Reducing the Time from Basic Research to Innovation in the Chemical Sciences: A Workshop Report to the Chemical Sciences Roundtable (2003)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

OVERVIEW OF TRENDS IN INNOVATION IN THE CHEMICAL INDUSTRY 7 1 Overview of Trends in Innovation in the Chemical Industry Richard M. Gross1 The Dow Chemical Company I would like to begin my remarks with an industrial perspective on the critical role of innovation. Everyone would agree with Peter Drucker’s statement: “Innovation is the fuel of corporate longevity. It endows resources with a new capacity to create wealth.” Although there are many forms of innovation in the business world, ranging from new business models to technological innovations, for technological innovations it is important to use the working definition of innovation developed by Joseph Shumpeter as an integral part of his economic model early in the 20th century: “Innovation is the first commercial use of new technology.” One key industrial perspective is that innovation differs from invention. There are many inventions still sitting on the shelf that are not creating value for shareholders, stakeholders, or society. It is with this perspective that I share my thoughts with you. When thinking about the workshop organizer’s request to give an industrial overview and to cover the trends in innovation—barriers, key success factors, and the like—I came to several conclusions. First, I wanted to help set the stage for the talks and discussions later today and tomorrow, while at the same time I wanted to be complimentary to the following talks without needless duplication. To do this, I decided to use both industry information and specific information from the Dow Chemical Company to illustrate what I see happening in the broader chemical industry. I’ll start with some broad trend data on chemical patents and, thus, on innovation in the chemical sciences. Next, I’ve selected three macrotrends occurring in the chemical industry to speak about, and I’ll finish with three critical innovation success factors that I call the three P’s—people, processes, and partnerships. My first source of information about innovation is the Council for Chemical Research (CCR) study completed in the year 2000. There were two main segments of this CCR study. First, a bibliometric study that looked at the strength of the chemical sciences in the United States was undertaken by Fran Narin and his colleagues at CHI Research. Second, Baruch Lev, of the Stern School of Business at 1R. M. (Rick) Gross is corporate vice president of research and development for the Dow Chemical Company. In this capacity he serves on Dow’s Corporate Operating Board, Human Resources Committee, Retirement Board, and Corporate Contributions Committee. Gross was a 1996 recipient of the Dow Genesis Award for Excellence in People Development. 7 

8 REDUCING THE TIME FROM BASIC RESEARCH TO INNOVATION IN THE CHEMICAL SCIENCES 12 Slower Chemicals 10 TCT (Years) Life Sci 8 Electronics Faster Info Tech 6 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 Issue Year FIGURE 1.1 The technology cycle time (TCT) for U.S. chemical patents is flat, indicating that the industry’s rate of innovation has decreased. Courtesy of the Council on Chemical Research, Copyright 2000. New York University, performed an econometric study on the return on the chemical industry’s R&D investment. Narin’s work was based on data from the U.S. patent database, which offers a global view by including the origin of the inventors. It is also dependent on several indices that CHI has developed to look at the impact of patents, the technology cycle time, and the tie to basic science. During the 1980s, the technology cycle time, a measure of the age of the earlier patents cited in a current patent, slowed significantly in the chemical sciences. During the 1990s, it remained flat—a clear indication of the slowdown in the speed at which the industry is innovating (see Figure 1.1). At the same time, the current impact index of U.S. technology—as measured by citations of U.S. patents by other patents—has strengthened in all areas, including the chemical sciences. A closer look at the data for chemical sciences shows that the U.S. impact is increasing, while Germany’s is flat and Japan’s is decreasing significantly. Another index analyzed by CHI was the science linkage, the citation of science publications on patents as opposed to the citation of other patents. The science linkage, the tie to more basic science, is increasing for the chemical sciences over the 15-year study period and is only outpaced by the life sciences. The science linkage data for the chemical sciences again show the United States significantly outpacing Germany and Japan. These trends help frame the importance of the chemical sciences in society and the comparative strength of U.S.-based chemical science innovation, as well as the opportunity in front of us. The three macrotrends I selected for today’s discussion are the impact of high-throughput research, global organizations, and market-driven research on the speed of innovation. 

