Page 62

4

Interactions of the Fusion Program with Allied Areas of Science and Technology

INTRODUCTION

Historically, the development of the fusion program involved both basic physics and the applied and engineering sciences. However, because the energy goal of the fusion program is an application, the contributions of this program to our understanding of fundamental physics are sometimes obscured. In reality, high-temperature plasmas are not only of great intrinsic scientific interest but also of great general interest in fields from astrophysics to material science, with the goal of basic plasma physics being to elucidate the linear and nonlinear properties of the plasma medium under the wide variety of conditions in which it is encountered in nature and in the laboratory. The fusion program should be credited with having played a major role in advancing our experimental and theoretical understanding of this “fourth state of matter.”

Many of the fundamental concepts of fusion science, ranging from linear stability theory to kinetic theory and nonlinear dynamics, as well as its experimental techniques, have close connections to other areas of physics. Plasma physics has, in fact, been the incubator for a number of key research areas in modern physics, and in some cases, plasma scientists became leaders of the emerging fields—solitons, chaos, and stochasticity are noteworthy examples. In addition, basic tools developed in the fusion program, from computer-based algebra to particle simulation techniques, have found widespread application in allied fields.

Historically the United States has been a leader in advancing all aspects of magnetic fusion and plasma science, particularly with respect to the ability of its fusion program to pose and answer deep scientific questions on issues relevant to both fusion and the broader scientific world. At a time when U.S. funding for the fusion effort has been cut back, it is important to ask what the forefront topics of interdisciplinary research are, whether the program is continuing to engage these critical topics, and whether it is advancing the computer and technological tools required to solve the critical problems facing plasma and fusion science. In the following sections, the committee discusses a number of these issues.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 62
Page 62 4 Interactions of the Fusion Program with Allied Areas of Science and Technology INTRODUCTION Historically, the development of the fusion program involved both basic physics and the applied and engineering sciences. However, because the energy goal of the fusion program is an application, the contributions of this program to our understanding of fundamental physics are sometimes obscured. In reality, high-temperature plasmas are not only of great intrinsic scientific interest but also of great general interest in fields from astrophysics to material science, with the goal of basic plasma physics being to elucidate the linear and nonlinear properties of the plasma medium under the wide variety of conditions in which it is encountered in nature and in the laboratory. The fusion program should be credited with having played a major role in advancing our experimental and theoretical understanding of this “fourth state of matter.” Many of the fundamental concepts of fusion science, ranging from linear stability theory to kinetic theory and nonlinear dynamics, as well as its experimental techniques, have close connections to other areas of physics. Plasma physics has, in fact, been the incubator for a number of key research areas in modern physics, and in some cases, plasma scientists became leaders of the emerging fields—solitons, chaos, and stochasticity are noteworthy examples. In addition, basic tools developed in the fusion program, from computer-based algebra to particle simulation techniques, have found widespread application in allied fields. Historically the United States has been a leader in advancing all aspects of magnetic fusion and plasma science, particularly with respect to the ability of its fusion program to pose and answer deep scientific questions on issues relevant to both fusion and the broader scientific world. At a time when U.S. funding for the fusion effort has been cut back, it is important to ask what the forefront topics of interdisciplinary research are, whether the program is continuing to engage these critical topics, and whether it is advancing the computer and technological tools required to solve the critical problems facing plasma and fusion science. In the following sections, the committee discusses a number of these issues.

OCR for page 62
Page 63 WHAT ARE THE DEEP SCIENTIFIC CONTRIBUTIONS THAT HAVE IMPACTED OTHER PHYSICS FIELDS? The core objective of the fusion energy science program is to reach a fundamental physical understanding of the behavior of high-temperature plasmas in the context of plasma configurations capable of plasma confinement sufficient for economic energy extraction. Fusion science aims to study the stability properties and transport behavior of such systems, and in order to conduct these studies, it has made progress in a number of topical areas that have had a broad impact on the larger scientific and industrial community. Examples of some cross-cutting research topics are stability theory; stochasticity, chaos, and nonlinear dynamics; dissipation of magnetic fields; origins of magnetic fields; wave dynamics; and turbulent transport. Stability Theory The complex plasma dynamics due to macroinstabilities observed in early plasma experiments motivated the development of powerful energy principles and eigenmode techniques for exploring the linear stability of plasma equilibria. The wide variety of instabilities in plasmas, which span an enormous range of spatial and temporal scales, defines the richness of the plasma medium and challenges us to understand its dynamics. Past research supported by the fusion program greatly improved our ability to predict the thermal pressure beyond which a plasma will disassemble. These predictions were confirmed in experiments in which the plasma temperatures exceeded those found in the core of the Sun. Experimental explorations led to methods that significantly increase the plasma pressure limits set by stability. It speaks to the quality of those studies that the stability analysis techniques developed by the fusion program—such as the energy principle, the notion of convective instability, and weakly nonlinear stability theory—are now essential tools not only in the field of plasma science but also in allied fields such as fluid dynamics, astrophysics, and solar, space, ionospheric, and magnetospheric physics. Stochasticity, Chaos, and Nonlinear Dynamics Understanding how magnetic field topology—the existence of bounding magnetic flux surfaces that lead to hot plasma confinement—is controlled by both local and global physical processes and how such bounding flux surfaces break up is a critical research topic for fusion. Interestingly, the same kind of issues also emerged in the description of how an essentially collisionless, unmagnetized plasma is heated; there, the onset of stochasticity needed to be understood in velocity space. This research followed on the heels of the pioneering Kolmogorov-Arnold-Moser (KAM) description of the onset of chaos and therefore contributed to the rapid progress in this field. A number of fundamental tools, including the standard map—now in common use in studies of the onset of stochasticity and chaos in far more general physical settings—came from plasma scientists. Studies of nonlinear wave-wave and wave-particle interactions relevant to both plasma confinement and transport played an important role in the development of tools for dynamical systems theory and nonlinear dynamics, and it was senior scientists trained in the physics of plasmas who developed the first published method for controlling chaos. Dissipation of Magnetic Fields A fundamental physics challenge has been to explain the observed very short timescales that characterize the release of magnetic energy in the solar corona, in planetary magnetospheres (including that of Earth), and in fusion experiments. Classical collisional dissipative processes are orders of magnitude too

