Bell Laboratories, Lucent Technologies
This chapter features the role of chemistry in the innovation of technologies that at first glance may not appear to be chemistry intensive, for example, electronics- and photonics-related advances. While the connection of these technologies that have revolutionized our way of life to disciplines such as electrical engineering, optical engineering, and computer science is readily discerned, materials chemistry also plays a significant role in their development and can be seen to provide an enabling foundation through materials and process design and development for desired functionalities.
Using the electronics industry as just one example, it is not an exaggeration to view semiconductor manufacturing facilities as large chemical factories. Chemistry has played a role in the scaling of silicon circuits for the past 50 years, from the invention of the transistor at Bell Labs (a device measured in inches) to current devices found in computers that have transistors with features as small as 130 nanometers. To put the size scale of current device features into perspective, an E. coli bacterium cell is about 1 micron by 5 microns in size, while a human hair is approximately 100 microns in diameter. In addition to decreased feature size, chemical processing has drastically reduced the cost of fabricating integrated circuits, thereby facilitating the computer revolution.
However, research and development related to electronics and photonics requires multi-disciplinary approaches in order to ensure the development of materials and processes that are compatible with overall system needs. The enabling chemical advances that have occurred in these areas have not been made in isolation. Clearly the process of invention or knowledge creation is not one that is readily amenable to scheduling. It is simply not reasonable to expect “invention on demand” or for creation of fundamentally new insights or phenomena to take place on “schedule.” On the other hand, innovation that uses existing knowledge and inventions to create new technologies can be facilitated. The most successful facilitation occurs through creation of an environment that encourages interactions among colleagues. In today's environment it is increasingly unlikely that a single individual will have the necessary breadth of expertise to turn an innovative idea into market reality. Multidisciplinary teams need to work in concert such that the real-time exchange of ideas, issues, and solutions results in concurrent development of multifaceted technologies. Their success depends largely on the commitment of the individuals to the overall program.
One of the key facilitators of advancements in microelectronics technologies was the formation of the Semiconductor Manufacturing Technology group (SEMATECH) in the 1980s. This organization provided a noncompetitive forum for discussion and identification of future trends and industry needs. More importantly, it served to establish research goals for this critically important industry sector and provided funding for precompetitive research in relevant areas. The Semiconductor Industry Association (SIA) Roadmap (see Figure 3.1) currently defines the proposed transistor feature size and related technology needs out to the year 2014. Additionally, it identifies roadblocks requiring research investment, thus fostering the necessary research and development to achieve manufacturable solutions.
Materials scientists and chemists alone cannot direct new technologies to commercially viable applications. Process and hardware engineers, tool manufacturers, and device designers need to be an integral part of the team developing new materials for electronic applications. Recent advances in microlithographic materials design and development serve to illustrate key factors required for successful multidisciplinary technology development. The overall performance of resists used for lithography is based on the radiation response, resolution, linewidth control, defect density, etching resistance, adhesion, supply and quality assurance, shelf life, cost, and other factors. Through understanding of these requirements from a device manufacturing perspective, each of these parameters can be translated into specific molecular characteristics. For example, reducing olefinic and aromatic moieties for a material that needs to be transparent at ultraviolet (UV) wavelengths prevents the film from being opaque to incident irradiation. In the 1980s, Ito, Willson, and Frechet conceived the idea of using catalytic processes to overcome limitations in available light flux at the wafer plane during resist exposure.
The performance of the resist is controlled through the chemistries associated with each component. Using one 193-nanometer resist concept as an example, each component—the matrix resin, the dissolution inhibitor, and the photoacid generator—must be designed to be compatible with each other, but equally importantly, they must be compatible with the overall device fabrication process. Table 3.1 lists a number of materials requirements and the associated desired molecular characteristics.