OVERVIEW OF TRENDS IN INNOVATION IN THE CHEMICAL INDUSTRY 9 The rapid growth in new high-throughput2 research tools yields both great benefits and significant pitfalls if not utilized correctly. It is imperative that everyone involved recognizes one important fact. It is simply said: “Since you can go so fast, you better be sure you are going in the right direction.” It is also imperative that your objectives and goals are well defined at each level and, most importantly, are understood by everyone. This is where the critical innovation success factors—people and processes— play a key role. People and processes are essential elements of creating alignment throughout an organization and are key to reducing innovation cycle time. High-throughput research involves how the research is done, not what research is done, and it clearly has the potential to impact the productivity of R&D. Dow became involved in high-throughput research in the late 1990s by partnering with Symyx, a company whose focus is the development of high-speed combinatorial technologies for the discovery of new materials, and bringing Symyx’s technology and expertise inside Dow’s large corporate R&D organization. One of our goals is to continue to leverage external high-throughput research expertise where appropriate, and we are building a large suite of multidisciplinary tools. The promise of high-throughput research is widely known, and Dow finds it to be a reality. For instance, in the case of a polyolefin catalyst process optimization, over 1,000 high-throughput experiments were run in 6 weeks. There were eight structurally diverse hits, the total time from the first designed experiment to pilot plant runs was less than 5 months, and the cost of the catalyst package was reduced by greater than 75 percent. Dow has additional examples illustrating a 10-fold decrease in cycle time, 3- to 4-fold decrease in personnel costs, and a significant reduction in the scale of reactants used and waste generated. The power is there, but prepared minds need to be thoughtful when setting the research direction before they begin. Without planning, much data can be generated without any knowledge gain. The second macrotrend of the industry is its move toward global organizations. A global organization is very different from a global company: the term “global company” denotes a location, whereas “global organization” defines a work methodology. In today’s world of specialization, there is a premium paid for being first in the marketplace. To be first in the world with a significant innovation requires global teamwork. Usually the team has a formal structure, but good teamwork among colleagues is just as effective. In fact, good teamwork and collaboration often equate to the ability to communicate effectively. The ability to utilize all of the advanced information technology and communications capability is necessary but not sufficient. Of paramount importance are the ability and willingness to share information freely. Those who do this well will benefit tremendously. The Wisdom of Teams: Creating the High-Performance Organization by Katenbach and Smith3 discusses the probability of collaborations as a function of distance. It contains both good and bad news. Unfortunately, once you move past the office or laboratory next door, the rate of collaboration frequency drops off rapidly in the first 90 feet and is only 20 percent of the “next door” collaboration rate. The good news is that past 90 feet the frequency rate changes very little (see Figure 1.2). Although this dataset stops at just over 1 mile, personal experience indicates that with today’s information technology and communication tools, collaborations over 1 mile and 1,000 miles are very similar. Low levels of collaboration at a distance are a real barrier to rapid innovation and represent a real opportunity for those who can find avenues for improvement. 2 A high-throughput research tool is any chemical or biological tool set that allows rapid parallel testing of multiple system parameters in a systematic approach. In this instance the author is referring to experimental systems, but these tools can also include computational approaches. 3J. R. Katzenbach and D. K. Smith. 1993. The Wisdom of Teams: Creating the High-Performance Organization. New York: Harper Business. 

