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.
Opportunities, Challenges, and Needs Information technology (IT) is a major enabler for the chemical sciences. It has provided the chemical scientist with powerful computers, extensive database structures, and wide-bandwidth communication. It permits imagination, envision- ing, information integration and probing, design at all levels, and communication and education modalities of an entirely new kind. The richness and capability of effective data management and data sharing already permit, and in the future will facilitate even more successfully, entirely new kinds of understanding. Combining modeling with integrated data may per- mit the community to predict with higher reliability issues of risk, environmental impact, and the projected behavior of molecules throughout their entire life cycle in the environment. Major progress on issues of societal policy ranging from energy to manufacturability, from economic viability to environmental impact, and from sustainable development to responsible care to optimal use of matter and materials might all follow from the integrated capabilities for data handling and system modeling provided by advances in information technology. This field is beginning to see the development of cooperative environments, where learning, research, development, and design are carried out utilizing both modeling and data access: this cooperative environment for understanding may be the most significant next contribution of IT within the chemical sciences. The questions posed by chemists and chemical engineers are often too com- plex to be solved quantitatively. Nevertheless, the underlying physical laws pro- vide a framework for thinking that, together with empirical measurement, has allowed researchers to develop an intuition about complex behavior. Some of these complicated problems will yield to quantitative solution as computational power continues to increase. Most will not, however, at least in the foreseeable 21
22 INFORMATION AND COMMUNICATION future. These will continue to be addressed by intuition and experiment. The data-handling and visualization capabilities of modern computing will increas- ingly become an essential aid in developing intuition, simple models, and under- lying physical pictures. Thus, information technology has changed and will con- tinue to change the way we think about chemical problems, thereby opening up new vistas well beyond the problems that can be addressed directly by large-scale computation. CURRENT STATUS Computational modeling currently excels in the highly accurate computation of small structures and some gas-phase properties, where experiment can be re- produced or predicted when adequate computational power is available. Tech- niques for the computation of pure solvents or dilute solutions, macromolecules, and solid-state systems are also advanced, but accuracy and size are serious limi- tations. Many efforts in chemical science have merged with biology or with ma- terials science; such modeling requires accuracy for very large systems, and the ability to deal with complex, macromolecular, supramolecular, and multicompo- nent heterogeneous systems. The nanosystems revolution is very small from the perspective of size, but huge from the viewpoints of chemical engineering and chemistry because it will allow wholly different design and manufacture at the molecular level. This will comprise a new subdiscipline, molecular engineering. Computational, communications, and data storage capabilities are increasing exponentially with time. Along with massive increases in computational power, better computational algorithms for large and multicomponent systems are needed if computations are to have a major role in design. Some of the major target opportunities for exploration in the near term include . Computational Methods: The implementation of theoretical models within software packages has now become excellent for certain focused problems such as molecular electronic structure or simple Monte Carlo simulations. Very large challenges remain for extending these methods to multiscale modeling in space and time. · Education and Training: The community has not made as much progress as might be desired in the training both of chemical professionals and profession- als in other disciplines. Training techniques tend to focus on particular packages, lack integration, and be concentrated too narrowly within particular subdisci- plines. Education in chemistry and chemical engineering has not yet utilized IT advances in anything approaching a reasonable way. · Databases: The database resource in the chemical sciences is rich but fragmented. For example, the structural databases for small molecules and for biological systems are in an early stage of development and integration. Often the data have not been verified, and federated databases are not in widespread use.
