Could America Afford The Transistor Today?John Carey
Not long ago, Michael Roukes was a top physicist at Bell Communications Research Inc. (Bellcore) in Red Bank, N.J. He had a modern lab and the freedom to study his specialty: the movement of electrons in bits of semiconductor material--research he says might lead to "the smaller transistors of the future." Then, in 1991, Bellcore stopped most long-term research in physics to focus on the immediate needs of its customers, the Bell phone companies. Dozens of scientists were let go, including Roukes, who landed a tenured post at California Institute of Technology. "It took years for Bellcore to build one of the best corporate research labs," Roukes laments, "but only a few months for it to be smashed."
Bellcore's new emphasis on quick results is part of the most sweeping redirection in U.S. science since World War II. For 50 years, the powerhouse team of the federal government, universities, and corporate labs has awed the world: U.S. research universities are No.1, and Americans dominate the Nobel prizes. But the cold war's end, pressures of global competition, and budgetary stringency are forcing a rethinking of how and why the U.S. does science. "We need a new contract between science and government that responds to the challenges and opportunities of the 21st century," says Representative Rick Boucher (D-Va.), chairman of the House subcommittee on science. To many, that means refocusing America's 2.5 million scientists, 700 government labs, and hundreds of university labs on near-term goals--such as boosting competitiveness or healing social ills.
Corporate labs already are leading the charge, from AT&T to Exxon to IBM. From 1986 to 1990, industry cut its long-term research funding by 15%, estimates economist Edwin Mansfield of the University of Pennsylvania. Texas Instruments Inc., co-inventor of the integrated circuit and other breakthroughs, is now pursuing just one big basic-research project: to develop tiny, superfast integrated circuits. DuPont Co., creator of nylon and Kevlar, has trimmed its long-term research from dozens of areas to "only five," says Alexander MacLachlan, DuPont's newly retired senior vice-president for research and development.
GET RELEVANT, OR ELSE. Researchers who flee to academe fear that the trend may spread there as well. Washington, which spent $76 billion on R&D last year, funds most of the nation's $26 billion in basic research, mostly in universities (charts, page 82). Now, Congress has given the National Science Foundation a startling order: spend 60% of its $2 billion research budget on projects deemed relevant to national needs. Depending on how "relevance" is defined, the effect could be dramatic. If the NSF doesn't pursue "activities where science leads to technology, products, and jobs," says Senator Barbara A. Mikulski (D-Md.), her appropriations subcommittee may slash its budget by up to two-thirds.
The Clinton Administration, meantime, is boosting spending at other agencies on "commercially important" technologies, such as clean, fuel-efficient cars and flat-panel displays. That change and Mikulski's ultimatum have researchers fearing a squeeze on fundamental science. Last fall, Congress voted to kill the $11 billion Superconducting Supercollider (SSC), the nation's largest basic-science project. And Clinton's proposed 1995 budget calls for a 2% hike in basic research--an inflation-adjusted cut.
The stakes involved in such decisions are huge. Until World War II, basic science was a low priority in the U.S., where industries such as textiles, chemicals, and steel were built upon the discoveries of European researchers. With the free world's survival suddenly challenged, however, Washington funneled millions into science projects that led to the atom bomb, advanced radar, and new medicines. "The war effort taught us the power of adequately supported research for our comfort, our security, our prosperity," wrote Franklin D. Roosevelt's science adviser, Vannevar Bush. In his classic 1945 report, Science--The Endless Frontier, Bush argued that public fund-
ing for science would lead to new products, new industries, and higher living standards.
To a large extent, Bush was right. U.S. wartime discoveries fueled growth in the aerospace, nuclear-power, drug, and electronics industries. Over the next four decades, fundamental research produced transistors, lasers, and molecular biology--the foundations of America's world-leading computer, communications, and biotech industries. For companies to curtail such work now is bad enough, scientists say. If Washington does, too--even as Japan boosts basic research spending by double digits annually--entire industries may never be born. "If we stop pushing the frontier, the whole innovation process will eventually crumble," warns James C. Sturm, professor of electrical engineering at Princeton University.
While granting that, however, critics see a flaw in the current approach: Just creating knowledge hasn't guaranteed profits for whoever foots the bill. "Congress has noticed that the Japanese have wiped us up in the automobile and electronics industries...with our own science," says Robert J. Hermann, senior vice-president for science and development at United Technologies Corp. VCRs, flat-panel displays, and compact-disk players, for instance, are based on U.S. research that wasn't exploited at home. The rising debate over how to avoid such failures adds up to "the greatest ferment over the role of science in the past 40 years," says House Science Committee Chairman George E. Brown Jr. (D-Calif.). That's because any errors will be long-lasting. A "new paradigm will emerge," says Venky Narayanamurti, dean of engineering at the University of California at Santa Barbara. "Hopefully, we will call it correctly, but we will only know that far in the future."
