in Boston who followed Micrion, said Finkelstein might have an edge. FEI and Micrion, he said, have long been trying to find a way to use elements besides gallium in their beams.
"If [NanoFab has] different ion sources already available, then they may be ahead of Micrion and FEI," he said.
Other key aspects to NanoFab's success, Opel said, are whether it can get its beam's diameter as thin as FEI's, which is down to five nanometers (18,000 times smaller than a human hair), and whether its machines can use gases such as oxygen or hydrogen in its beams, which are more powerful -- something that has not been achieved by either company.
Finkelstein said his beam is 10 times larger than Micrion's, but with a few modifications he'll be able to match it. His company is also in the process of testing the use of gases in the ion beams.
"The answer is yes, I can do all this," Finkelstein said.
But until the machines have undergone more testing -- something he expects will happen in the coming weeks or months -- Finkelstein isn't going on a full assault of the semiconductor industry to push his wares.
"I'm out their talking with them," he said, "and they're answer is, 'Prove it to me, and we'll talk.'"