Turning Gold Dust Into Clean Air

BURLINGAME, CALIF. -At age 82, William Miller might be excused for kicking back, playing with grandkids or simply puttering around. He's doing none of those things. Instead, Miller is leading a start-up with two big ambitions: to clean up diesel emissions and to spread a new approach to designing a key ingredient in countless chemical reactions, namely catalysts.

"I like to be an explorer. This is an exploration," declares Miller.

And that approach--the exploration of both a new technology and of the opportunity for using that nascent tool--is at the core of genuine innovation.

Over the course of his 53-year career, Miller has been a computer scientist, chairman of a major software maker, an adviser to both venture capitalists and government leaders, and a teacher. These days, he is chairman and co-founder of Nanostellar, a four-year-old, 22-person start-up in Redwood City, Calif. Miller is too savvy to use the hyperbolic language common to most start-up founders. But he exudes quiet confidence that Nanostellar has a shot at making a genuine difference, both in reducing greenhouse gas emissions and in changing commercial chemistry.

Nanostellar is developing novel chemical catalysts that promise big improvements over existing ingredients. The company's first product: fine powders of precious metals--gold, platinum and palladium--that when used to coat a filter for a diesel truck or car can reduce its toxic emissions by as much as 40% over existing catalytic converters. At Nanostellar's heart is a computer program that predicts how different compounds will work under specified conditions. Think of it as a design tool for chemists: "We have CAD-CAM tools for mechanical engineers, computer-aided design for circuit makers," says Pankaj Dhingra, Nanostellar's chief executive. "The impact our tool could have on the world of chemicals is absolutely humongous."

How does a septuagenarian wind up starting a company at the very forefront of new technology? By never retiring in the first place.

Back in 2003, Miller was juggling a collection of assignments, including advising groups at Stanford University, where he served as Provost, and a few company boards. He had always been an overachiever: He was the last hire of Stanford's famed Fredrick Terman and one of the first five advisers to venture group the Mayfield Fund, when it started in 1969. Over the years, he also chaired Borland Software, worked as chief executive of SRI International and hobnobbed with China's leaders, including Zhang Zemin and Li Peng. Colleagues often turned to him for advice--as did fellow Stanford professor KJ Cho.

Cho and longtime collaborator Jonathan Woo had been brainstorming about how to take to market what seemed like a promising tool--a way to design chemical catalysts on computers. Other fields were already using similar tools, including biochemists who analyze protein structures via computer models. Early this decade, Cho and Woo realized that the convergence of more powerful computers, better design algorithms and rapid advances in materials science--particularly in the excruciatingly small realm of nanomaterials--was finally promising to open up catalytic chemistry to computer modeling.

Miller's advice: Build a product, not just a tool. He also tested their ambitions with a well-honed set of questions: "Does the technology work? Can it scale? Is it robust? And who wants it?"

That last question, Miller says, can be a killer. "If the answer is 'Everyone,' then my next question is: 'Name one.' "

Together with Miller, Cho and Woo scrutinized what fields could be significantly improved with better catalysts. Scrubbing the toxic emissions created by burning carbon-based fuels was high on their list. But gasoline auto makers have poured billions into refining the catalytic converters used in consumer vehicles; utility plants, swaddled in regulations, are cautious adopters of new technology.

The diesel trucking industry, however, still used an expensive approach to reducing its emission problems, namely catalytic converters that rely on the highly reactive--and highly expensive--material platinum. The computer model that Cho and Woo developed suggested intriguing alternatives, including catalysts made of nanosize particles of gold and alloys of gold, palladium and platinum.

Since then, Miller and Woo have continued to build Nanostellar into a full-fledged company. Miller took the business lead, helping raise $26 million in venture capital from investors that include 3i Technology Partners, Khosla Ventures, Monitor Ventures and others. He helped hire a chief executive for Nanostellar and other key managers. "It's more acceptable now to tell company founders they can't run a company," he says wryly. Woo is chief technology officer.

Nanostellar has developed three variants of catalysts using gold, platinum and palladium. Its work is winning technical accolades: Late last year, the start-up was picked as one of 39 "technology pioneers" for 2008 by the World Economic Forum--a honor that has in past years gone to the likes of Google and Napster.

