Harold McMaster could change the world by inventing the first cost-efficient solar-energy panel. Or he could break his company trying

Big things can happen in unlikely places--such as Solar Cells Inc. (SCI), in the nondescript engineering building on the campus of the University of Toledo.

Last fall, after 11 years of dogged labor, SCI's scientists had what amounted to a eureka moment when they perfected a technique for depositing photovoltaic cells--semiconductors that convert sunlight to electricity --on glass to create a solar panel. The breakthrough amounted to a huge advance in manufacturing efficiency, as it allowed SCI to coat one two-foot-by-four-foot panel every 30 seconds. That left rivals eating a lot of dust. BP Solar, a subsidiary of the oil giant British Petroleum and probably SCI's closest competitor, requires six hours to coat a panel two feet by five feet.

There are other reasons that SCI's panel stands out. It seems to be at least as efficient as other panels--and possibly more so--in converting sunlight to electricity, and unlike many panels, it suffers no apparent degradation in the natural environment. Ken Zweibel, who heads the Thin Film Partnership program at the Department of Energy's National Renewable Energy Laboratory (NREL), in Golden, Colo., says there are many promising solar companies and technologies, but SCI is clearly the pick of the litter. "Right now Solar Cells is the one thin-film company that has the potential to revolutionize the [solar] industry," says Zweibel. He estimates that SCI's technology may have as much as a five-year lead on the pack.

SCI's breakthrough has coincided with the awareness that, by the turn of the century, fossil fuels must yield to sources of energy that are cleaner, more plentiful, and less politically volatile. In 1997 worldwide demand for solar energy grew a torrid 40% over the prior year. John Browne, CEO of British Petroleum, thinks global warming is a fact, not a hypothesis, and he intends to make BP Solar a $1-trillion business within 10 years. Royal Dutch Shell foresees a $250-million investment in solar energy over the next 5 years. Both companies have visited SCI on many occasions, as have a number of utilities and glass manufacturers. SCI receives letters daily from villages in the developing world inquiring about the prospect of solar electrification. "They're all now starting to come to us. We're on our way," boasts SCI president Michael J. Cicak. "It's incredible. The only people who can screw this up now is us."

Success and acclaim are causing raised expectations for the small, research-intensive company, and a screwup is not beyond imagining. After all, in 12 years SCI has consumed $35 million ($12 million of it from U.S. taxpayers) and has yet to return a dime to investors. SCI, for all its shining promise, seems to be caught in the classic small-company squeeze between the ongoing seduction of cutting-edge research and development and the hard inevitability of the marketplace. Steven Johnson, the company's director of market development, readily admits that "success for us now is as much a maturing of the organization as anything else."

If there is doubt about the company's future, some of it may be laid at the feet of SCI's dominant--and dominating--founder, Harold A. McMaster, who is invariably defined in hopeful terms such as "visionary" and "genius." McMaster, already credited with being a pioneer and a revolutionary in the glass industry, seems bent on repeating the feat in the solar industry. "Harold's a very creative man who likes to see big things happen in a short period of time," says Bob Nicholson, a process engineer and technician who worked at SCI from 1990 to 1995. He argues that McMaster is the rare inventor and technologist who, like Bill Gates, seeks "hyperleaps" that others often cannot even discern, much less essay.

But in his Gatesian quest to make the hyperleap into the 21st century and solar-energy history, McMaster may have made a fateful choice. In search of entrepreneurial immortality, his critics argue, he may have doomed his very creation, SCI, to a premature death. Eschewing opportunities that might put SCI on sound financial footing in the short term, McMaster has sought instead to rapidly scale up a technology in which the science is complex, the learning curve is steep, and the likelihood of hitting technical snags is considerable.

But that kind of risk taking bespeaks McMaster's bent and the culture of a company where there are few half measures. Alan McMaster, SCI's vice-president of advanced engineering and Harold's son, says, "Harold's vision is to do this on a large scale. We've always thought big. That's been our story all along."