In this example the matrix resin uses polymers containing alicyclic moieties, maleic anhydride, and acrylate derivatives. The alicyclic units provide for transparency and etching resistance, while acrylic acid and its derivatives can affect differential solubility. Maleic anhydride is incorporated into the polymer because it facilitates metal-ion free synthesis; metal contamination, even at minute concentrations, can degrade device performance. The dissolution inhibitor is a cholic acid derivative; it is a UV transparent, readily available steroid that occupies a large volume fraction of material, thereby enhancing the differential solubility in aqueous base of exposed and unexposed areas of the resist. The photoacid generator is a triflate diaryliodonium salt that is miscible with the resist components and generates a strong acid upon exposure to UV light while producing relatively nonvolatile byproducts upon irradiation so that the optics in the tool are not damaged. Each component was designed with broad understanding of manufacturing process requirements and hardware and device constraints: a necessary requirement for avoiding pursuit of materials chemistry concepts that would ultimately prove not to be manufacturable.
The concept of using UV wavelengths for lithography (deep-UV lithography) was introduced in the mid-1970s with the demonstration that poly(methyl methacrylate) could be imaged upon UV irradiation. Furthermore, Bowden and Chandross demonstrated imaging of poly(1-butene sulfone) upon exposure to 185-nanometer light.2 Although deep-UV lithography was not practical at this point, the work led to interest in the design and development of short-wavelength sensitive materials for lithographic applications. Ultimately, in the early 1980s, chemically amplified resist strategies were introduced, leading to the commercial introduction of 248-nanometer resists in the early 1990s. As interest in using still shorter wavelengths for lithographic images grew, research associated with the design of materials chemistries for 193-nanometer exposure was initiated. While the introduction of 193-nanometer resists was made possible by the knowledge gained in the early 1980s, serious work in materials development began in the early 1990s. Although attempts were made to shorten the development cycle, the time cycle for market introduction of a materials technology remained at roughly 10 years from initiation of direct applications-focused research. Resists using lasers at still shorter wavelengths will have even larger challenges associated with the resist materials properties, and it is likely that future timescales covering exploratory research, applications-oriented research, applied development, and market introduction will remain in the range of a decade.
Looking at another materials-intensive technology (one that is yet to be commercialized), similar trends surrounding multidisciplinary team interactions and time lines emerge. Conducting polymers have long been of interest for both the fundamental research community and the commercial sector. The first demonstration of semiconducting behavior in plastics took place in the late 1980s in Garnier's laboratory in France. These exciting results led to a surge of interest in such materials fueled by the prospect of fabricating thin, flexible, lightweight devices using low-cost printing techniques. In 1997 the first printed all-plastic transistor was demonstrated,3 and in 2000 the first large-scale complementary circuit with 864 transistors was fabricated.4
More than a decade after Garnier's initial discovery, a team of researchers from Bell Labs and E Ink Corporation envisioned a prototype plastic display that could be considered the first demonstration of electronic paper. This program was initiated in late 2000, and within a year the first “plastic paper” was demonstrated using a 256-transistor back plane fabricated on a sheet of mylar that was then laminated onto a similar sheet of an electrophoretic display material.5 This achievement was the result of a technology-focused, multidisciplinary team of scientists and engineers working together from the outset to identify not only the end goal but also each step required to reach that goal. The need to understand the interplay between device parameters, materials performance, and process technologies was implicit. Further technology-focused research will be required before “plastic electronics” will reach the commercial sector. To be successful, a cohesive multi-disciplinary team is a requirement.
Our communications infrastructure relies heavily on advanced materials chemistries. From the manufacturing processes used to fabricate optical fiber cables to molecular beam epitaxy techniques for the creation of nanoscale heterostructures that enable many optical devices, innovations in materials chemistry have played a role. An example of a recent technological achievement that relates to optical communications systems is the MEMS-based (microelectromechanical system) Lambda Router. The Lambda Router is an optical system developed at Lucent Technologies for switching narrowly focused beams of light in the core of an optical network. The idea for using micromachines in light-wave networks was conceived around 1990, and the first example was a mechanical device that deflected light. Over time, technology concepts matured and evolved into the present system that uses complex arrays of small MEMS mirrors. In a fully connected system, light can be directed from one point to any desired point in the optical system.