Customers are emerging, too--even though Miller ruefully says that business has taken a bit longer to develop than he and his co-founders had originally hoped.

In conjunction with a manufacturing partner based in Pennsylvania, Nanostellar sells powders that are used to coat the honeycomb-like filters used in diesel vehicles. One European customer is using its powder to improve emissions in existing vehicles; a European car maker is planning to incorporate Nanostellar's powder in a new fleet of diesel-based cars.

"Technically, every class of engine that runs differently should have a different catalyst," Miller observes. A UPS truck engine, for instance, is different from the machine driving an 18-wheeler. How much time an engine runs "cold" changes its emissions, too, he adds.

Miller remains cheerfully upbeat about how Nanostellar's technology can improve the world. And with any luck, this won't be his last venture. "Wait until I tell you about my next idea for a start-up," he quips.

Sphere: Related Content

Column: What It Takes to Succeed

Letter From Silicon Valley

11.20.06, 6:00 AM ET

BURLINGAME, CALIF.--There is only one person who is both on the Forbes Rich list and the Computer History Museum’s Hall of Fellows--essentially the all-star list for computer science pioneers. (Hint: He probably isn’t your first guess.)

The chasm between who collects the big bucks and who invents new technology struck home for me in the past week. On the one hand, there’s YouTube, the Web’s current repository for video clips. YouTube made use of technology developed by others, notably Macromedia. The company went from startup to score ($1.65 billion in Googlebucks) in a dizzying 20 months. (Cashing out those winnings will take a bit longer; Google has tucked 12.5% of the equity due to YouTube--about $224 million--into escrow for a year until the dust kicked up by copyright lawyers scrambling to sue the company settles.)

On the other hard, there’s privately-held NeoPhotonics, based in San Jose, Calif., whose roots stretch back ten years. If business continues to ramp up, the company hopes to get off what has been a financial roller coaster and turn cash flow positive next year.

The scientists who built NeoPhotonics have invented things you’ll never see but that could change your life. Topping the list: better batteries for medical devices and optical chips for communications networks. The batteries are used in implanted defibrillators, devices that get an irregular heartbeat back in sync. The optical components sit at the end of fiber optics, repackaging data sent via light waves into electronic ones and so opening the spigot of data flowing from the Internet into local homes.

NeoPhotonics is a case study of how much persistence it takes to build a business around genuine invention.

The company started in 1996 with the promise that new nanomaterials could store energy more efficiently than conventional techniques. At the time, Sean Bi, a Chinese-born post-doctoral student in physics at the Massachusetts Institute of Technology, was working on a unique way to create powders of nanoscale particles by shining laser light through gases. Nanoparticles are almost like new elements: They have different properties than bigger versions of the same molecules. (Super tiny things have much bigger surfaces than they have volume--and this changes many properties.)

An MIT professor introduced Bi to a successful Japanese-born physicist turned businessman, Nobuyuki Kambe. Together, Bi and Kambe spent a year refining the technique. Pete Thomas, a venture capitalist then with Institutional Venture Partners, was intrigued. “They created a platform technology that could be focused to develop batteries,” he says. Thomas provided $1 million in seed funding, enough to enable them to turn lab bench experiments into a process that could churn out big batches of nanosize-particle powders. They called their company NanoGram.

Led by Tim Jenks, who joined NanoGram as chief executive, the company started developing batteries that scored high for efficiency. Yet as they experimented with the properties of their powders, Bi and Kambe realized their novel batteries released charge fast--ideal for implantable defibrillators, less so for consumer devices like cell phones.

Jenks was hunting for other applications for the powders, too. Optoelectronics was hot; NanoGram’s scientists soon realized they could create perfect coats of glass on top of silicon wafers--a boon for those making optoelectronic components. In 2002, they divided the company into three: NanoGram Devices, which kept the battery-making technology for medical devices; NeoPhotonics, which went after the optical devices market; and a third firm, NanoGram Corp., which managed the intellectual property.