OPPORTUNITIES, CHALLENGES, AND NEEDS 23 · Ease of Use: The field has a very bad record of resuscitating "dead code" code that has been modified so that some branches will never be ex- ecuted. Too many of the software packages in current use, both commercial and freeware, are very poor from the viewpoints of reliability, user friendliness, port- ability, and availability. Maintenance and improvement of codes is not well handled by the field in general. . Problem Solving: Particular areas such as drug design, molecular struc- ture prediction, prediction of materials properties, energetics of particular reac- tions, and reaction modeling have shown major successes. These successes will provide the springboard for designing the Collaborative Modeling-Data Environ- ments that constitute a major theme of this report. . Optimization: Large-scale nonlinear optimization techniques for continu- ous and discrete variables are just beginning to make their way into every part of the chemical sciences, from the molecular level to the enterprise level. At the molecular level these techniques are being used to design molecular structures, while at the enterprise level they are being used to optimize the flow of materials in the supply chain of the chemical industry. Issues such as sensitivity analysis, parameter estimation, model selection, and generalized optimization algorithms are particularly important but are not yet in common use. · Supply-Chain Modeling: In this crucial area, security, privacy, and regu- latory issues must be addressed. Supply-chain structures are characteristic of large-scale problems of importance both within chemical science and engineer- ing and across the larger society. Just as the impact of such work will be major, so will the importance of assuring accuracy, security, privacy, and regulatory com- pliance. Often the data used in such simulations are proprietary and sensitive; thus, access to both the data and the models themselves must be controlled so that the effectiveness and societal acceptance of such modeling are not jeopardized. CHALLENGES In the future, modeling and data management will become unified. Research, development, education, training, and understanding will be done in a compre- hensive multiscale problem-posing and problem-solving environment. Targeted design, as well as our understanding of simple systems, can profit by investiga- tion within such a holistic research environment. Although the exponential increases in capability and utilization cannot con- tinue forever, the advances already attained mean that the capabilities available to the chemical sciences are not limited to commodity computing and the constraints of more, faster, and cheaper cycles. Major challenges exist in maximizing the ability of chemical scientists to employ the new methods and understandings of computer science: . Develop interdisciplinary cooperation with applied mathematicians and
24 INFORMATION AND COMMUNICATION computer scientists to approach chemical problems creatively. Particular targets would include use of informatics, data mining, advanced visualization and graph- ics, database customization and utility, nonlinear methods, and software engi- neering for optimization and portability of chemical codes. · Develop computational chemical sciences, as a focused subdiscipline of chemistry, chemical engineering, computer sciences, and applied mathematics. Cross-disciplinary training grants, planning workshops, symposia, programming groups, and other forms of interaction sponsored by professional societies and funding agencies should materially help in bringing computer science and ap- plied math expertise to solving advanced problems of chemical modeling and data utilization. . Better integrate the two disciplines of chemistry and chemical engineer- ing. These face nearly identical issues, concerns, advantages, and IT-based chal- lenges. Integrated software, communications, database, and modeling capabili- ties can be used as a pathway for closer linking of these disciplines. · Generate graphical user interfaces. Current advances in computer sci- ence permit adaptation of existing complex codes into a larger problem-solving environment. Tools for semiautomatic generation of new graphical user inter- faces could facilitate calling or retrieving data from these extant programs. Even- tually, one can envision an object-oriented approach that would include inter- faces to a suite of components for integrating chemical science modeling tasks such as electronic structure calculation, smart Monte Carlo simulation, molecular dynamics optimization, continuum mechanics flow analysis, electrochemical pro- cess modeling, and visualization. The disciplinary expertise in chemistry and chemical engineering is a key element in making the interfaces broadly useful within the discipline. · Develop more general methods for manipulating large amounts of data. Some of the data will be discrete, some stochastic, some Bayesian, and some Boolean (and some will be unreliable) but methods for manipulating the data will be needed. This integrative capability is one of the great markers of human intelligence; computational methods for doing such integration are becoming more widespread and could transform the way that we treat the wonderful body of data that characterizes the chemical sciences. . Help the IT community to help the chemical world. One issue is the need to modify the incentive and reward structure, so that optimization experts or data structure engineers will want to attack chemical problems. Another issue is the rapid exporting of IT offshore, while national security and economic consider- ations require that extensive IT expertise and employment also remain in the United States. A national award for IT professionals working in the chemical sciences would be useful as one component for addressing these issues. Another possibility would involve new postdoctoral awards for IT experts to work within the chemical sciences.