The connection to chemistry comes about through the materials, process, and packaging technologies used to fabricate the devices. Microelectromechanical devices are fabricated using silicon-based processing, and thus the processing, packaging, reliability, and manufacturability all depend on a chemical knowledge base. Integrated cross-disciplinary teamwork is essential to ensure understanding of all relevant systems parameters and to minimize costly late-stage design changes while helping to focus the system and chip designs toward a viable commercial product.
The implementation time line for MEMS-based mirrors is very similar to those discussed above. Using the Texas Instruments planar digital mirror device as an example, the development to product implementation cycle time was 5 years, preceded by 10 years of active technology-focused research, not to mention the very relevant silicon-based process research that preceded the development of MEMS technologies. Once again, elapsed time from research concept to market implementation was in the range of 10 to 15 years. Notably, the time frame from concept to market implementation for the Lambda Router using a three-dimensional MEMS mirror design took only 18 months. In this case the team had a clear goal with defined technology milestones and was well coupled to a customer through marketing and sales professionals who were integrated into the team. However, the aggressive development cycle for the Lambda Router would not have been possible without preceding enabling research and development efforts related to MEMS mirror arrays and other optical devices. Even these efforts were backed by the demonstration of microgears in the 1990s based on MEMS technology, which was additionally supported by 30 years of integrated circuit research and development experience.
Using the three examples described above as references, there is no question that materials chemistry plays a critical role in the development of advanced technologies. Each case has relied heavily on both fundamental and long-range technology-focused research for enabling ideas, concepts, and inventions. While it is not reasonable to expect that the creation of fundamental insights can be dictated by a predetermined schedule, once relevant inventions and enabling technological concepts are realized, the innovation process can be facilitated through the use of integrated multidisciplinary teams to rapidly capitalize on research insights. These teams need to communicate effectively, understand the value that each individual brings, trust each other, and, perhaps most importantly, understand market needs. In a similar vein, marketing, sales, development, and manufacturing organizations need to have knowledge of the latest scientific findings as well. Decreasing the time line for innovative new technologies requires more than just carrying out the requisite materials chemistry research and development; it involves identifying and strengthening the connections among all the stakeholders. The end product as well as the steps to achieving that end product must be defined with a view toward manufacturability.
Mary L. Mandich, Lucent Technologies: First, who started the Semiconductor Industry Association (SIA)? Also, is there an equivalent organization in the more mainstream chemical industry?
Elsa Reichmanis: I can't directly speak to the origins of SIA, but SEMATECH was formed in the mid-1980s in an effort to regain U.S. leadership in semiconductor technology. Funding was provided by the Department of Defense and leading semiconductor industries.
SEMATECH was started in direct response to a feeling that silicon-based technology was being taken out of the country, that Japan was largely dominating the industry. We wanted to be sure that we retained a healthy industry in the United States. Interestingly, SEMATECH has become more of an international organization as companies have become more global in nature.
I am not aware of a comparable activity in the chemical industry. There is a battery-oriented consortium that the automotive industry belongs to. I know there is a similar packaging consortium, but this is again more silicon technology based. I think that because the chemical industry is so diverse, any chemical consortium would have to be technology oriented rather than broadly chemical in scope.
Hans Thomann, ExxonMobil: Historically, Bell Labs is known for championing the individual as a team. I am wondering, in the context of the examples you gave such as plastic electronics, if there were many years of research during which ideas were created for applications of that research. Are you using innovation processes to facilitate a conversion from this past mode of research guidance to the present market-driven research, and what tools are you using to help your multifunctional teams reach those goals?
Elsa Reichmanis: That's an interesting question. I can't say that the research management at Bell Labs uses a defined process for encouraging innovation. I think that we have a dual personality. We have a very strong individual contributor component where individuals are rewarded for their accomplishments. At the same time, many of those individuals traditionally have—and I would say more so today—also participated in team-based activities. If you look at the publications coming out of Bell Labs, I think you're going to be hard pressed to find many papers with a single author.