But Ken Zweibel is quick to note a basic irony at the heart of the SCI story. SCI is a company whose "technology is so good it probably can't run off the rails." And yet, in the transition between "being a pioneer to building an industry that will be in the trillions--how do you get from here to there? Can the entrepreneur really survive? There are so many different ways to fail."

The Man
The SCI story begins and ends with Harold McMaster, who stands barely five feet tall, speaks in an offhand mumble, and most days heads home for a nap after lunch. But such outward traits constitute a handy disguise for McMaster's ambition and urgency. A lack of time drives him to score big. McMaster, who has nearly 60 years of business experience and 100 patents to his name, knows how to win. He's built three companies that have done cutting-edge work in the glass industry, and he has become rich in the process. But none of that will buy him years. "It's too bad I'm 81," he confides. "Five-year programs don't interest me. I want to see this happen in six months."

In Toledo, the seat of the U.S. glass industry, McMaster enjoys--literally--iconic status. There is a bust of him in the lobby of McMaster Hall, the University of Toledo physics building. (He has been a generous donor to the university.) Students ritualistically walk by and touch the nose for good luck before exams.

That sort of idolatry is light-years from the farm in northwestern Ohio where McMaster grew up, and where he discovered that husking corn in February, with "icicles running off my nose," was not his idea of a great career choice. After receiving his masters in nuclear physics from Ohio State, McMaster landed a research position at one of the country's premier glassmakers, Libbey-Owens-Ford, in Toledo, where in time he, arguably, came to know more than just about anyone about how to temper glass. Tempering involves heating and then rapidly cooling the glass. The process compresses the glass, adding tensile strength. The glass can be shaped. If broken, it shatters like rock salt, leaving no sharp edges. Not a bad advance if you happen to be a maker of auto glass, patio doors, and the like.

McMaster founded Glasstech in 1971 to commercialize his work in tempered glass. Glasstech machines currently produce 80% of all the automotive glass and 50% of all the architectural glass in the world. In 1987, McMaster sold the company for $100 million.

McMaster may have gotten rich, but that hardly dulled his drive. During the 1970s, as the country slogged through the energy crisis, he began staring hard at the sun during retreats to his second home in Arizona. He calculated that he could create a solar panel so efficient that a field of such panels arrayed over 2,000 square miles of Arizona desert--about 2% of the state's landmass--could sate the entire country's appetite for heat and light. He could solarize every power plant in the country and end the hegemony of oil.

Backed by 57 local investors in Toledo, McMaster founded Glasstech Solar in 1984, which he then retained after the 1987 sale of the parent. Glasstech Solar began working with thin-film deposition, a technique for coating glass with photovoltaic, or PV, cells. (See "Why the Buzz?" below) That meant depositing on glass a semiconductor material called amorphous silicon to create a layer of PV cells. For five years Glasstech Solar worked with amorphous silicon to little effect, running through $12 million in the process.

Suddenly, in 1990, McMaster boldly changed course. Glasstech Solar was reincarnated as Solar Cells Inc., and McMaster offered to pay back any of his investors who wanted out. Then he raised another $15 million--$2 million of which was his own--to put into the company, and started again with a relatively unknown but risky semiconductor material, cadmium telluride.

The Vision
McMaster's penchant for taking the road less traveled is reflected in the ambitious strategy he has set for SCI. He aims to use SCI's leading technology to crack open the market for solar energy by flooding that market with solar panels. That will drive down the cost of the panels and within five years will make solar energy cost-competitive with other sources of energy.

That is a huge leap given that global demand for PV cells in 1997 amounted to just 125 megawatts, a small figure compared with the 600 megawatts that one ordinary nuclear plant can easily generate. Most of the 1997 production of PV modules went to niche markets far off the grid, where cost is not as big an issue. Peter Meyers, director of R&D at SCI between 1990 and 1994, recalls how McMaster would refer to the pilot production machine the company built in 1993 as a "toy"--even though refinements have since brought its current potential to about 20 megawatts, or roughly one-sixth of last year's global solar-energy output. McMaster had far grander designs in mind. "Harold was talking about building a machine that would produce 500 megawatts a year," Meyers recalls.