As the tech bubble burst, NeoPhotonic snapped up another technology-centric company, 14-year-old Lightwave Microsystems, which was on the verge of closing shop. The technologies of the two companies meshed well, but the business liabilities did not: NeoPhotonics filed for Chapter 11 in 2003.

Most of NeoPhotonics investors did not lose heart, however. Thomas continued to sit on the board. The battery company was sold to Greatbatch for $45 million; NeoPhotonics emerged from bankruptcy. With additional financing, Jenks merged the company with a Chinese-based manufacturing operation and began building a portfolio of optoelectronics devices. Revenues were $55 million last year and have most recently exceeded $25 million per quarter. Altogether, NeoPhotonics has amassed a treasure chest of 128 U.S. patents, along with almost 100 overseas patents and few dozen more pending.

Was it worth it?

Yes, say Thomas, Jenks and Bi, in separate conversations.

“The technology created the road for NeoPhotonics,” says Thomas, now with ATA Ventures. "The proof that patents are valuable is that you're winning in the market with products that use that technology," which NeoPhotonics is, Thomas asserts.

It also led to other products such as the batteries for defibrillators and equipment for producing literally tons of nanoscale particles, he observes. “Where NeoPhotonics goes in five years time couldn’t have happened without the early nanotechnology work,” he says.

Jenks is also upbeat. Millions of new subscribers for high-speed Internet connections could mean very high double-digit growth for NeoPhotonics, he says. Although its beginnings have been tumultuous, Jenks believes the company's deep technological strength will make future growth steadier.

And Bi, who is now an executive with software maker ArcSoft says he is grateful for the chance to have developed new technology--and for the opportunity to see how it takes just the right combination of people, technology, market readiness and funding to bring an idea out of the lab and into production.

“Commercializing fundamental technology generally takes a longer time” than building on existing technology, Bi notes. “Should you wait for someone else to do this hard work?”

“This is what risk capital is about,” Thomas says. Such experiences test the mettle of entrepreneurs and inventors--but does create long-term value, he says.

And that guy who made both the Forbes list and the Computer Museum’s Hall of Fame? Intel's co-founder, Gordon Moore.

Sphere: Related Content

FORBES: Bright Ideas

FORBES MAGAZINE

A handful of young companies

are producing new ways

to harness solar energy.

The sun never shines in the basement laboratory on the University of California, Berkeley campus, where graduate students Delia Milliron and Ilan Gur spend their days. But that isn't stopping them from doing some of today's most advanced work in solar-cell technology.

In one corner of the lab Gur brews toxic ingredients to make a half-vial of crystals, each measuring a millionth or so of an inch, of cadmium selenide, a light-sensitive semiconductor. He covers a postage-stamp-size piece of glass with the nanocrystal goop and hands it over to Milliron. In a mini-vacuum-chamber she adds tiny strips of aluminum. Finally they carry the tiny chip into a black room and shine light on it:The chip spews back power--1.5 milliwatts for this fledgling solar cell, which on some bright future day could translate to 15 watts per square meter.

Even brighter days are ahead. The pursuit of an improved photovoltaic cell, the cornerstone of solar energy production, has the whiff of a gold rush about it. Since 1999 shipments of photovoltaic cells and modules have been growing an average of 30% a year, reaching $3.5 billion in 2002, according to Clean Edge, a San Francisco research group.

Manufacturers, led by Japanese companies such as Sharp, are convinced that within sevenyears traditional solar-cell technology will deliver power as cheaply and conveniently as steam turbines. The economics of solar, once derided as hippie wishful thinking, are getting pretty compelling. "When I started in the early 1970s, the going price for solar modules was about $200 per watt," recalls Arthur Rudin, director of engineering for Sharp's solar systems division in Huntington Beach, Calif. "Today the average price for solar modules is about $5 a watt."

That is, a solar cell that generates a watt of power when the sun is at its peak for four hours a day sells for about $5. Figure in night, clouds, winter, various supporting hardware and the depreciable lives of cells, and you find that sun-made electricity is still pretty expensive, between 20 cents and 50 cents per kilowatt-hour. Get the $5 cost down to $1, though, and the sun could compete with natural gas.