While we are awarding the individual, at the same time we have a very collaborative environment that encourages team approaches to understanding and solving problems. We also have an open-door policy. We have wide corridors that are very long, and most people leave their office doors open. You can easily talk to your neighbor. You can find somebody with the right expertise down the hall or in another building to talk to about a problem that you can't solve yourself. The culture is one that encourages the interaction of communication.
Hans Thomann: Are you using any formal processes for innovation, such as targeted innovation sessions?
Elsa Reichmanis: We don't use formal innovation processes such as those described in the preceding presentations. We do continually examine our portfolio of activities and determine what relates to the business from the advanced development, applied research perspective and from the very long-term research perspective. We need to have a mix of both, and we have a broad continuum of activity. In reality the spectrum of fundamental research to productization requires different modes of working at different stages—there is an evolution from what could perhaps be more individual blue-sky research to a team-based, problem-solving approach. To be successful, we need flexibility.
I'm in a laboratory that has a mission for doing research that should produce applications 5 to 15 years in the future. Company executives accept that long-term research in the advanced technology arena is necessary for the company to continue to be a provider of advanced technologies in the future. Otherwise, we won't have a business.
Even in a very tough economic climate for Lucent, there is resounding support for long-range fundamental research that does not necessarily have a direct application to the business. The research will generate knowledge, and that knowledge will generate ideas for technology advancement.
David E. Nikles, University of Alabama: SEMATECH produces road maps that identify the key problems to be solved within approximately 10 years. Is that why the corporate research timescale seems to be 10 years?
Elsa Reichmanis: I don't believe so. Historically, the timescale has always been 10 to 15 years from research concept to product implementation. I don't think the 10-year SEMATECH road map is defining this cycle.
David E. Nikles: We've heard about the innovation cycle from three big companies that have established businesses. How long was that from discovery to marketplace for the polymerase chain reaction for replicating DNA? Do smaller companies have a faster pace?
Elsa Reichmanis: I have read a lot about start-ups and smaller companies having a faster pace and being the “innovators of the future.” I don't entirely buy into that. I think you can have that same rapid time to market if you have the right environment in a large company. For example, Lucent's MEMS Lambda Router took only 18 months from idea to delivery of product, and I don't think of Lucent as a small company.
If we start relying on small companies to produce advanced technologies, the industry will be hurting 10 or 15 years down the road. Small companies are very focused on the development end of technology and ensuring that they have a revenue stream. They don't have the resources to do research in broad areas. Companies need to have a balanced portfolio that includes a number of different avenues for doing fundamental research in and of itself; fundamental research as it may apply to a technology, industry, and applications; and applied research and advanced development activities. We need to have a balanced perspective of the research path.
Robert A. Beyerlein, National Institute of Standards and Technology: My question is very close to Hans Thomann's. I'd like for you to comment further on what you think are the essential elements at or near the beginning of the innovation process.
I would like to mention that in Allen Bard's Priestly Award address he expressed the opinion that new science is originated from one or two investigators working largely in isolation. At the same time, he made comments showing his skepticism about the trend he saw toward funding collaborative centers by the federal government and less emphasis on supporting the individual investigator.
You gave three powerful examples of the route from innovation to effective implementation in the marketplace and how, early on, teams were and are important. In the spirit of Allen Bard, could you comment further on what the essential elements at or near the beginning of innovation are and how these have changed?
Elsa Reichmanis: First we have to differentiate innovation from invention and discovery. I agree with Allen to some extent that invention and discovery can be done either by an individual scientist working in isolation or by groups of scientists working together.
The electronic paper idea and the work we've done in organic semiconductors have had components including work by individual scientists, but there have also been components of a multidisciplinary team approach. For example, chemists have been associated with the design and characterization of the semiconducting organic materials, developing the understanding of what functionalities are needed and other factors.
The chemists have worked very closely with the physicists and the device engineers interested in understanding how these devices work. In turn, they have all worked with process engineers to define reproducible fabrication processes. Each question, such as “How can a device be designed for better performance?” or “How can the material be processed in a more cost-effective manner?,” is answered by an individual inventor. However, all the scientists come together as a team to provide a technology objective. I don't completely agree with what Allen Bard said, and I don't disagree either. I think there needs to be a balance between individual and teamwork. I believe the academic sector should place more emphasis on teaching how to successfully work in teams.