McMaster's ambitious plans for SCI flow directly from his grand success at Glasstech, which, in effect, created the market for tempered glass. "People said we'd sell 5 machines worldwide," says SCI's Cicak, Glasstech's former president. "We've sold 400 in the U.S. alone." The Glasstech product, broadly defended by a host of patents, virtually monopolized the fast-growing markets it helped create.

Norman Nitschke, a cofounder of Glasstech, says SCI, similarly well-defended by patents, is set on a similar course. "Harold's aim is to make a large-production machine that will provide many times the current world demand," he says. "He expects the market to develop as a result." To realize that vision, McMaster foresees integrating SCI's panel-coating process into "float lines" in the glass industry. (A float line is the industry's equivalent of an assembly line.) That way, every glass plant could turn out windows one day and switch to producing solar panels the next--all on the same float line.

And doing that is a slam dunk if you listen to Cicak. "We know every president, every CEO, every guy in charge of the major glass companies in the world," he says. "The engineering and science on this are done. Now it's just a question of upscaling the machinery and marketing the product."

If McMaster sees the glass industry as, in effect, his joint manufacturer, he looks to the domestic utility industry--volume driven and very price sensitive--as his prime customer. Hence, McMaster says, it's imperative to ramp up production in order to drive down price and meet the huge market demand that will be inevitable. "You have to ask yourself, 'Will this technology allow you to process 500 tons of glass a day?" says McMaster. "If it can't do that, it's a waste of time to do it."

Not So Fast
But wasted time and squandered opportunity, some contend, have been part of the legacy of McMaster's big-vision strategy. Buried in the SCI story is the existence of a generation of key technologists and managers who worked hard and with success under McMaster and who have since drifted into exile, their efforts to steer the company along a different course having proved ultimately fruitless. One of them, Bob Nicholson, recalls how the tightly knit crew, starting in late 1990, transformed what was an old machine shop cluttered with school desks into an operation that turned out world-class work. "We literally shoved stuff out of the way and painted the floor," he says. "Within two years we had produced what could become the most efficient panel in the world." According to SCI, the company's ability to coat a panel--the "deposition rate"--proved to be nearly 600 times faster than any competitor's.

The leader of that focused effort was Jim Nolan, a physicist and SCI's vice-president for operations from 1990 to 1993. Precise and methodical, Nolan also had the personal touch. Nicholson recalls rushing to meet a deadline and then getting home to find a phone message from Nolan, simply thanking him for his hard work. Peter Meyers remembers how Nolan once coolly entered a crucial board meeting and claimed a technical milestone had been reached--while the engineering group was still working on it. By the time the meeting was over, the team had, in fact, succeeded.

Steve Kaake, once an electrical engineer at the company, summarizes the sentiments of several others interviewed: "Jim was the best boss I'll ever have," he says.

Nolan, in contrast to McMaster, sought to commercialize SCI's state-of-the-art technology quickly by getting panels into the field, where he knew they would earn kudos from customers. Once the panels were available, they would create both a reputation and cash flow for SCI. But manufacturing panels at this point meant challenging the SCI orthodoxy. Think small, Nolan urged, not big. "I felt Harold's course was too big a step and the probability of success was low," Nolan recalls. "I didn't want to bet the company on that approach."

By early 1993, says Nolan, SCI had the ability, on a limited scale, to produce panels that would catch the market's fancy. While McMaster was thinking in terms of 500 megawatts annually, Nolan was squeezing 200 kilowatts of real product out of the pilot production machine designed and built by his team. In the meantime, he drafted plans for a full-production machine with 10-megawatt capacity. Though small by McMaster's standards, it would produce enough PV modules to satisfy about 10% of world demand, and it would cost $10 million to build, an investment within reach for SCI at the time. Moreover, as the company gained experience on that machine its yield and uptime would increase. Nolan estimates that he could have increased the capacity of such a machine to 30 megawatts within six years, "with little additional capital expense."