The best commercial solar cells today convert 15% of sunlight into electricity, but they are made of brittle, expensive materials such as silicon. Cells can be made from cheaper, more malleable plastics, but those typically now turn a mere 3% of light into electricity. "We know it's possible to do better," contends Paul Alivisatos, the chemist who oversees that lab at UC-Berkeley.

In March 2002 Alivisatos put his reputation--and his venture capitalists' money--where his mouth is, turning over much of his research to a Palo Alto firm called Nanosys. The company has amassed $75 million and is devoting much effort to figuring out how toembed nanofilaments of semiconductors in cheap, bendable plastic sheets. Nanosys' goals: 10% efficiency, $1 per watt.

Solar energy cells are typically made of crystalline silicon, the kind used in computer microprocessors, sandwiched between electrodes (see box).Silicon's orderly atomic structure allows electrons to shoot rapidly through the crystal to the electrodes. But silicon is expensive and must be processed in elaborate clean-room facilities, just like computer chips. Organic materials such as light-absorbing polymers are easier and cheaper to process, but typically fritter away 97% of the sun's energy because the electrons must navigate a tortuous path.

Solar entrepreneurs have many ways to attack the problems. In Los Gatos, Calif. a nine-person startup called Solaicx is building a furnace for making cheap crystalline silicon substrate. Pour in elemental silicon and out will come little bricks that are the foundation material for 90% of today's solar modules. "We have far less waste than the usual manufacturing process, lower labor costs--we're just attacking every factor of production," says Robert Ford, chief executive of the company.

He says that Solaicx's furnace may be able to cut the cost of making wafers by as much as 50% when the company is in full-scale production, about two years from now. He bets that Solaicx will make it possible for solar modules to generate power at 8 cents per kilowatt-hour by 2007.

Nanosys and two other new firms, Konarka and Nanosolar, are taking the radical approach of replacing chunky solar-cell modules with rolls of flexible plastic that have photovoltaic elements built in. "You don't have to invent a new factory to do this. You use coating and printing machines," says William Beckenbaugh, chief executive of Konarka. Production costs would drop by maybe 80%.

Which light-sensitive nanomaterials to use? Nanosys' solar-cell work is based on Alivisatos' tinkering with so-called nanorods in his Berkeley lab. When embedded in plastic, the rods act like roads for the electrons, shortening their route toward the electrode andproducing power more efficiently. Early batches of nanorods were as organized as pickup sticks. In a happy accident Milliron and other Berkeley postdocs made an errant batch of nanorods that looked like jacks--with four legs--rather than rods. No matter which end was up, one leg always pointed toward the electrode. Another perk:Nanorods can be "tuned" to absorb light from different parts of the spectrum just by growing them bigger or smaller. Although Nanosys is still exploring just the right shape for its nanorods, it has committed to a deal with Matsushita Electric Works to make plastic photovoltaic laminate for materials like Spanish-style roof tiles, slated for the Japanese market by late 2006.

In Lowell, Mass. two-year-old Konarka is using titanium dioxide nanocrystals coated with a thin layer of a light-sensitive dye. The dye-sensitized solar cells suck up photons even in dim light. Konarka's work is based on a decade of research by Michael Graetzel at the Ecole Polytechnique Fédérale de Lausanne. Konarka's plastic-based solar modules have already shown efficiencies of up to 6.5%, beating the scores of thin- film (or amorphous) silicon solar cells, Beckenbaugh says. (For a story on a thin-film enthusiast, see p. 86.) Konarka has teamed up with Groupe Electricité de France, ChevronTexaco and Eastman Chemical to make products by 2005. Earlier this year Aisin Seiki, in Japan, demonstrated a solar module that's more efficient than the best silicon versions and can be made for one-tenth the cost.

Nanosolar, in Palo Alto, is building "brushes" of semiconducting metal oxide nanowires that it says improve efficiency even more. Its work is based, in part, on technology developed at Sandia National Laboratories, with the help of former students from Alivisatos' laboratory. R. Martin Roscheisen, the company's founder, says Nanosolar has a working prototype and plans to raise funds to build a factory.

http://members.forbes.com/forbes/2003/1124/222.html

--##--

Sphere: Related Content