10 REDUCING THE TIME FROM BASIC RESEARCH TO INNOVATION IN THE CHEMICAL SCIENCES Percentage with at least 1 collaboration/week 30 25 Colleagues separated by 90 feet 20 are no more likely to interact than colleagues separated by several miles 15 10 5 0 next door 30 feet 90 feet > 1 Mile Colleagues' Locations FIGURE 1.2 Although the collaboration rate of employees decreases rapidly over short distances, the rate remains nearly constant for all distances over 90 feet. At Dow we have standardized our workstations globally. I can go to any of the 50,000 workstations around the world and immediately get to my personalized desktop. The use of NetMeeting, remote network control of experiments, and the sharing of complex spectra and other data globally are standard at Dow, as they are across the chemical industry. All of this has had a large impact on the rate of collaboration. The last macroindustry trend that will be covered is the increase in market-driven research. Sir Henry Tizzard, a physicist and scientific advisor to Winston Churchill during the war, recognized early on the importance of working on radar technology. This was a remarkable observation that is as important now as it was then. Tizzard said: “The secret of science is to ask the right questions, and it is the choice of the problem more than anything else that makes the man of genius in the scientific world.” To paraphrase Tizzard, the secret of innovation in the chemical industry is to ask the right questions, and it is the choice of the right market opportunity more than anything else that drives the speed of innovation. Identifying the unmet or latent needs in the marketplace and then bringing the full power of basic or fundamental research to bear on the specific opportunity delivers results. This is not a description of what some people might think of as applications R&D, which is taking an existing product and tailoring it for a specific application. For instance, taking an existing latex formulation and reformulating it for a paper coating opportunity in Europe is an applications R&D activity. Rather, this is a description of the identification of a significant unmet need that requires new materials or a new system to meet the need. Contrary to what some believe, this does not demand a less fundamental approach. In fact, to a large degree, the profitability and sustainability of a company’s market position will come from intellectual property and the protection it provides based on fundamentals and on new knowledge derived from basic research. The market-driven research trend will be illustrated by using some examples from Dow. Although I will move through some detailed information rather quickly, I’ll do that in order to paint a larger picture, which is most important. 

OVERVIEW OF TRENDS IN INNOVATION IN THE CHEMICAL INDUSTRY 11 In the mid-1990s Dow decided to grow its presence in the advanced electronic materials business segment. In an effort to better understand the semiconductor market, one of Dow’s top scientists spent 6 months in the marketplace developing the knowledge and understanding required to identify business opportunities where Dow’s technical strengths could be leveraged to create a sustainable market position. There is no doubt that everyone is familiar with Moore’s law, the doubling of data density per integrated circuit every 2 years. The performance of integrated circuit devices, historically limited by the characteristics of the transistors, is today limited by the electrical characteristics of the interconnect. The needed improvements in the interconnect performance are achieved with copper and a reduction in the insulator dielectric4 constant due to the associated reduction in the interconnect capacitance, the cross-talk, and the power consumption. The scarcity of efficacious insulation candidates prompted the Semiconductor Industry Association to identify the criticality of low-k dielectric material development. Thus, in June 1995, Dow made a business commitment to invent a new material specifically tailored for the interconnect application. Specific performance targets were defined based on interactions with the industry, experience gained through Dow’s earlier benzocyclobutene-based systems, finite element analysis of the anticipated interconnect structures, and principles of material sciences. Molecular modeling was used to predict the dependencies of dielectric constant, mechanical toughness, and thermal stability on the polymer repeating unit structure and cross-link density. The computational chemists worked diligently long before work was done in the laboratory. The computational output was used to focus the targets of the synthesis activities. A synthesis team composed of experts leveraged from throughout the company produced samples from several chemistry families. SiLK resin is a solution of low molecular weight, aromatic, thermosetting polymer. The polymer’s molecular weight and solution concentration were tuned to enable precise and convenient deposition by spin coating, a technique universally used by the industry for the deposition of photoresist materials. After deposition on a wafer, the polymer is thermally cured to an insoluble film that has a high glass transition temperature. The polymer has good mechanical properties at process temperatures, which is required for the application, and it is also resistant to process chemicals. The most important aspect of this project was the time line. In mid-1996 the specific polymer composition was selected, and in April 1997 Dow publicly announced what became known as SiLK Semiconductor Dielectric. In April 2000, IBM reported the complete integration of the SiLK dielectric and copper wiring and announced its intent to commercially fabricate integrated circuits using SiLK resin. All of the critical innovation success factors were important in driving this rapid innovation time line. A vast array of external partnerships ranging from universities and institutes around the world to fabrication equipment suppliers and customers were involved. Without the “SiLK network,” the project would not have been completed in such a rapid time frame. The next generation of ultra low dielectric constant material will be a porous SiLK structure. This work is currently being done in partnership with IBM and was started under the National Institute of Standards and Technology Advanced Technology Program. The approach was to template less than 10- nanometer closed pores in the SiLK thermoset matrix. This allows the porous structure to be compatible with the SILK spin-on equipment already owned by integrated circuit fabricators, thus extending the SiLK dielectric through many generations of integrated circuits. 4 The dielectric constant, k, is a measure of the ability of a material to conduct electrons. A low-dielectric constant, or low- k, material is an important part of any electronic circuit because it is used to insulate the copper pathways of the circuit and thereby increases the performance of the device. 