OCR for page 62
Page 64 weak to explain the timescales observed. The difficulty lies in the extreme range of spatial scales, from the macroscopic to the microscales associated with kinetic boundary layers, and in the necessity to include kinetic processes to provide collisionless dissipation. An emerging understanding based on theory, computation, and basic experiments is linked to the mediating role of dispersive waves, which act at the small scales where the frozen-in condition is broken. For the first time, the predictions of energy release rates in fusion experiments are consistent with observations. One consequence of the fast release of magnetic energy associated with magnetic reconnection in some fusion experiments is the evolution to a minimum energy state (the Taylor state, named after the U.K. physicist who made the theoretical prediction), where the magnetic field is partially self-generated by the plasma. The resulting “dynamo” action is related to magnetic dynamo processes in astrophysical systems such as the Sun, the planets, and the galaxy, discussed next. Origins of Magnetic Fields The origin of cosmic magnetic fields is another of the fundamental puzzles of physics. Magnetic fields are ubiquitous in nature. Objects as diverse as planets, stars, galaxies, and galaxy clusters show evidence of magnetic fields and show phenomena that are intimately tied to the presence of these fields, but—with the possible exception of Earth's planetary dynamo—their origin is not understood. Although theory and simulations have made great strides in elucidating the basic physical principles of magnetic field generation, the key missing piece has been controlled laboratory experimentation. With a few exceptions, astrophysicists, geophysicists, and applied mathematicians led the most important past developments in dynamo theory and simulations; now, with the recognition that fast dynamo processes also play an important role in fusion devices, fusion researchers have begun to play an important role as well. On the experimental front, the dynamo in the reversed-field pinch device remains among the few laboratory demonstrations of a turbulent dynamo. These fusion experiments are pushing the subject into previously unexplored areas, including the development of magnetic dynamo theory beyond the single-fluid MHD regime. These studies will be of direct relevance to the origins of magnetic fields on the largest scales in the universe, seen in galaxies and clusters of galaxies. Wave Dynamics The plasma state is unique in the rich variety of waves that are supported by it. Waves in plasmas not only appear spontaneously as a consequence of instabilities but also can be generated to control plasma temperature and currents. Understanding how waves propagate and are absorbed in nearly collisionless plasma was a key scientific goal for the fusion program and has had an important impact on understanding phenomena in space plasma physics. Building on Landau's idea of the wave-particle resonance as a mechanism for collisionless dissipation, fusion scientists developed models to describe the absorption of high-power radio-frequency waves from kilohertz to multigigahertz and benchmarked the predictions in fusion experiments. Waves could then be used to engineer the phase space of particle distribution functions. Waves can now be excited in plasmas to generate intense currents or to accelerate particles to high energies, a technique that may be applied to a future generation of high-energy accelerators. The nonlinear behavior of waves has also been an intrinsic component of the science of plasma wave dynamics, and this knowledge has spread widely to many other branches of physics. Indeed, such ubiquitous concepts as absolute and convective instabilities, solitons (nonlinear waves that persist through collisions), and parametric instabilities were extensively developed within the fusion context. Important industrial applications include the use of radio-frequency technologies for plasma processing in

OCR for page 62
Page 65 semiconductor manufacturing. These ideas emerging from the fusion science program have also had an impact in less obviously allied science areas: for example, plasma physicists introduced the idea of using solitons for commercial high-speed communications. Turbulent Transport Understanding transport driven by turbulence is critical to solving such important problems as the accretion of matter into black holes, energy transport in stellar convection zones, and energy confinement in fusion experiments. Both experiment and theory have shown that gradients in pressure, angular momentum, or other sources of free energy drive small-scale turbulent flows that act to relax the gradients. This “anomalous transport” process is to be contrasted with classical transport, which arises from two-particle Coulomb interactions in magnetic fusion plasmas (as well as from photon diffusion in some astrophysical systems such as stellar interiors). The identification of anomalous transport in fusion experiments (and the corresponding theoretical work) sparked the recognition of its importance in space science and astrophysics, fields in which concepts such as anomalous transport, heat flux inhibition, and turbulent heating are now common language. This cross-fertilization continues: experimental work in fusion has shown that turbulence can be spontaneously suppressed and a transport barrier formed, and that the responsible mechanism is linked to the development of local zonal flows, which shred the vortices driving transport. These experimental developments, together with theory and simulations, have intimate connections with work on zonal flows in laboratory fluid dynamics (e.g., the Couette flow) and planetary physics, in which similar transport barriers and processes have been investigated. WHAT HAVE BEEN THE FUSION-SPECIFIC CONTRIBUTIONS OF THE UNITED STATES TO THE WORLD PROGRAM? In addition to the examples of fundamental science issues discussed in the previous section, a number of significant science discoveries closely linked with fusion can be identified. The United States has traditionally played a central role as a source of innovation and has to a greater extent than other countries sought to understand at a fundamental level the physical processes governing observed plasma behavior. This role is reflected in the U.S. dominance of plasma theory, which is an essential tool for unraveling the complexities of plasma dynamics. The U.S. fusion program has advanced the world effort in a number of key scientific areas: In the late 1950s, the basic energy principle for MHD stability was largely derived in the United States, particularly using the powerful Lagrangian variational δW approach. Also in the late 1950s, U.S. scientists developed techniques for modeling collisions by Coulomb forces in a way suitable for large-scale computations and showed the existence of a large variety of kinetic waves in a magnetic field (now known as Bernstein waves). During that same period, U.S. scientists invented important magnetic confinement schemes such as the stellarator. Later, the favorable properties of the spherical torus configuration were demonstrated theoretically (and have been confirmed in recent experiments in the United Kingdom). In the 1960s, the United States also led the effort to understand resistive instabilities in magnetically confined plasmas, which led to the identification of modes, such as the tearing mode, that grow on a timescale intermediate between the MHD and resistive timescales. Resistive-MHD theory and computer simulation of tokamak instabilities in the 1970s were able to reproduce the sawtooth phenomenon and also many features of tokamak disruptions. The phenomenon of “second stability,” in which increasing plasma pressure surprisingly leads to greater stability, was a discovery of the U.S. program. Also in the 1970s, the United States led the effort to develop a variety of methods for the radio-