Those who argue that science should be more relevant do have some good ammunition. "Over the past 40 years, the grounding between research and real problems has become weaker," says John Seely Brown, director of Xerox Corp.'s Palo Alto Research Center (PARC). For instance, discoveries in such areas as quantum physics laid the groundwork for the transistor and the laser. But new advances in high-energy physics are so arcane that even scientists don't see many payoffs. That's partly why Congress axed the SSC.
BLOCKBUSTERS. Beyond that, the path to innovation is changing in favor of manipulating existing technologies. Or so finds a study by Don E. Kash, professor of public policy at George Mason University in Fairfax, Va. Kash compared the 30 products having the highest world sales in 1970 with 1990's top 30. The 1970 list was dominated by chemicals, petroleum derivatives, metals, plastics, and drugs--many the results of lab discoveries. By 1990, however, assembled products--consumer electronics, computers, cars, planes, and telecommunications gear--led the top 30. Currently, "70% of new ideas come out of manufacturing and marketing, not R&D," Kash says. "There's still a hell of a benefit from a breakthrough. But that's not where the action is."
This realization arrives just as the cold war's end prompts Congress to scrutinize a federal R&D budget that is 60% defense-related. The inquiry will uncover a discomfiting truth: "A peacetime economy can't absorb [discoveries] nearly as rapidly as a cold war one," says Bellcore President George H. Heilmeier. "We've got this marvelous research engine that generates more technology than the system can invest in and deploy."
UNKIND CUTS. More and more profit-conscious executives agree. After several years of cutbacks, major U.S. companies spend less than 22% of R&D on long-term projects, vs. nearly 50% by their counterparts in Japan, estimates the industry-sponsored Council on Competitiveness. This year may bring the unkindest cuts of all: In a survey by the Industrial Research Institute of 253 big R&D spenders, 41% said they would reduce total R&D in 1994, vs. 20% that plan increases. Three times as many plan to cut long-term research funding as to raise it. At TI, says Senior Vice-President Pallab Chatterjee, the average payoff required of long-term research has "come down from 10 years in the mid-1980s to about 5." At some companies, horizons are much shorter: just two years at Communication Intelligence Corp., a Redwood Shores (Calif.) handwriting-recognition company, says John S. Ostrem, R&D vice-president.
Hold on! says Robert W. Lucky, vice-president for research at Bellcore: These trends mean "we're going too far in the short-term direction." Worse, argue experts such as Eastman Kodak Chief Executive George M.C. Fisher, such steps are based on a fallacy. Many failures of U.S. companies to roll out new products lie not in research, they say, but in flawed business visions. Xerox invented key elements of the personal computer only to see Apple Computer Inc. commercialize it. IBM developed high-speed RISC microprocessors, but Hewlett-Packard and Sun Microsystems made them a success. American Telephone & Telegraph Co., despite its vaunted--and expensive--Bell Labs, today is challenged by upstarts MCI and Sprint, which do little research.
Instead of examining itself, "management is taking the stick to researchers," says physicist and IBM veteran Jerry M. Woodall. Adds Ralph E. Gomory, president of the Alfred P. Sloan Foundation and former senior vice-president for science and technology at IBM: "We need to do better translating science into industry. But you don't do that by killing science."
In fact, scientists argue, corporate cutbacks are leaving two widening gaps in R&D. One is in "mezzanine" research--developing a prototype from a promising idea, such as using holograms to store vast amounts of computer data. This research, which can be basic or applied, is "high-risk, high-cost," says Linton Salmon, program manager for micro-electromechanical systems at the NSF. "That's why everyone is getting out of it."
VACUUM CLEANUP. The next gap comes below the mezzanine--in the "ground floor" research that leads to new ideas in the first place. It doesn't have to be blue-sky. Bell Labs invented the transistor in 1947, for example, while trying to find a more reliable replacement for vacuum tubes. But researchers are free to explore promising or perplexing avenues, even those far from the starting point. In this research, obstacles typically present opportunities. "A barrier may tell me more about how some piece of the world works," explains Brown of Xerox' PARC. Given industry cutbacks and the proposed earmarking of more federal funds for goal-oriented research, "it's now legitimate to ask who will do long-term, more basic research in the future," says Bellcore's Heilmeier.
Take one example. In the late 1980s, scientists at Massachusetts Institute of Technology's Lincoln Laboratory were making gallium-arsenide transistors for fast circuits when a temperature malfunction caused the material to be laden with too much arsenic. That made it useless for transistors, but Jerry Woodall at IBM's Thomas J. Watson Labs found that the arsenic had formed little lumps within the material. Not only has that uncovered a "whole new physics phenomenon," he says, but the material's unique properties are leading to such devices as ultrafast light detectors for the Information Superhighway.
Problem is, this work no longer fits Big Blue's immediate business needs, according to Woodall. So he left last July for Purdue University--where he now worries that industry and government support for this kind of project could dry up. "If nobody supports blue-sky research, 10 years from now we won't have things like my new devices," he says. Scientists across the country tell of similar fates for other ideas such as longer-lasting rubber, new types of memory chips, microscopic machines, and innovative ways to capture color images, both electronically and with chemicals.