12 REDUCING THE TIME FROM BASIC RESEARCH TO INNOVATION IN THE CHEMICAL SCIENCES The results to date are spectacular. Today we can routinely achieve our goal of closed pores at less than 10 nanometers. To our knowledge, this is a world first in thermoset resins. Another example that illustrates market-driven research tied to basic research is in the area of polymer light emitting diodes (PLEDs). One of the key attributes of PLEDs is their simple structure compared to that of liquid crystal displays. The chemistry is simple, versatile, and scalable. In short, the chemistry is elegant. Most importantly, the chemistry is tunable across the entire color spectrum. PLEDs were identified as a significant market need by industry leaders. We believed that Dow’s expertise could be leveraged to create new knowledge and thus to provide technology options with a proprietary position. Again, this example illustrates the importance of early partnerships. In this case the partnerships were with Richard Friend at Cambridge University and with CD Tech, Inc., also in Cambridge. This type of innovative partnership occurs across the spectrum of business sectors. Dow’s interest in biotechnology began in the Agricultural Sciences business but broadened into nonagricultural applications in the late 1990s. We were interested in using corn plants as production facilities for monoclonal antibodies for use in human therapeutics. The projected growth rate for monoclonal antibody production was high, and a shortfall in production capability was expected—in the year 2005 the shortfall could be as high as 30 percent of demand. Dow determined that we didn’t have the fundamental science base for such research, so we created an alliance with a start-up company, Epicyte Pharmaceuticals, Inc. Our vision is to meet the growing demand for monoclonal antibodies by combining the power of Epicyte’s expertise in expressing antibodies in plants with Dow’s expertise in the agricultural sciences of corn as well as our overall strength in engineering science and production capability. Combining this new fundamental science with the scientific and engineering strength of Dow Chemical Company clearly was the fastest way to the marketplace. These three examples illustrate market-driven research that requires substantial fundamental research to generate the new knowledge that can, in turn, generate intellectual property and provide protection in a long-term market position. I would now like to focus on three critical success factors paramount for rapid innovation from basic research to the marketplace—the three P’s. The first critical factor is people. I believe people are the main determinant in successful innovation. The right people are needed at the right time and in the right place. People define the environment, and to have an innovative environment requires the right people. Each person not only possesses a skill set that is typically the focus, but they also have a specific mindset. It is critical to have team members appropriately deployed against the different stages of innovation that match their makeup. The Myers-Briggs and KAI testing methodologies are tools that describe the personal profile of specific individuals. It is interesting to watch employees relate their test results to what they have felt and experienced on different R&D assignments throughout their careers. In most cases there is an amazingly high degree of correlation. However, sometimes people are attracted to specific types or stages of research based on neither their mindset nor their skill set. They may be attracted by some perception they have about themselves or about that particular part of research. For instance, there are some people who believe that discovery is more exiting and highly valued than other aspects of research in the development process. In these cases the people are square pegs in round holes. It is therefore very important to move the individual to an area that matches his or her profile, on both skill set and mindset. Proper resource deployment is critical for rapid innovation. The heart of innovation is ideas. To quote a historic Dow R&D leader, John Grebe: “Ideas are among God’s most precious gifts; without them we’d still be living in the Dark Ages. They separate man from all other creatures.” He went on to say: “Listen carefully and keep an open mind. Perhaps you can 

OVERVIEW OF TRENDS IN INNOVATION IN THE CHEMICAL INDUSTRY 13 convert a bad idea into a usable one.” Grebe clearly understood the power of ideas, and he understood the even greater importance of listening with an open mind. Fifty years later Arnold Penzias was asked in an interview with Business Week what made a top- notch research laboratory. Penzias’s answer, concerning the building layout, floored the reporter. Penzias knew the value of ideas and, more importantly, the value of sharing ideas. He knew that no one talks in elevators, so there was no reason to have them. He understood that corridors needed to be long enough to provide opportunities for spontaneous sharing and wide enough that people felt comfortable lingering and building on ideas. Idea sharing is a critical aspect of creating an innovative environment, as is listening