OCR for page 62
Page 66 frequency heating of plasmas to thermonuclear temperatures. In the late 1970s and early 1980s, U.S.-invented methods for sustaining plasma currents by waves stimulated noninductive current drive programs on essentially all tokamak facilities worldwide. In the 1980s, discrete Alfvén-wave instabilities in tokamaks, including those that could be driven by alpha particles, were identified, with the United States playing a leading role; these modes were first observed on U.S. tokamaks in the 1990s. U.S. experiments first identified the ion-temperature-gradient instability as a driver for thermal transport in toroidal systems. The understanding of the role of sheared flow in reducing transport was also led by U.S. scientists. In the 1990s, the United States made the first observations of the so-called “reversed shear” mode of enhanced confinement in tokamaks. The United States also played a leadership role in applying modern computational tools to plasma problems, including particle-in-cell methods and gyrokinetic and gyrofluid simulations. The application of these computational tools, together with theoretical advances in the understanding of turbulence and associated anomalous transport, has led to the beginnings of predictive capability for energy storage in magnetic containers, as described in Chapter 2. WHAT ARE THE FUTURE FOREFRONT AREAS OF INTERDISCIPLINARY RESEARCH? A wealth of exciting new areas of research can be identified for high-temperature plasma physics. Some examples primarily of interest to fusion science have already been given. Other areas include the following: New methodologies for advanced diagnostics and modeling. Fusion research is among the most advanced disciplines in terms of the integration of experiments, diagnostics, theory, and numerical simulations. Advances in this area will have significant impact on allied disciplines, including astrophysics and space physics. Material science. The development of materials that can withstand the huge energy and neutron flux of a fusion environment, as well as the large energy flux of present experiments, is an enormous challenge. The material must survive for long periods, release minimal impurities into the plasma, and be minimally radioactive. The development of new composite materials is under way, as are new techniques such as flowing liquid walls. In many areas, the richness of new research areas relates to the deep connections between plasma physics and other physical science disciplines and complements work done purely in the fusion domain. Promising interdisciplinary research areas include the following: Nonneutral plasmas. The acceleration of energetic dense beams of ions or electrons involves the physics of nonneutral plasmas, where a number of collective effects play important roles. Techniques developed in plasma physics have played—and continue to play—a central role in the design and optimization of particle accelerators and colliders, from the heavy-ion accelerators used for nuclear studies, to the electron accelerators used as synchrotron radiation sources for the exploration of the structure of materials, to the high-energy accelerators and colliders used to explore the fundamental constituents of matter. A very different regime of very-low-energy, nonneutral plasmas, namely the confinement of nonneutral plasmas in Malmberg-Penning and Paul traps, is yet another modern beneficiary of plasma physics research and has led to an exploration of the fascinating physics regime of crystallized plasma.

OCR for page 62
Page 67 High-energy-density plasmas. With the advent of high-power laser facilities, such as those at the University of Rochester, the Naval Research Laboratory, and (under construction) Lawrence Livermore National Laboratory, it becomes possible to explore entirely new regimes of plasma physics in which momentum and energy exchange between photons and matter dominates the physics. This regime of radiation hydrodynamics, which plays important roles in areas such as astrophysics, is experimentally largely unexplored (see Figure 4.1 ). Nonideal plasmas. In such plasmas, sometimes also referred to as strongly coupled plasmas, n-body interactions (n > 2) become competitive with two-body interactions. Such plasmas arise in the context of high-energy-density matter, which occurs in high-power laser experiments as well as in nature (for example, on the surface of neutron stars). Relativistic plasmas. The dynamics of relativistic plasmas, as occur in astrophysical jets and in the magnetospheres of compact stellar objects such as neutron stars, as well as in laser-driven plasmas, is largely unexplored by experiment (see Figure 4.1 and Figure 4.2 ). Dusty plasmas. Such plasmas arise in a wide range of physical circumstances, including planetary rings and disks, the atmospheres of cool stars, and the interstellar medium, as well as in terrestrial circumstances. Extremely weakly coupled plasmas. In some physical circumstances, such as accretion onto black holes, there is evidence that coupling between electrons and ions (mainly protons) may be virtually entirely suppressed; in such “advection-dominated accretion flows” (ADAFs), accretion occurs without the usually expected radiative emissions. The plasma physics of such flows is not well understood. It would constitute a major advance if the appropriate physical conditions could be experimentally attained and the physics in this regime explored. Multiphysics/multiresolution simulations. From the computational perspective, plasma physics is unique in that while the complex (plasma) physical processes do span a large dynamic range in both time and space, these ranges are not so large as to be unattainable by present or future large-scale computations. For this reason, plasma experiments, in both the kinetic and fluid regimes, may provide outstanding validation testbeds for developing kinetic and fluid codes for other disciplines, including astrophysics and space physics. Another new area of interest for experimental (fusion) plasma research is its connection with astrophysics as fusion experiments directly relevant to important astrophysics problems become more sophisticated and as theory and simulations capable of connecting experiments to astrophysical observations become more sophisticated. The principal areas of interest are the origins of magnetic fields (the dynamo problem) and the dissipation and reconnection of magnetic field lines. To a significant extent, existing fusion experiments (such as tokamaks and reversed-field pinch devices) have been used to study both dynamo action (especially the alpha effect) and reconnection. In addition, there has been a new effort to construct more specialized experiments to explore these issues. Examples include the following: Liquid metal dynamo experiments, which try to answer fundamental questions about “slow” and “fast” magnetic dynamos, including nonlinear limiting of the alpha effect and the effects of nonlinearities on turbulent diffusion, and Reconnection experiments, ranging from the low-β to the high-β limits and from the collisionless to the fully collisional limits, which try to answer the question of whether (and how) “fast” reconnection can take place and explore the transition from single-fluid MHD to the kinetic regime (see Figure 4.3 and Figure 4.4 ).