Many exciting new discoveries will be at the interfaces between disciplines. To work at the interface we need to be able to facilitate interactions and understand how to talk to our colleagues. Physicists, biologists, and chemists don't necessarily talk the same language. We need to learn how to communicate with each other effectively in order to be able to drive innovation in the future.
Lawrence H. Dubois, SRI International: As an alumni of Bell Labs and a current resident of Silicon Valley, I can tell you that the SIA Roadmap is an incredibly powerful tool. It drives a number of different industries to focus on where they want to go and a number of different disciplines to develop new kinds of resists, etching tools, lasers, deposition chemistries, and the like. Having said that, the SIA Roadmap also stifles creativity. If you come up with an idea or a concept 5 years too early, nobody cares. It will sit on the shelf until it's time for it to be on the road map. There are clearly pluses and minuses to road maps.
Elsa Reichmanis: That's happened to Bell Labs a lot, too. We've come up with inventions too early.
Kenneth A. Pickar, California Institute of Technology: Let me just throw another stone on this one, too. Gordon Moore, who predicted that data density would double every 18 months, would be the first to tell you it was not a stroke of genius on his part. Things like the road map are a self-fulfilling prophecy. Creativity may have been stifled, but maybe if you look at how the business has expanded, it's hard to see how it could have been done any better.
Your talk is astonishing to me. I read the financial section, and your business is in meltdown. It is the worst catastrophe in the history of the telephone since Alexander Graham Bell. To hear you talk, Elsa, it's like business as usual in the research laboratory. That's just amazing. Who do you talk to out in the businesses when they're out there firing people?
Elsa Reichmanis: Life has certainly changed, but Bell Labs really benefits from the executive level's belief—meaning Patricia Russo, CEO, Lucent Technologies; Bill O'Shea, president of Bell Labs; Jeff Jaffe, president of research and advanced technologies for Lucent; and the board—in supporting long-range, fundamental research.
This belief is demonstrated by their commitment of dollars to the research organization. We have laboratories with a shorter-term focus that are more aligned with interactions with the business unit, meeting not next-generation but perhaps second-generation, third-generation needs.
Mary Mandich is involved in one of those efforts, where the bulk of her work, but not all of it, involves a shorter-term, applications-driven research program. It's not a development program but is somewhere between applied research and applications-driven research. There are also some fundamental research aspects along the way that could enable implementation of a new technology.
On the other hand, the Physical Research Lab that I work in is looking at the longer term. We are very well funded, particularly in light of the company's economic situation. We don't have enough money to do all we'd like, and we are a much smaller organization than we were 5 or 10 years ago. We have to worry about whether we can maintain adequate critical mass to do what we want to do and to provide technology to the company 5 or 10 years from now. The company appreciates our research efforts and continues to support us.
I was at a Kellogg technology management program a couple of years ago, and one of the presenters said there are two kinds of chief executive officers: those who don't really understand technology and therefore don't believe in it and are very critical of research activities and those who take technology as a matter of faith. For a technology business to exist 10 or 20 years from now, the business must have a research program. Otherwise, new technologies will not be available to the company and it won't be competitive and stay in business in the long term.
Believing in, valuing, and needing technology is a matter of faith.
Elsa Reichmanis is director of the materials research department at Bell Laboratories, Lucent Technologies, in Murray Hill, NJ. She has also been elected president of the American Chemical Society, the world's largest scientific society, for the term beginning January 2003.
M.J. Bowden and E.A. Chandross. 1975. Poly(vinyl arene sulfones) as novel positive photoresists. Journal of the Electrochemical Society 122:1370–1374.
National Academies Press (US), Washington (DC)
National Research Council (US) Chemical Sciences Roundtable. Reducing the Time from Basic Research to Innovation in the Chemical Sciences: A Workshop Report to the Chemical Sciences Roundtable. Washington (DC): National Academies Press (US); 2003. 3, The Chemistry Innovation Process: Breakthroughs for Electronics and Photonics.