to the ideas of others. Work processes take many forms, from the simplest of structure techniques to the most complex multifunctional processes. They are important, and the more complex the innovation task is, the more important they are. Work processes capture the best practices over time, standardize those practices for everyone on the team, and provide a common language for everyone to use. Additionally, they are key to defining success on multiple levels. Many people are familiar with work processes, including Bottom Line Innovation, TRIZ, and the various Six Sigma elements. These standardized techniques for work processes have gained a broader use throughout the chemical industry. At the macro level, much of industry is working with a stage-gate process.5 Stage-gate processes are business activities, not functional R&D processes. If rigorously used, stage processes are beneficial and can focus employees on the critical scientific technology needs for success. Work processes provide a useful framework to do industry-wide benchmarking to evaluate internal performance versus best-in- class standards. This is useful to identify areas that need improvement and to understand what is both internally and externally possible. The third critical factor of successful innovation is partnerships. Data show that there are an increasing number of partnerships in the chemical industry and across all industries in the United States. There are three factors driving partnerships that I would like to mention. The first is industry restructuring. The increased degree of specialization has clearly left many companies without a full hand of cards. Many companies have downsized their R&D organizations or eliminated their central or corporate R&D capabilities entirely. Therefore, necessity has driven a fraction of the increase in partnerships. Second, understanding the importance of being first in the marketplace causes a company to focus on the speed of getting new products to market. With the rapid rate of change in the marketplace and thus in the industry, it is virtually impossible to have all the right skill sets internally at the right time. Partnerships allow a company to put the required skills together before they are needed, regardless of whether they’re internal or external. Finally, when there is focus on meeting unmet customer needs rather than pushing the technology and science interests of the company, both internal and external people become more willing to uti- lize all the required skill sets. Dow’s partnerships with other laboratories have more than tripled over the past 4 years. I want to close my comments on partnerships by emphasizing the importance of the National Institute of Standards and Technology Advanced Technology Program and others. There are many strong points to these programs, but one that does not get enough recognition is the ability to 5 The stage-gate process is the process by which a new project is evaluated at multiple points in its development. For the project to progress from one stage to the next, it must first pass through a gate—a decision-making point where the choice to continue, kill, hold, or recycle the project must be made. This stage-gate process streamlines the innovation process. 

14 REDUCING THE TIME FROM BASIC RESEARCH TO INNOVATION IN THE CHEMICAL SCIENCES provide a framework for large and small companies to better collaborate and cooperate in the spirit of providing solutions and benefits for society more rapidly. In closing, I want to end where I started by discussing the three macrotrends in the chemical industry. First, high-throughput technology makes it imperative that your goals and objectives are clearly defined and understood by all. Second, global organizations and global teamwork reflect how we work. Innovation processes must be able to work across distances and do so rapidly. Third, the secret of innovation in the chemical industry is asking the right questions and the choice of the right market opportunity that drives that speed of innovation. In fact, a proprietary and profitable market position will usually only come through new knowledge that comes from basic research. The critical innovation success factors for the macrotrends above are the three P’s: people, pro- cesses, and partnerships. People define the innovative environment. Work processes capture best prac- tices and standardize them for all members of the team. Finally, partnerships provide the required knowl- edge, understanding, and skill sets in real time to really drive rapid innovation—speed counts in the 21st century. DISCUSSION Hans Thomann, ExxonMobil: First, do you believe that high-throughput experimentation has had a bigger impact on innovation or invention? In either case, what do you anticipate for the future? Richard M. Gross: Clearly, what I have tried to capture is that high throughput is unmistakably going to impact invention. We’ve also been able to take the output of high-throughput research to the marketplace, so the impact on innovation is definitely there. For the future we must have prepared minds. High throughput is not about what work we do; it is about how we do the work. It needs prepared minds that are skillful