OCR for page 62
Page 68 ~ enlarge ~ FIGURE 4.1 These two images show two views of the famous Crab Nebula, a remnant of a supernova seen in A.D. 1054. The optical image (top left) shows the highly filamented state of the remnant's gaseous interior that is the result of synchrotron radiation from energetic electrons spiraling in the magnetic field of the remnant. The Chandra X-ray Observatory image of the Crab Nebula (bottom right) shows, for the first time, the x-ray inner ring within the x-ray torus, with the suggestion of a hollow-tube structure for the torus and x-ray knots along the inner ring and (perhaps) along the inward extension of the x-ray jet. Courtesy of (top left) P. Scowen and J. Hester (Arizona State University) and Mt. Palomar Observatories and (bottom right) NASA, Chandra X-ray Observatory Center, and Smithsonian Astrophysical Observatory (SAO).

OCR for page 62
Page 69 ~ enlarge ~ FIGURE 4.2 A Chandra X-ray Observatory (ACIS-S) image (top) of x rays from the quasar PKS 0637-752. The contours show an overlay of the 8.6-GHz emission measured with the ATCA. One tic mark is 1 arcsec, which corresponds to 9.2 kpc in the plane of the sky for the redshift 0.652 of this object. The x-ray jet, which closely correlates with the radio emission from 4 to 10 arcsec west of the nucleus, is the largest and most luminous detected to date. The radio polarization fraction and E-vector direction are shown in the bottom panel. Clearly, major events are taking place 11 arcsec west of the quasar, where the radio jets bend, the x-ray emission drops, and the radio emission becomes unpolarized, followed by realignment of the magnetic field direction. Courtesy of NASA, SAO, D. Schwartz (SAO), and the PKS 0637-752 Consortium.

OCR for page 62
Page 70 ~ enlarge ~ FIGURE 4.3 This solar image was taken with the TRACE telescope, using a normal incidence soft x-ray mirror whose band pass is centered on an emission line of Fe IX; the image clearly shows the highly complex structuring caused by solar surface magnetic fields and the remarkable “plasma loops,” which indicate million-degree solar gases trapped by these magnetic fields. Courtesy of NASA and the Stanford-Lockheed Institute for Space Research.

OCR for page 62
Page 71 ~ enlarge ~ FIGURE 4.4 This image shows the interior of the MRX reconnection experiment. The pair of large coils produce oppositely directed magnetic field lines. The bright area between the coils (in an X-shaped pattern) is the region where magnetic energy is being released into the plasma through the reconnection process. Superimposed is a snapshot of the measured magnetic field. Courtesy of M. Yamada (PPPL). Reprinted by permission from Journal of Geophysical Research, 1999, vol. 104, p. 14529. DOES THE FIELD MAINTAIN LEADERSHIP IN KEY SUPPORTING RESEARCH AREAS? Quite aside from spawning new ideas that gain currency in other allied fields of physics, fusion research also involves the use and development of tools that it shares with other physical science disciplines. An important question is, To what extent has fusion research maintained the links required to maintain the flow of scientific information between it and allied disciplines? To illustrate the issues involved, the committee has focused on three areas—computational physics, applied mathematics, and experimental tools and techniques for producing and diagnosing plasmas—as examples of how these interactions have evolved.