Nolan was driven by his belief that most start-ups work without a safety net. "A small company gets only a few chances to make mistakes," he says. Therefore, the risks it takes should be calculated. "Let's use the technology we had to enter the market," he recalls.

Furthermore, any company burning through capital, as SCI was, has a fiduciary duty to produce a return for investors in a reasonable amount of time. SCI seemed to have no timetable, perhaps because most of SCI's investors were members of McMaster's Ohio community and trusted him implicitly. Also, most of them were relatively well off; if their SCI investment went south, it wouldn't kill them.

Nolan argued that the company should start manufacturing in limited quantity and gradually scale up. The science concerning solar thin films was complex and largely untested. The leap from the lab directly to the large 100-megawatt machine that McMaster envisioned was fraught with unforeseen technical challenges. Confirms NREL's Ken Zweibel, "The technical infrastructure is not there for PV. It's being invented as we go along."

Cost
But surmounting such challenges has been second nature to Harold McMaster throughout his career. In fact, the challenges have made life worth living. McMaster knew that putting SCI on the map depended on driving down production cost in an effort to push the technology into mainstream markets.

Solar energy is currently most viable in specialized applications. "The 'off grid' locations or developing countries are where the market is now," says Rick Yocum, SCI's former director of marketing and former president. Going after those markets, he adds, would be a way for SCI to "build credibility."

McMaster sees such markets as little more than a costly nuisance. "We've been diverted by these niche markets," he says. He thinks that the gradual approach to the market championed by Nolan would have sent the company deeper into the red. "It would have been tough to make a profit," McMaster claims. He says that by his calculations, operating losses would have totaled $9 million over five years. Making panels in limited quantities for specialty markets amounted to death by a thousand cuts. To McMaster, the way out lay in driving down the cost by manufacturing in large volume. PV module manufacturers currently produce a watt of solar energy for $3 to $4. SCI's breakthrough could bring that cost down near $1. But, says McMaster, "if you can't get your cost of production down below $1, it's not worth it." He predicts that in five years SCI will be able to produce a watt of solar energy for 60¢.

But producing at such low cost requires a higher-volume approach--which, in turn, means running through the machine a continuous ribbon of molten glass, from which solar panels would be cut one after the other. Nolan--and many of the key employees--advocated a more conservative method that involved processing one panel at a time.

Though Jim Nolan emphasizes that his differences with McMaster were strategic, not personal, he says he could never influence the chairman's mind. "Harold's thinking dominates that company, and Harold is not inclined to change his mind. I'd be surprised if he ever decided to go after the small markets before we hit the home run." Still, as an officer and an investor, Nolan felt he had an obligation to make his opinion known. He says he tried "20 different ways" to tell McMaster he was making a strategic error, finally going so far as to put his thoughts in a letter to McMaster.

Steve Kaake, the electrical engineer who worked under Nolan, recalls once trying to divine with Nolan, over a round of beers after work, what drove McMaster. "We couldn't figure it out," recalls Kaake, who labels McMaster's technological approach "10,000 times harder" than Nolan's.

Bob Nicholson, the mechanical engineer in Nolan's group, suggests a simpler explanation for McMaster's doggedness. "Time was running out on Harold," says Nicholson. You go down in history if you're the guy who made solar energy affordable, he says, " not if you're the inventor of the curved windshield."

Control
McMaster says that by staying the R&D course and holding out for the big markets, SCI is making the very progress that it needs to allow the company to get a continuous line up and running by the year 2000. "The breakthroughs we've had the last two years have put us so far ahead," he says. "Without them we didn't have anything to offer."

By the end of 1995, with five of the original seven key hires gone, SCI, like Glasstech Solar before it, was running out of cash. In January 1996, McMaster issued more stock, which he then bought himself; that boosted his stake in SCI from 22% to 67%. He specifically stipulated that the money be used to develop the high-speed-deposition process. He also brought in his faithful lieutenant from the Glasstech days, Mike Cicak, 62, to be SCI's new president.