and knowledgeable in the area to guide it and to set goals and objectives. Without a thoughtful approach, you can just be very busy generating a lot of data without making any progress on new knowledge. I’m hopeful for the future. Robert A. Beyerlein, National Institute of Standards and Technology: First, you mentioned that with restructuring activities going on in many companies, they might not have the full resources needed. If the company’s science base and resources are not complete for the job at hand, how do you address that? Also, what is the process of implementing innovation that allows the company to focus on unmet market needs? I’m curious about how you identify those unmet market needs. Richard M. Gross: Let me take these questions in reverse beginning with unmet market needs. The first thing is to work with the marketplace, not specific customers. Customers tend to be focused on their needs, not necessarily the broader market needs. It is beneficial to work with a large array of customers or players in the market, so the entire picture can be seen. Second, we have found it extremely effective to have one of our most senior scientists involved in that activity. This eliminates the hands-off, nonscientific minds coming back and trying to describe things to the scientific mind. Having a senior scientist involved also puts someone out there who understands the possibilities of science. They then frame and see the problems differently. Having scientific expertise in the marketplace and working with the broader marketplace instead of a specific customer is key to identifying unmet market needs. The first question was how to handle a lack of scientific resources. We are blessed in my company 

OVERVIEW OF TRENDS IN INNOVATION IN THE CHEMICAL INDUSTRY 15 because we still have a corporate R&D organization that represents 25 percent of our 7,000 employees in R&D. We’re very mindful of keeping the basics well tuned. I think the industry has to pick and choose what expertise it has internally and what expertise it will employ externally. Then it must build those relationships and know the sources it will use for the expertise it lacks. Dow does that. To anticipate all the needs of the future is impossible. We go around the world to institutes and universities from China to Russia and everywhere in between. We look for the things that add the pieces to the puzzle so we can put the whole picture together. Michael Schrage, MIT: A few years ago at these kinds of innovation seminars we would have been talking about over-the-wall issues. You take R&D and throw it over the wall to manufacture. You talk about the rise of alliances and the importance of intellectual property. I’m wondering whether Dow and other companies in the industry have experienced a comparable over-the-wall experience in which you have the scientist working on R&D while the agreements are negotiated over the wall by intellectual property lawyers with boilerplates, who may not necessarily appreciate some of the tradeoffs involved in the collaboration and cooperation issues. How sophisticated have the lawyers needed to become, and is this going to be a topic of greater contention or cooperation with intellectual property (IP)? Richard M. Gross: Good question. First of all, if there are any IP attorneys in the room, my apologies for my answer. I do not find the IP attorneys to be the most important people. Many companies have created a group to manage their own intellectual assets. At Dow the Intellectual Asset Management Group is part of the R&D function. This group is the interface between the laboratory scientists and the business leaders. The group members understand the business objectives, the business needs, the IP attorneys, and the whole corporation. These folks are well skilled in looking at this strategically. The way we look at IP is strategically, not as an after-the-fact activity. We spend a lot of time developing and purchasing data-mining technology, so we can look at the topography in order to know where we want to land. Actually, our intellectual asset management folks use the data-mining technology and other patent- mapping technology to guide our research. This is a strategic thrust for us, and I think it is increasingly becoming a strategic thrust for the industry, not an after-the-fact response type of function. Henry F. Whalen, Jr., American Chemical Society: You mentioned that you had a partnership with IBM to develop this porous dielectric with less than 10-nanometer closed pores, but you also said that this truly started as an ATP program. I know Mary Good will talk about ATP later, but would you be where you are today if it hadn’t been for ATP? Richard M. Gross: No. Mary L. Good, University of Arkansas at Little Rock: I wish I had asked that question. Everybody wants to know about processes, but let’s go back to your point because it’s so very important. What is the industry’s thought today on the fact that the enrollment in undergraduate chemistry is going down like a rocket and that graduate enrollment has an increasing number of foreign students? Is that a good thing? What is the industry’s thought about that, and what are we going to do about making a difference? 