OCR for page 62
Page 72 Computational Physics Computations are playing an increasingly important role in nearly all areas of science. The effective use of computers requires a close interplay between physical reasoning, approximation, applied mathematics (including algorithm design), and computer science. To ensure that its computational tools remain at the forefront of current technology, a physical discipline must maintain strong interdisciplinary connections to applied mathematics and computer science. In fusion research, computations have from the early days of the program been an essential tool to understand the highly nonlinear processes characteristic of high-temperature plasmas. Indeed, fusion science played a central role in pushing the development of computations as an ancillary tool for theoretical physics. Thus, fusion research was clearly identified with the frontiers of computational science from its beginnings, the early work on kinetic codes being an outstanding example. The present fusion program is facing a number of grand challenge computational problems, including the dynamics of short-scale turbulence, which controls energy, particle, and momentum transport in fusion confinement systems, and the evolution of larger scale instabilities, which lead to the rapid loss of magnetic containment on a large scale (disruptive events). Both areas of research are characterized by strong dynamical nonlinearity and a very large range of space- and timescales. In addition to the grand challenge problems, codes have been developed to explore plasma equilibrium and stability in complex geometries; radio-frequency wave propagation and dissipation for exploring heating; and current drive and real-time analysis of data from experiments. Other examples of areas where computations have played a crucial role were discussed in greater depth in Chapter 2 . A fundamental question is whether scientists in the fusion program have developed the computational tools required to explore the complex nonlinear dynamical processes that underlie plasma behavior. Some of these issues were discussed in Chapter 2. Here, the committee briefly discusses two of the grand challenge topics, turbulence and macroscale dynamics, commenting on the computational techniques employed and areas where greater resources or more effort is required. Scientists studying transport have developed novel magnetic-flux-tube-based coordinate systems for dealing with the extreme anisotropies in the turbulence that develops along and across the magnetic fields confining the hot plasma. The gyrokinetic and gyrofluid techniques, which average over the rapid ion and electron gyromotion, have been developed to uncover the slowly growing instabilities that characterize the fluctuations that occur as a result of the pressure gradients in magnetically confined plasma. These turbulence codes have pushed the frontiers of modern computation in terms of size and complexity and challenges for load-balancing and multiresolution calculations; the sophistication of the physics characterizing these codes is comparable to that of any of the forefront computational areas in modern physics. Visualization techniques developed for understanding the dynamics of large volumes of data in complex geometries are also pushing the state of the art. In addressing the dynamics of instabilities at scales up to the system size, codes have been developed to effectively treat the complex geometries of modern fusion containment experiments, which is important since the shape of magnetic surfaces can strongly affect the stability and dynamics of large-scale instabilities. In addition, implicit time-stepping techniques have been developed to allow the longtimescale evolution of slowly growing dissipative instabilities. The scientists in the fusion program have been less effective in developing advanced techniques for handling the strong nonlinearities that can develop in plasma systems (Godunov, Flux Corrected Transport (FCT), and related approaches), nor have they been at the forefront of the development of adaptive mesh and similar techniques for addressing the extreme range of scales that is an intrinsic aspect of the dynamics of high-temperature plasmas with very weak dissipation. Stronger links between the fusion community, the astrophysical and space

OCR for page 62
Page 73 physics community, and the fluid dynamics community would benefit all of these communities. The macroscopic simulation codes are now being ported to the parallel architectures of the most powerful computing platforms, but this task should receive higher priority, because advanced computation in this area requires the effective use of these machines. Finally, as discussed in Chapter 2, the description of the macroscopic dynamics of high-temperature fusion plasmas requires the use of non-MHD physics at small spatial scales. The inclusion of the requisite kinetic physics and the scale-lengths required to model this physics while at the same time describing the macroscales is a major computational challenge for the program. The development, for example, of adaptive mesh techniques for including this physics would significantly strengthen the physics basis of large-scale modeling of fusion systems and would probably have an impact on the space and astrophysics communities. Computation is clearly an important force in advancing the understanding of fusion plasmas. In recent years the rapidly increasing ability of computational physicists to successfully model the phenomena observed in experiments has led to a greatly increased demand for active simulation of ongoing experiments. As a consequence, the computational challenges facing the field exceed the resources available, and not enough young scientists are entering the field. The insufficient support also translates into a relatively weak effort to develop new codes that use modern object-oriented programming techniques and into inadequate computer facilities. The latter issue is addressed more fully in the next section. The DOE has recognized that the lack of support for computation and associated theory is hindering the ability of the program to interpret and understand the plasma dynamics being measured in experiments. In response, it has initiated the Plasma Science Advanced Computation Initiative, which directs new resources into selected areas of computation of importance to the program. This is an essential step toward rectifying a significant deficiency in the program and is strongly endorsed by the committee. The creation of several centers of excellence in plasma science, as recommended by the committee, will also facilitate increased funding in this important area. Access to State-of-the-Art Computational Platforms The most powerful computational platform available to scientists in the fusion program is the CRAY T3-E at the National Energy Research Scientific Computing Center (NERSC). This machine is not adequate for the frontier computations in transport (which require the treatment of the cross coupling of electron and ion scales, as discussed in ) or macroscopic dynamics (which also requires the treatment of both macro- and microscales). In January 2001, a new IBM-SP, the “Glenn Seaborg” machine (Phase 2, with 2048 processors), became available at NERSC. The new machine, which is comparable to the machines now in use by the DOE DP laboratories, is a significant improvement over the CRAY T3-E. However, the DOE DP laboratories are in the process of procuring even more powerful machines, which will not be accessible to fusion researchers. To develop the predictive capability required for performance predictions in present and future machines, the fusion program will need to keep pace with advances in computational power. Community Codes The development of community codes is an essential product of large scientific programs such as fusion since such codes facilitate the exploration of physics issues by all elements of the community,

