For archive reference only . . . DO NOT LINK TO! Could not get permission from Scientific American to place on website.

http://www.sciam.com/article.cfm?colID=6&articleID=000E6185-09E0-1E40-89E0809EC588EEDF

March 2003 issue of Scientific American

The Relentless Storm

Bell Labs weathers the worst crisis of its 78-year history

By Gary Stix

For years, David Bishop has served as a standard-bearer for the postdivestiture Bell Labs. Trained as a condensed-matter physicist, Bishop demonstrated how someone who spent the formative years of his career doing high-temperature superconductivity experiments at one of the nation's top industrial laboratories could make the transition to overseeing early-stage product development. In the mid-1990s, as the emphasis on market-oriented research was growing, Bishop managed a group that fabricated microelectromechanical systems (MEMS), which contain tiny mirrors that can change the direction of optical signals. The initial research on MEMS resulted in his heading a team of about 100 people that built the LambdaRouter: a switch that could take a wavelength from one optical fiber and route it to hundreds of other pathways in a network.

The product was a showpiece of innovation at the laboratories. But in the summer of 2002, as the depression in the telecommunications sector reduced demand dramatically for new long-haul optical pipes, the LambdaRouter was pulled off the market. Not much interest lingered in a switch equipped to handle 10 terabits (trillions of bits) of switching capacity. Speaking of this experience, Bishop invokes the perfect storm, which, along with the nuclear winter, is constantly repeated as a metaphor for the telecommunications industry's financial implosion of the past two years or so. "Never before in the history of the company has its survival been so actively discussed," Bishop laments.

From the moment of the AT&T divestiture in 1984, questions arose about whether the unparalleled mix of scientists and engineers that produced the transistor, the laser and the fractional quantum Hall effect could survive outside the shelter of a monopoly. The push for market relevance at Bell Labs began just a few years afterward and has continued to emerge with the morphing of corporate parenthood from AT&T to Lucent, which later cast off its microelectronic, fiber and business-networking divisions.

Through spin-offs, layoffs and attrition, Bell Labs Research--the locus of the company's basic science investigations--has diminished from 1,200 employees in 1997 to about 500 today. A three-year-old Bell Labs Research facility in Silicon Valley was shuttered in 2001. The umbrella organization--Bell Labs, which includes the development side of Lucent's business--has shrunk from 24,000 in 1999 to 10,000 today. Overall R&D spending has dropped from $3.54 billion in the company's 1999 fiscal year to $2.31 billion in fiscal 2002, although as a percentage of dwindling company revenues it has actually increased.

The current crisis, exacerbated by numerous missteps by Lucent upper management, is the worst since the laboratories were founded in 1925. Some outsiders question whether basic research at Bell Labs will survive, the rationale for its existence having been frittered away over time; for instance, the spin-off in 2001 of Lucent's microelectronics division into Agere Systems undercut some of the justification for maintaining a physical sciences group, a linchpin of the research division. "Bell Labs Research is currently misaligned with Lucent's future, so ultimately it's going to be disassembled," says Greg Blonder, a venture capitalist who spent about 15 years at Bell Labs.

The physicists, materials scientists, chemists, mathematicians, engineers and even some biologists who are members of the core research team reject that argument, contending that the organization has a new role to play in staging a turnaround. In the past few years, many of these scientists have begun to work more closely with product developers than at any time before in the labs' history. Laboratory managers battled to alter the ivory tower mind-set of basic researchers beginning in the early 1990s. But officials assert that collaborations between Bell Labs and the business units have never been undertaken in such a systematic manner as they are now.

For his part, Bishop has managed smaller projects since the LambdaRouter was put on hold, including development of automated methods for assembling optical components. Lucent is also attempting to market its intellectual property more broadly. Government agencies and Ford Motor Company, among others, are evaluating quantum cascade lasers, designer light emitters invented at Bell Labs, as chemical sensors. World-class chemist Elsa Reichmanis worked at Bell Labs for about 15 years developing chemicals for semiconductor manufacturing, but this expertise was no longer needed after the Agere spin-off. She now leads a team that is lending know-how, along with Lucent patents, to DuPont and Sarnoff Corporation to help create organic light-emitting diode displays.

Basic scientific investigations have not disappeared either, as a greater focus on applied research has emerged. "We're still in the damn good science business," Bishop says. The emphasis on the practical sometimes works backward from application to science--scheduling algorithms for wireless networks have helped address nettlesome theoretical questions, for example. Research by Bishop and his colleagues on MEMS went into making a type of sensor that measures a quantum-mechanical effect called the Casimir force. Two scientists from unrelated disciplines can still strike up a collaboration over cafeteria hamburgers or sushi and begin work on a project the same afternoon, a difficult proposition at universities, where the need to seek grant money constrains such impromptu alliances. This atmosphere prevails despite a recent scandal that led to the firing of physicist J. Hendrik Schön over misrepresented data about organic electronics and high-temperature superconductivity.

Bell Labs's continued existence obviously depends on its parent's survival. "I think what's critical for Lucent is to show better success in commercializing R&D, whether that's done by Bell Labs or wherever," observes Nikos Theodosopoulos, a financial analyst with UBS Warburg who holds stock in Lucent. Too often Bell Labs inventions--from the Unix operating system to advanced chipmaking techniques--were ones that ultimately furnished as much or even more benefit to other companies as they did to AT&T and its offspring.

For the most part, other companies have eschewed de novo research in favor of different models--for instance, buying smaller companies or tapping research from national laboratories or universities. But Jeffrey M. Jaffe, president of Bell Labs Research and Advanced Technologies, defends Lucent's approach. "Developing technology in house is more efficient than making acquisitions," he says. "Companies pay premiums for acquisitions--and at times have difficulty integrating them."

Even if Jaffe is right--and other research leaders might disagree with his assessment--the monopoly-era notion that research should originate in the organization that ultimately brings it to market has changed unalterably. The demands of commercial research require a heterogeneous mix of collaborations extending far beyond any single company. The danger, however, is that without the critical mass of scientists engaged in undirected pursuits, pathbreaking telecommunications technologies will not emerge. "The problem with not doing research is that you never know what you're going to lose. You never know what you might have had that would have changed things in some way," says Robert Lucky, a former research executive at both Bell Labs and one of the AT&T progeny, Bell Communications Research (later Telcordia). The National Research Council has recruited Lucky to head a study group this year to determine whether the U.S. research base in telecommunications is being eroded. When the participants begin examining the merits of new research models, one thing is certain: Bell Labs and its more than 40,000 inventions will serve as a frame of reference against which all alternatives will be compared.