With a background in the auto and glass industries, Cicak is from the bare-knuckles school of business. He recalls with some pride that, as an executive at American Motors, he helped set up a sting that caught workers pilfering parts from a warehouse near Boston.

"This is the next Intel," Cicak frequently announces to all within earshot. He then goes on to explain that a share of stock in SCI valued at $10,000 today will be worth $100,000 in three years--and $1 million in 10 years. That assertion, offered one day at a conference-room lunch, drew slightly embarrassed looks from SCI's scientists and engineers.

It is Cicak who gives ample voice to McMaster's vision. "I want to build a 100-megawatt line. We think the whole world will come to us because of our ease of manufacturing and low cost." But building a 100-megawatt line would cost $100 million. Where would SCI get that kind of money? Cicak says that first he would get customers to sign long-term contracts, obligating them to purchase a certain amount of production off the line; then he would borrow money. (At press time, Cicak claimed, there were three companies in as many countries negotiating to buy panels from SCI, including competitor BP Solar.)

Cicak's presence accentuates the philosophical fissure that runs down the middle of the company. Jim Nolan and some of his crew have moved on, but they have been replaced by a new generation of scientists and engineers quietly fretting about the prospect of opportunity sacrificed on the altar of the Big Vision.

Steven Johnson, who is trying to build infrastructure and a culture at SCI, worries that it will be difficult to attract semiconductor-technology talent to Toledo, which he describes as "a classic midwestern manufacturing town."

Gary Dorer, vice-president of technologies, acknowledges: "We're struggling in switching over to being a manufacturer. The researcher is never happy with the product; it's hard for us to stop pushing buttons." He says that endlessly pursuing research with the idea of finally producing the perfect product is "like Intel waiting until it gets the Pentium 10 developed before it releases a chip."

A White Knight?
By early 1998 the tentativeness inside the company must have seemed palpable to would-be customers and partners. SCI's technology seemed the talk of the industry, but no one was stepping up. In fact, ardent courtship followed by disappointment seemed the norm. According to Cicak, two years earlier AFG, a major manufacturer of architectural glass, was set to enter into a joint venture with SCI. But its Japanese parent, Asahi Glass Co., worried about the environmental impact of cadmium telluride and nixed the deal at the 11th hour; Shell's interest seems to have waned. Cicak says he wined and dined the president of Enron's solar division at the Toledo Club, but then Enron chose a competing technology.

But then, just as suddenly, Cicak says, a white knight appeared in the form of a major local utility, Detroit Edison, that looked like it meant business. Detroit told Steve Johnson that if it didn't start a joint venture with SCI, it would invest in another solar company. In mid-February, just four months after that first meeting, Detroit began showing serious interest. It wanted 60% of any joint venture and it wanted to control the technology. Cicak says, "They are talking to us about building an assembly line of 20 to 30 megawatts. But we want to build a 100-megawatt line. That's the key--we want to control our destiny."

Ten days later Detroit prevailed upon SCI to sign a standstill agreement--but only after it agreed to give SCI the opportunity for equal partnership in any joint venture.

The Detroit deal provoked a half sigh of relief. "It's been a bunch of years and people have heard a lot of talk," said Johnson in early March. "If Harold's needs are met, this makes all the sense in the world. Otherwise, it will be a long negotiation." Detroit's interest seems promising. After all, it's a major player in the industry.

But Johnson also cautions, "This is a takeover, not a joint venture--and that's a significant cultural issue for Solar Cells. The cultural memory here is only one of control. No one here can remember not running the company. This idea is only beginning to sink in."

McMaster admitted as much. "These large companies want to be in control of the technology." But he is realistic, it seems, about SCI's chances without a white knight. "Given the kind of money we need," McMaster concedes, "we can't bootstrap our way any longer."