OCR for page 62
Page 74 even those that do not have the resources needed to develop code of the complexity that characterizes most plasma problems. Community equilibrium and stability codes are widely available and shared. The TOQ is a twodimensional equilibrium code, and VMEC can do both two- and three-dimensional equilibria. MHD stability codes that are widely available include PEST and ERATO/GATO. Kinetic stability codes include FULL and GS2. All of these stability codes can handle the complex shapes that characterize modern plasma confinement machines. A second class of codes analyzes data from experiments to reconstruct equilibria and analyze energy and particle transport. Examples include TRANSP and EFIT. Very few of these codes are available on a Web site with appropriate documentation so that they can be simply downloaded and used. Web-based, open-source access to the standard tools of plasma science should be developed by the fusion program to facilitate progress in the field. In many cases, the codes could be upgraded using modern programming to facilitate their future enhancements. In the last several years there have been several cross-institutional efforts to develop code, including the following: NIMROD, a multi-institutional team project to develop a nonideal MHD code. The goal of NIMROD is to exploit modern computer methods to model the large-scale dynamics of plasma confinement systems. A similar effort (MH3D) is centered at the Princeton Plasma Physics Laboratory (PPPL). The funding of both efforts has been enhanced through PSACI. The National Transport Code Collaboration (NTCC), a DOE-promoted effort to construct a modular transport code for the community. The idea behind the modules is to allow disparate (largely Fortran 90) codes to be interfaced, much in the spirit of modern modular codes. Applied Mathematics There has traditionally been a strong link between mathematics, especially applied mathematics, and fusion research. In the past, this link was very strong in areas such as stochasticity and chaotic dynamics; multiscale analysis, asymptotics, and boundary layer theory; and weakly nonlinear dynamics. In recent years, these links have developed in the context of problems without any obvious direct application to fusion energy science, so they fall into the category of work on general plasma physics, which has been relatively poorly supported by the DOE fusion program. The funding for the NSF/DOE program in plasma science, now in its fourth year, supports research in basic plasma science and is a positive development, but the overall funding for this program remains small (see Appendix B for figures). The early links between plasma science and mathematics, which led to significant advances in nonlinear physics (as discussed earlier in this section), are not being maintained despite opportunities for cross-fertilization. Two examples are cited here. First, there has been considerable mathematical interest in the topological structure of magnetic fields. In the fusion context, this interest derived from studies of the confinement properties of magnetic configurations. More recently, it has been realized that field topology is connected to the relaxation and reconnection of magnetic fields, a topic of great importance to fusion science. The sense that fusion researchers are disconnected from this sort of work is reinforced by the fact that important conferences on topological fluid dynamics have not had many participants from the fusion community. A second example is the explosion of work on singularities in continuum systems—for example, How do singularities form in systems described by partial differential equations with smooth initial data? Such problems are of considerable current interest in applied mathematics,

OCR for page 62
Page 75 fluid dynamics, and astrophysics, but the interaction of these fields with fusion research, which has related problems, is again notable by its absence. These two examples are emblematic of the larger problem—namely, that fusion science has become disconnected from many of the advances and activities in modern applied mathematics. A number of areas in mathematics have seen dramatic advances over the past few decades, ranging from modern differential geometry and topology to research in the existence and stability of solutions to partial differential equations. Many of these advances have direct applicability to plasma physics. However, the evidence (based on, for example, the topics listed for discussion at recent meetings of the American Physical Society's Division of Plasma Physics) shows that the interaction of plasma science practitioners with these advances is weak and largely confined to practitioners who are increasingly peripheral to the core fusion research program. Experimental Techniques The fusion program has developed techniques to produce plasmas and to diagnose their properties. Each of these two endeavors has spawned new tools needed for high-temperature plasmas. Production Techniques Using tools developed in the fusion program, plasmas can be produced with temperatures from 104 to 108 K and densities from about 109 to 1015 particles/cm3. Moreover, techniques have been developed to adjust the spatial structure of the plasma—the magnetic field, the plasma temperature, and density profiles. Even aspects of the velocity distribution of particles are adjustable. These techniques have enabled a new area of research—the study of high-temperature plasmas. High-power heating techniques have been developed to deposit megawatts of power into the plasma. Intense beams of neutral atoms (up to about 500 keV in energy) have been developed for this purpose. A wide array of techniques for electromagnetic wave injection has been invented for heating; these employ a huge array of plasma waves spanning the frequency range from tens of kilohertz (where Alfvén waves propagate) to 100 GHz (the electron cyclotron frequency range). A wide assortment of electromagnetic wave sources has been employed or developed (such as high-frequency gyrotrons), as well as a wide assortment of antennas and launching structures. These techniques have also been used to drive current in the plasma and to adjust the current and magnetic field spatial structure. Mega-amperes of plasma current can be driven by these techniques. To fuel the plasma, pellet injectors have been developed that shoot frozen hydrogen pellets into the plasma at high speed. This assortment of plasma production tools has application in many areas, including electromagnetic wave sources, accelerator physics, plasma processing, and beam physics. Diagnostic Tools The challenge in diagnosing hot plasmas is to remotely or nonperturbatively measure a large number of fundamental plasma properties. This has required the invention of new methods as well as the extension of existing techniques to new regimes. Techniques using laser injection into plasmas exploit the plasma effect on the laser scattering, the phase angle, and the electric field polarization. Laser scattering has been developed to determine the electron temperature, electron density, and collective density fluctuations. Laser interferometry is used to determine electron density. Measurement of the Faraday rotation of the laser beam yields information on magnetic field. A wide array of spectroscopic techniques, from the visible to the x-ray spectrum, are used to diagnose impurity ion dynamics. The