A project at Xerox shows how shortsighted this can be. Instead of doing R&D just to improve the printheads on its copiers and laser printers, Xerox' PARC has pioneered a way of making thin transistors from a material called amorphous polysilicon. The work has led to new technologies for flat-panel displays and scanning--plus a laser printer. "From going into a problem deeply enough, we can now project entering three new market segments," says Brown. "If we were just preoccupied with printheads, we would never have seen these other opportunities."
There are signs that Washington's techno-wonks see the threat to long-term science. "We have focused on the commercialization problem, with the sense that research is just fine," says one Administration analyst. "Now it's time to focus on research." So while it pushes a $1 billion plan to develop a clean, efficient car, the White House, in memos, also has told government agencies to make "vigorous support for basic research a top priority." And in late January, it held a high-powered public meeting to discuss how to keep the U.S. ahead in basic science.
While no comprehensive vision has yet emerged, BUSINESS WEEK interviews with 75 government, industry, and scientific leaders turned up many interesting ideas. Experts advocate a balancing act involving cuts in areas where progress has slowed, such as high-energy physics, plus new investments in disciplines where rapid advances promise payoffs, such as molecular biology and information sciences. Meantime, says Phillip A. Griffiths, director of Princeton's Institute for Advanced Study, it's critical to maintain the scientific talent to power up research if once-sleepy fields are transformed by discoveries. To do that, Sloan's Gomory advocates a two-part approach: Make sure that the U.S. stays on the cutting edge in all major areas of science. Then, in those areas that promise hefty returns, boost funding for results-oriented work.
SOLOMONIC SCIENCE. Though still broad, that approach would lead to painful adjustments, not least of all for universities. Big increases in science spending led to a 41% jump from 1977 to 1989 in the number of U.S. scientists with doctorates--not all of them grade-A researchers. "One outstanding scientist can do 100 or 1,000 times more than someone who is almost as good," says Caltech President Thomas E. Everhart. In any case, this army of scientists is pursuing too many mediocre projects, adds Bruce Alberts, president of the National Academy of Sciences. The solution, experts say, is to eliminate redundant work, fund excellence, and demand more results. Such moves might produce better science while freeing money for such efforts as the Commerce Dept.'s Advanced Technology Program (box), which promotes mezzanine research.
Crafting just the right policy will be difficult and risky. The ultimate goal must be to keep basic science, mezzanine research, and product development healthy--while qtill producing short-term results. "We are talking about the wisdom of Solomon here," says Mary Lowe Good, Commerce Under Secretary for technology. But compared with charting a new R&D path for the nation, Solo- mon's task seems simple: He had only to decide which of two women was the mother of the child they both claimed.
WHERE RESEARCH IS IN RETREAT
As an unpredictable blend of serendipitous discoveries and blind alleys, science is hard to categorize. Since much fundamental new knowledge has practical applications, for instance, it's difficult to label research as basic or applied. But distinctions can be made. Here's where the U.S. is heading in various types of research, which totaled $160.8 billion in spending in 1993:
TYPE OF RESEARCH 1. FUNDAMENTAL UNDIRECTED RESEARCH 2. FUNDAMENTAL DIRECTED RESEARCH 3. MEDIUM- TO LONG-TERM RESEARCH 4. PRODUCT DEVELOPMENT
WHAT IT IS
1. Science driven by curiosity and the search for basic knowledge.
2. Long-term research exploring basic phenomena--but with a specific goal in mind. In many cases, researchers still have freedom to follow unexpected avenues.
3. Basic and applied research that tries to transform bold ideas into prototypes of products. It's expensive and high-risk, since many ideas won't pan out.
4. Turning prototypes into products and making existing products better. Involves developing improvements in manufacturing processes and meeting customer needs.
WHO DOES IT
1. Federally funded scientists and graduate students at research universities.
2. Universities backed largely by government grants, national laboratories, and such large corporate labs as Bell Laboratories or IBM's Thomas J. Watson lab.
3. Mainly company laboratories, but some universities as well.
4. Engineers, manufacturing experts, and marketing departments of company business units.
1. Growth of funding, largely from Washington, has slowed. The number of scientists has risen much faster, so that the odds of finding money are falling.
2. Significant cutbacks in industry labs since the mid--1980s and slower growth in government grants.
3. Major cuts by big players such as DuPont, Texas Instruments, Bell Labs, and Bellcore.
4. Competitiveness pressures have forced U.S. companies to become more efficient. Development projects are getting a larger share of R&D spending.
1. Congress is pressuring scientists to link more research to national needs. Researchers fear that this will undercut U.S. science and competitiveness.
2. U.S. companies may be shortchanging their future by failing to make fundamental advances in technologies that will underlie future products.
3. U.S. companies will lose out to companies in countries such as Japan that are boosting the resources they put into such research.
4. Some worry that cuts in research will slow down the flow of new ideas and eventually stunt development as well.