16 REDUCING THE TIME FROM BASIC RESEARCH TO INNOVATION IN THE CHEMICAL SCIENCES Richard M. Gross: I think industry is unanimously struggling to understand what we can do not just to help at the undergraduate level but to help all kids have a fuller appreciation of science earlier in life. Quite frankly, my biggest concern is not with finding future employees. My biggest concern is having a science-knowledgeable society to vote. As issues become more technical, I have huge concerns about the people who are voting. I’m concerned about who we’re going to hire when I’m retired, but more importantly I’m concerned about the populace. The key from our vantage point is the uplifting of science teachers. It is clear to me that if a teacher has not had an impact by approximately the fifth grade, the student is lost. When you have such a large population of elementary science teachers who are not trained in science and haven’t the foggiest idea of what they’re teaching, it is little wonder that the kids are not turned on by science. We’re involved with the National Science Resource Center, the partnership between the National Academies and the Smithsonian Institution. Hands-on, inquiry-based science has to be the answer, and it has to be taught by well-informed and well-trained teachers. Joseph S. Francisco, Purdue University: I was struck by your SiLK example. What I thought was very interesting was the role of computational chemistry in this process. To what extent were the computational chemistry explorations important in terms of reducing the time line, and where do you see it being utilized more in this high-throughput process in the future? Richard M. Gross: Joe, thanks for the question, because two things come to mind. The answer is that computational chemistry is huge. Whether it’s designing a new material for the soles of Nike shoes or whether it’s SiLK, the computational chemist could stand here and give you example after example where we’ve never gone into the laboratory until we’ve spent hours, weeks, months in front of the workstation. I have an interesting story about SiLK. When they were doing the computational chemistry, one of the hits they came up with was a material where the published dielectric constant was wrong. The computational chemist brought this forward, and everybody said, “That’s not right, because we know what the dielectric constant is. It’s published.” It turned out the published source was wrong. Computational chemistry is so important. The three to five families of compounds that were identi- fied by the computational chemists were the only thing that the synthetic chemists worked on. SiLK is indeed out of one of those organizational families. They had material in the marketplace within 6 months of having the original go-ahead for pursuing the thought. This was before the computational work had been done—6 months and they had samples in the marketplace. There was a second part to this question, Joe? Joseph S. Francisco: How do you see it going forward in the future? Richard M. Gross: When we talk about high-throughput research, most people think about combinatorial chemistry in the laboratory. When we think about high-throughput research, we also think about the computational dimension and, most importantly, linking them together. This is an area where you can share information instantaneously when working in teams globally. The guys in Europe can be working while you are home sleeping and vice versa. Robert W.R. Humphreys, National Starch and Chemical Company: I applaud you distinguishing between a company that has offices and plants in every part of the world and a company that truly works globally. If the problem were as simple as having computer terminals that speak the same language in every part of the world, we would have solved the problem long ago. What is unique about Dow’s training of culturally different people around the world to work together in global teams? 

OVERVIEW OF TRENDS IN INNOVATION IN THE CHEMICAL INDUSTRY 17 Richard M. Gross: We have a research assignments program in which people hire on to a special assignments program. Most companies have it. We have made that global in recent years. We have new hires that not only do 5- to 6-month assignments around sites in the United States but also go to Europe. This is because we’ve got 1,400 researchers in Europe. It’s our second-largest area in the world. In their first 2 years, they can have a 6-month assignment in another part of the world. We also have a number of the European analytical chemists come over and do a 6-month assignment in the United States. This program develops networks, which gets back to the people issues that are really at the heart of the matter. Robert W.R. Humphreys: Is that done through corporate? Richard M. Gross: Not necessarily. The businesses have a special assignments program as well. In fact, the special assignments program is primarily through the businesses. Mary L. Mandich, Lucent Technologies: When I saw market-driven research, I immediately thought about financial market drivers. How do you think the forward-looking strategy of corporate-supported research and innovation is affected by market trends and stock prices? Richard M. Gross: Dow supports our traditional materials that have been around for a very long time. R&D supports them with ever-lower costs and ever-lower resources, because we continually improve our processes and the way we service the industry. Everyone in the chemical industry is looking for those areas where we can use our capabilities to answer society’s problems. Each company has different capabilities and a different focus. Yes, we are a chemicals company, but we’re largely a materials company as well. We are trying to identify the growth areas with large enough scope and scale where we can bring our expertise. I believe that the chemical industry was the first knowledge-based industry. I believe we’re still a knowledge-based industry. I believe at the end of the day those in the chemical sciences sell knowledge. I think the challenge is that we all have to look for those marketplaces where we can generate new knowledge to help society as well as be rewarded financially.