OCR for page 62
Page 76 injection of beams of neutral atoms forms the basis for an assortment of measurements, many recently developed. Emission of light from the beam atoms provides a measure of electron density fluctuations. Through a charge exchange interaction between the beam atoms and impurity ions, the impurity ion emission is enhanced, providing a spatially localized diagnosis of impurity ions. The Stark splitting of emission lines from the neutral atoms (produced by the motional electric field experienced by the atoms moving through the magnetic field) yields a measure of the magnetic field in the plasma. Scattering of the beam atoms by protons in the plasma (a Rutherford scattering process) yields information on the temperature and velocity of the protons. Many of these and other diagnostics have been employed to measure detailed aspects of plasma turbulence, as well as equilibrium plasma quantities. The development of novel plasma diagnostics to measure new quantities continues as a vital activity. HAS THE FIELD BEEN RECOGNIZED FOR ASKING AND ANSWERING DEEP PHYSICS QUESTIONS? It is the committee's subjective impression that the fusion science field has not received adequate credit for past contributions to other scientific fields and, in spite of doing quality science (as described in this report), is not widely respected as an area of fundamental physics. There is, moreover, evidence (below and in the preceding section) that an inadequate level of formal and informal interaction is taking place between fusion scientists and the rest of the scientific community. The committee supports its conclusion with the following additional data: Of the 2185 members of the National Academy of Science, 10 members have worked in fusion research; no one in the field has been elected to the Academy for 13 years. This statistic probably reflects the fact that while the early accomplishments of plasma science received widespread recognition, the increasing isolation of fusion science from the rest of the science community has prevented similar recognition of the more recent accomplishments in the field. The National Medal of Science has been awarded to 110 recipients in the physical sciences. Two of the recipients—L. Spitzer and M. Rosenbluth—were specifically rewarded for their fusion-related work. Three recipients—S. Buchsbaum, M. Kruskal, and E. Teller—had worked in plasma physics or fusion at some time in their careers. Between 1960 and 1998, there were 180 winners of the Department of Energy's E.O. Lawrence Award. Fifteen of the awards were fusion-related: 10 for magnetic fusion and 5 for inertial fusion. The American Institute of Physics has published Physics News Update for the past decade. From 1990 through May 2000, the number of physics-related headline items was 1627. Of these, seven (0.4 percent of the total) concerned magnetic fusion, two (0.1 percent) concerned inertial fusion, and four (0.2 percent) concerned general plasma physics. ARE PLASMA SCIENTISTS WELL REPRESENTED AND TRAINED AT THE NATION'S MAJOR RESEARCH UNIVERSITIES? Essential to the health of any scientific field is the training and influx of high-quality students on a continuing basis. The major research universities are essential to the maintenance of this process and therefore to the long-term health of the field. There are disturbing demographic trends in plasma science at the major universities that raise significant doubts about the ability of the field to sustain itself over the long term.

OCR for page 62
Page 77 Information was provided to the committee from an independent survey 1 of the physics and applied physics faculties at 25 top U.S. universities; the survey indicates that 4 percent of the faculty members are in plasma physics, a number that can be compared with the 6 percent of all members of the American Physical Society who belong to its Division of Plasma Physics. Admittedly, the survey could have been usefully expanded to include more universities and departments, since plasma scientists are also to be found in astrophysics, applied mathematics, and various engineering departments. The disturbing result from this survey is, however, that at these 25 universities, out of roughly 1300 physics faculty, there are only three assistant professors in plasma physics, well below the level at which the plasma physics faculty at these institutions can be sustained. These three assistant professors in plasma physics account for less than 2 percent of the assistant professors in the physics departments at these institutions. This could herald the eventual disappearance of plasma physics from physics departments at the nation's top universities. According to the American Institute of Physics' 1997 Graduate Student Report (admittedly already somewhat out of date), 4 percent of graduate students with three or more years of study at Ph.D.-granting physics departments were enrolled in the subfield of plasma and fusion physics, consistent with the fraction of the physics faculty in this subfield. Slightly more recent numbers from the National Research Council on physics and astronomy doctorates by subfield show that 55 out of a total of 1584—or 3.5 percent—were in plasma and high-temperature physics in 1998. Furthermore, students who now graduate with Ph.D. degrees in this subfield are much less likely than before to enter a career in fusion energy science. To illustrate this, about 80 percent of the plasma theory Ph.D. students who graduated during the 1980s from the University of Texas (a major center for fusion and plasma research in the United States) went into fusion positions, whereas only 20 percent of those who graduated during the 1990s did so. This suggests that the pipeline for staffing future plasma physics research is not being filled. One likely consequence of these demographic shifts is that the fusion field will be increasingly underrepresented at universities, with the demographic balance of Ph.D. scientists shifting to the national laboratories. High-quality fusion science certainly has been and is being done at the national laboratories, and it should be borne in mind that good science can be done in large projects as well as in smaller, university-scale projects. Furthermore, there are many examples of collaborations among fusion scientists at national laboratories and universities, and more such interactions could be encouraged. Nevertheless, the survival of plasma science as a viable field is dependent on maintaining its strength in the university environment. CREATION OF CENTERS OF EXCELLENCE IN PLASMA AND FUSION SCIENCE The relatively small number of assistant professors at the large research universities forces the committee to conclude that the field of plasma science is held in low esteem in the academic community at large, an opinion that ripples out to influence students, news media, and policymakers. Given the committee's view that the science coming out of the program is of the highest quality, this is a disturbing conclusion and one for which remedies must be devised and implemented. While the budget cuts in the program that occurred in FY96 may partially explain the trend, the committee feels that there are deeper problems in the program. Two issues have emerged: (1) the relative isolation of the field—scientists outside the program are apparently not aware of the level of scientific discovery in the program—and (2) the meager degree to which science issues are driving the 1 Survey study conducted by Kenneth Gentle, University of Texas at Austin.