At press time no deal had been finalized, but McMaster was hopeful. Perhaps Detroit would affirm his hyperleap, and he could prove his doubters wrong.

Coda

In a rational world, every solar company would have long since been banished to Missouri. Few industries have generated so much hype, consumed so much money, and ended up with so little to show for it. Investors, bloodied for so long, are now eager to get even. "In the next five years there will be a lot of turmoil in this market," cautions NREL's Zweibel. The breakthroughs will be matched in scope only by the write-offs.

Zweibel worries that SCI is a beautiful orphan abandoned in a hard, bottom-line-driven world. "I hope whoever buys them has their head screwed on right."

In the meantime, though, Zweibel, who has worked in thin-film solar technology for nearly 20 years, keeps telling himself, "SCI will make it because they have to make it." He imagines a can't-miss powerhouse combination made up of SCI, Detroit Edison, and BP Solar together bringing all the elements of technology, capital, and distribution to the market. He imagines the cornerstone of a major industry of the next century about to be laid. That's the ideal, which for now clashes with the company's current fix, exemplified by Harold McMaster's stubborn genius. Says one observer of the company: "If everybody bought into his vision there would be no near-term research on a small scale in this industry. He's trying to hit the ball out of the stadium when he's already hit a 450-foot home run."

Both Johnson and Zweibel believe that if SCI had commercialized the technology five years ago, it would have three or four suitors today, not one. "Even if they had just one or two megawatts of production out there, they would have absolutely taken the PV world by storm," says Zweibel. But that is not the case. As SCI, for all its continued promise, faces the prospect of an arranged marriage, time is running out for its founder and his prized company.

"Right now they're not in control," says Zweibel. "Their destiny is not in their hands."

Edward O. Welles is a senior writer at Inc.


The R&D Trap

Many small companies get seduced by their technology. They lavish time and money on research--at the expense of pushing forward to market their goods and make a profit. It seems that Solar Cells Inc. has fallen into that trap. To date, the company has spent 12 years and $35 million and has yet to make a dime of profit. The company, guided by a founder-cum-inventor who owns a majority of the shares, has opted to push the envelope on its technology to exploit markets that are potentially huge--but have yet to mature. For founders who hear the siren song of R&D, here are a few things to bear in mind:

Give up to get back. Small companies tend to zealously guard their research; they are leery of dealing with big companies, which typically seek exclusive rights to their technology. But they overlook a big, seemingly voracious company's know-how and muscle to help move an idea quickly from research to market. Ken Zweibel, manager of the Thin Film PV Partnership at the National Renewable Energy Laboratory, in Golden, Colo., says, "Small companies with good backing do the best." To him, choosing a good partner and not husbanding your research make the difference.

If a small company's technology is good enough, the company need not bankrupt itself to prove it. A company can usually get someone else to bankroll its research. SatCon Technology Inc., a developer of advanced electromechanical equipment based in Cambridge, Mass., has forged a number of alliances with large companies based on their willingness to pay. To date, SatCon, which was founded in 1985, has spent $1 on research for every $15 kicked in by potential partners. SatCon chairman Dave Eisenhaure says many small companies see partnering with much larger ones as "a penalty they have to pay. We see it as an opportunity."

Less is often more. Early-stage companies have a tendency to throw a lot of resources at a research project. But added resources often result in delay and loss of focus. Computer Associates International Inc., one of the largest software companies in the world, is well known for systematically extracting good value from its R&D. It does so by taking a lean and counterintuitive approach. Chairman and CEO Charles B. Wang typically assigns fewer programmers to a project than competitors would. If the project runs late, Wang starts subtracting programmers from it, not adding them. "It's my belief that writing programs is an art form. It's not science or engineering," he says. Wang says a late project usually means that "an artistic person is so bogged down with administering and managing the project that he or she is not doing creative work." Taking people off the project allows them to refocus on their tasks.