OPPORTUNITIES, CHALLENGES, AND NEEDS 25 The preceding challenges involve bringing tools from information technol- ogy, in a much more effective fashion, to the solution of problems in the chemical sciences. There are also major problems within the chemical sciences themselves, problems that comprise specific challenges for research and technology develop- ment. The largest-scale challenges deal with the ability of chemistry and chemi- cal engineering to address major issues in society. We should focus on major issues: · Provide stewardship of the land. We need to develop new methods in green chemistry for manufacturing in the twenty-first century, and we need to accelerate environmental remediation of sites around the world that have been polluted over the previous century. Examples include o incorporating computational methods into sensors for real-time analysis and assimilating sensor-measurement-information data in simple yet techni- cally correct formats for use by public policy decision makers; and o expanding our environmental modeling efforts especially atmospheric and oceanic modeling to account for the impact and fate of manufactured chemicals; to assess how changes in air and water chemistry affect our health and well-being; and to develop alternative, efficient, and clean fuel sources so that we need not rely on imported hydrocarbons as our major energy source. . Contribute to betterment of human health and physical welfare. This chal- lenge extends from fundamental research in understanding how living systems function as chemical entities to the discovery, design, and manufacture of new products such as pharmaceuticals, nanostructured materials, drug delivery de- vices, and biocompatible materials with lifetimes exceeding patients' needs. This challenge is especially relevant to computing and information-based solutions, because we will need to develop smart devices that can detect and assess the early onset of disease states. Moreover, we need to convey that information in a reli- able manner to health care professionals. We will also need to develop large-scale simulations that describe the adsorption, distribution, metabolism, and excretion of as-yet-unsynthesized drugs. We need to do this to account for naturally occur- ring phenomena such as bacterial evolution that make extant drugs ineffective, and we must be able to do it in quick response to potential homeland security disasters that affect our health and lives. . Ensure an informed citizenry through education. Our democratic society depends on an informed citizenry, and a major challenge facing the chemical sciences community involves education. This includes reaching out to our col- leagues in other disciplines who are beginning to study how things work at the molecular level. A need exists to integrate seamlessly chemistry with biology and materials science and to strengthen the connections between chemistry and chemi- cal engineering. The challenges described above concerning environmental,
26 INFORMATION AND COMMUNICATION health and welfare issues are also challenges to the community of educators. The educational challenge includes public relations, where it will be necessary to ad- dress the negative connotations of the word "chemical." Educational efforts that use such information technology tools as websites and streaming video could help to deliver accurate information and to inform the level of discourse within the public domain. · Facilitate more thoughtful decision making. We need informed policy and decision makers to guide our society in times of increasingly complex techno- logical issues. Examples include international policies on health, disease, envi- ronment, natural resources, and intellectual property, to name a few. Issues that arise include mining databases for relevant information, developing and evaluat- ing model scenarios, and assessing uncertainty and risk. The goal "every child a scientist" should be attainable, if by "scientist" we mean someone with an under- standing of what science is. . Protect and secure the society. Of all the scientific and engineering disci- plines, the chemical sciences and technology have perhaps the strongest influ- ence on the economy, the environment, and the functioning of society. The com- munity of chemical engineers and scientists must retain responsibility both for maintaining the traditional intellectual and methodological advances that chemis- try and chemical engineering have provided and for responsibly managing the footprint of chemistry. Issues of privacy and security are crucial here, as are scru- pulous attention to responsible care and continued responsible development. Us- ing information technology to achieve these goals will greatly enhance our ability to protect and secure both the principles of our discipline and the civilization in which that discipline is used and practiced. Finding: There are major societal and civic problems that challenge the chemical community. These problems should be addressed by chemistry and chemical engineering, aided by IT advances. These societal issues include providing stewardship of the land, contributing to the betterment of human health and physical welfare, ensuring an informed citizenry through education, facilitating more thoughtful and informed deci- sion making, and protecting and securing the society. The chemical sciences need to develop data-driven, natural teaching- and information-capture methods, preferably including some that do not require equa- tion-based algorithms (what might be called natural learning). The community should develop means for using assembly-knowledge methods to produce weighted, customized data-search methodology (which would, effectively, corre- spond to designing a chemistry-engineering Google search engine). The commu- nity also should utilize such popular IT advances as web searching and 3-dimen- sional graphics developed for games. Finally, a verification or standardization scheme for data is needed. Within the overall challenge themes just expressed, it is useful to focus on
OPPORTUNITIES, CHALLENGES, AND NEEDS 27 Sampling is a key bottleneck at present in obtaining accurate results in molecular modeling simulations. Obtaining convergence for a complex condensed-phase system is extremely challenging. This is the area in my opinion where prospects are most uncertain and where it is critical to support a lot of new ideas as opposed to just improved engineering of existing approaches. Some advances will come about from faster hardware, but algorithmic improvement should contribute even more if sufficient effort is applied. Richard Friesner (Appendix DJ several specific, targeted challenges that the chemistry and chemical engineering community will need to address: . Multiscale methodologies are crucial. The community needs to incorpo- rate advanced theoretical ideas to generate nearly rigorous techniques for extend- ing accurate simulations to deal with phenomena over broad ranges of time and space. It also needs to attack such methodological problems as model-based ex- perimental design, virtual measurement, quantum dynamics, integration with con- tinuum environments, dispersion energetics, excited states, and response proper- ties. All of these are part of the multiscale modeling capability that will be crucial if the chemical community is to utilize advanced IT capabilities in the most effec- tive and productive fashion. Optimization must go beyond studies of individual systems and deal with chemical processes. The community must develop effective design, planning, and control models for chemical processes. It will be necessary to address the integration of these models across long time scales, as well as their accuracy and utility. The community must develop enterprise-wide optimization models. Using advanced planning and scheduling methods, the models must allow fast and flex- ible response of chemical manufacturing and distribution to fluctuations in mar- ket demands. The economic implications and societal importance of this effort could be enormous. . . Finding: The nation's technological and economic progress can be ad- vanced by addressing critical needs and opportunities within the chemical sciences through use of new and improved information technology tools. Bringing the power of IT advances to bear will greatly enhance both targeted design through multidisciplinary team efforts and decentralized curiosity- driven research of individual investigators. Both approaches are important, but they will depend upon IT resources in different ways. Among the needs to be addressed are:
28 INFORMATION AND COMMUNICATION · Take strategic advantage of exponential growth of IT resources, which will revolutionize education, research, and technology in chemical science and . . englneenng. · Develop a rational basis for dealing with complex systems. · Embrace simulation, which is emerging as an equal partner with expen ment and theory. Utilize web-based collaborative problem solving: Collaborative Model- ing-Data Environments). . . . Recognize the increasing importance of multiscale, multiphenomena com- puting. · Maintain support for the fundamental aspects of theory that will remain essential for progress. Support an approach to algorithm development that is application driven. Facilitate the development of technologies for building reliable and se- cure multiscale simulations. Multiscale computational methodologies are crucial for extending accurate and useful simulations to larger sizes and longer times. Fundamental advances in the formalisms and methodologies, as well as in algo- rithmic software and visualization advances, will be needed to make effective multiscale modeling a reality. Finding: To sustain advances in chemical science and technology, new approaches and IT infrastructures are needed for the development, sup- port, and management of computer codes and databases. Significant breakthroughs are needed to provide new means to deal with complex systems on a rational basis, to integrate simulations with theory and experiment, and to construct multi-scale simulations of entire systems. It will be necessary to · develop methods for semi-automatic generation of user interfaces for codes and design modules in the chemical sciences with the eventual goal of a semiautomated object-oriented modeling megaprogram~s) containing modules for specific capabilities; · develop reliable error estimators for computational results; . develop enhanced methodology for data mining, data management, and data-rich environments, because databased understanding and insights are key enablers of technical progress; and · develop improved methods for database management, including assur- ance of data quality.