Checks and balances. Many small companies lack clear separations of power between board, management, and shareholders. "They don't have the necessary checks and balances," says SatCon's Eisenhaure. If board members are passive investors or friends of the founder, with little expertise in what the company does, management tends to be less accountable. When Dave Giddings became the chairman and CEO at Diametrics Medical Inc., a newly public company specializing in medical diagnostics, he joined a board that had recently shrunk in size from nine to seven people--and raised the qualifications of those seven people. "Each board member has occupied a CEO-level position on the customer or supplier side," says Giddings. "These people are not only knowledgeable about the industry, but they are also accustomed to making decisions within operating companies."

Keep things separate. Dave Giddings says that many young technology companies suffer from "an embarrassment of opportunities." They have to be disciplined enough to figure out which research projects are worth doing. Dave Eisenhaure says that although SatCon's research budget is less than $10 million, R&D has already been separated into three divisions, each of which must submit budgets and business plans routinely. SatCon distinguishes its various research initiatives from one another and institutionalizes accountability. "Otherwise," he warns, "you can end up with a real mess."

Right mission, right needs. Eisenhaure believes that early-stage companies face a simple and clear-cut choice: management must ask if it is running a lifestyle company or a company whose mission is to build shareholder value. Often overlooked is that a small entrepreneurial company inevitably reflects the needs of the entrepreneur, according to H. Irving Grousbeck, a consulting professor of management at Stanford University's Graduate School of Business. And founders frequently have very different agendas from the rank and file. Moreover, they resist change. "The key is to have an open discussion about the needs and the objectives of the people in the company," says Grousbeck. Otherwise, he says, there is no "congruence of goals," which results in a loss of direction for the company.


Why the Buzz?

One distant day, much of the world's electricity will probably be generated by solar energy. Sunlight will be the energy source of choice because it is abundant, long-lived, and nonpolluting. Trouble is, right now it's not economic. The company that can make solar energy cost competitive with other fuels will reap the bounty of a market measured in the trillions. Right now, the leader in the quest for that bonanza is Solar Cells Inc.

Two decades ago electricity generated by solar energy cost more than $100 a watt by the time it was installed and producing power, making it suitable for only the most exotic, cost-insensitive uses such as solar panels on spacecraft. The figure has since fallen to an average of $7.50 per watt (installed), and SCI's recent breakthrough will at least halve that figure. But that's not enough. For solar energy to be used on a broad basis, the number must fall to around $1 a watt. (A watt is a unit of measure of the power available at any point in time. For example, a 100-watt lightbulb requires 100 watts of power to light.)

For SCI to reach the magic $1 figure, it must improve in three areas: cost of manufacturing, conversion efficiency, and stability.

Cost of manufacturing: SCI's breakthrough last fall allows it to coat a panel with PV cells nearly 600 times faster than its rivals. While SCI's process requires depositing 4 layers of semiconductor material on a panel, most competitors need 11. Having more steps entails more capital costs, more manufacturing time, and a greater likelihood of error, leading to lower manufacturing yields.

Conversion efficiency: This measures the amount of solar energy hitting the panel that can be converted to electricity. Working with cadmium telluride, Solar Cells has already achieved a 9.1% conversion-efficiency rate with its full-size modules. After two decades of R&D, solar companies using amorphous silicon have struggled to reach 7.5%. That is not good enough. At an 8% conversion rate, cadmium telluride begins to compete with other solar technologies. SCI projects reaching 10% within four years, which would make it "a world beater" in the words of vice-president of business development Steve Johnson. Ultimately, Johnson expects SCI to raise its conversion efficiency to 12% or 13%.

Stability: This measures how fast the solar panel degrades--that is, loses its ability to convert sunlight to electricity. In effect, it's another way of projecting the failure rate in a field of panels. Amorphous-silicon panels currently lose 25% of their conversion efficiency in the first six months of use and stabilize thereafter. SCI's cadmium-telluride panels, in the field for three years, show no loss and are projected after 20 years to still have 100% of their original conversion efficiency. They seem to be more reliable.