OCR for page 62
Page 78 programmatic progress toward fusion. The lack of recognition on the part of outside scientists is reflected in the relatively small number of accolades received by plasma scientists, as discussed previously. The effective communication of the intrinsically interesting and broadly important science being uncovered in the program is essential if the program is to successfully refurbish its image with scientists and policymakers outside the field. Communication must be enhanced on a broad front. The DOE is actively supporting a distinguished speakers program through its Division of Plasma Physics, a positive but very small step. A fundamental issue, discussed extensively in Chapter 3, concerns the organization of experimental concepts in terms of their progress toward a reactor rather than their progress in addressing important scientific goals. Given the tendency of the program to emphasize performance issues, it is perhaps not surprising that the external community has not fully appreciated the scientific accomplishments of the program. The reorganization of the program around scientific issues would more effectively communicate the science in the program and enhance the ability of the program to address the cross-cutting science issues more effectively. One way of generating interest and excitement in a field is to offer funding opportunities that will attract the best talent. In a flat funding situation, the maintenance of a continuing flow of talent into the field will be challenging but nevertheless essential to the health of the field—the challenge is to provide appropriate security for productive scientists while at the same time attracting individuals with new ideas and enthusiasm into the field. The committee concludes that the main research groups have remained far too static for the health of the field. To illustrate this point, the list of universities that have major theory groups has not changed in more than 20 years: the University of Maryland, New York University, the Massachusetts Institute of Technology, the University of Wisconsin, the University of Texas at Austin, the University of California at Los Angeles, the University of California at San Diego (UCSD), Cornell University, and the University of California at Berkeley. The strength of the plasma physics programs at Berkeley and Cornell has gone down slightly, while that of the program at UCSD has grown. While the work at all these places is excellent, the committee is concerned that new groups have not spontaneously formed and grown with time. The formation of strong new groups, while not the sole measure of excellence, certainly reflects the success of a field in attracting talent and in making the case that it is ripe for identifying and exploring new science issues. To address a number of the science issues discussed in Chapter 2 and Chapter 3 and the programmatic issues discussed in this chapter, the committee recommends the creation of several interdisciplinary centers (see below). On the scientific side, many of the issues in the fusion program are now sufficiently complex that they require closely interacting, critical-mass groups of scientists to make progress. For example, understanding the dynamics of plasma turbulence and transport requires the development of appropriate physical models; computational algorithms for treating disparate space- and timescales, as well as complex magnetic geometries; efficient programming on massively parallel computing platforms; and an understanding of nonlinear physics (energy cascades, intermittency, phase transitions, avalanches). Tight coupling with a parallel experimental effort is required to challenge theoretical predictions. No single scientist or small group of practicing scientists has the breadth of knowledge required to tackle such large and complex problems. In the area of theory and computation, the absence of closely interacting teams of critical mass is inhibiting a concerted attack on a number of central science issues confronting the fusion research program. The loose collaborations established by the program from time to time have not succeeded in fostering the close working relationships required to address the most challenging topics. New “centers of excellence,” each of which would either serve as the node for a distributed network of collaborators or undertake scientific explorations of significant magnitude at one site or would do

OCR for page 62
Page 79 both, could create a new focus on scientific issues within the U.S. fusion program. (See discussion and recommendations in Chapter 2 and Chapter 3.) Potential focus topics for such centers include turbulence and transport, magnetic reconnection, energetic particle dynamics, and materials; other topics would emerge in a widely advertised proposal process. Because the topics listed above have such broad scientific applicability in allied fields, collaborations could be set up with scientists having expertise of great value to the plasma science and fusion research effort. An explicit goal of the centers should be to convey important scientific results to both the fusion community and the broader scientific community. The mere announcement of opportunity for fusion centers of excellence would signal to the broader scientific community the fusion community's intent to significantly bolster the scientific strength of the field. FINDINGS AND RECOMMENDATIONS Findings 1. In the key scientific areas being discussed, there is a history of intellectual exchange between the plasma physics community and the broader scientific community in areas such as MHD, nonlinear dynamics, instabilities, and transport. 2. The current fusion program is relatively weakly coupled to programs in the rest of the physics community. Most of its external coupling takes place in areas such as space physics and (plasma) astrophysics, which are themselves poorly represented in physics departments. Believing that the education of all graduate students in physics should include a course in plasma physics, the committee is distressed by the small percentage (roughly half) of departments with faculty in plasma physics. 3. The future representation of plasma science at universities is threatened by an apparent lack of new blood. Of a total physics faculty of roughly 1300 individuals at the top 25 physics research universities, there are only 3 assistant professors of plasma physics. The committee is concerned that the very low overall rate of replacement of plasma physics faculty threatens the future of this field. 4. In the area of theory and computation, the dearth of critical-mass, closely interacting teams of researchers from different institutions is inhibiting a concerted attack on a number of important science issues faced by the fusion research program. The loose collaborations established from time to time by the program have generally not succeeded in nurturing the close working relationships required to address the most challenging topics. No single scientist or small group of practicing scientists has the breadth of knowledge required to tackle such large and complex problems. Recommendations A systematic effort to reduce the scientific isolation of the fusion research community from the rest of the scientific community is urgently needed. Program planning, funding, and administration should all encourage connectivity with the broad scientific community. The community of fusion scientists should make a special effort to communicate its concepts, methods, tools, and results to the wider world of science, which is largely unaware of that community's recent scientific accomplishments. There are numerous examples in federally funded research programs of formal coordination mechanisms having been established among related programs in different agencies. In some instances this coordination can optimize the use of funding. Perhaps more significant, dialogue among the leaders of

OCR for page 62
Page 80 these government research programs can encourage interactions among the various scientific communities, foster joint undertakings, and raise the visibility of the discipline as a whole. The fusion science program should be broadened both in terms of its institutional base and its reach into the wider scientific community; it should also be open to evolution in its content and structure as it strengthens its research portfolio. Clearly, this issue can be approached in a number of ways. One good way would be to set up competitive funding opportunities of sufficient magnitude to elicit responses from potential new institutional participants. The creation of centers of excellence in fusion science (proposed below) and the greater involvement of the National Science Foundation in fusion and plasma science are other ways to broaden the institutional base of fusion science. A larger proportion of fusion funding could be made available through open, well-advertised, competitive, peer-reviewed solicitations for proposals. Fusion program peer review could involve scientists from outside the fusion community where appropriate. The evaluation and ranking of proposals by panels that include individuals with appropriate expertise in allied fields would broaden the intellectual reach of the grant review process. Plasma science research not immediately related to the quest for a practical source of power from fusion (the fusion energy goal) should play a more influential role in the fusion program, at an appropriate budget level. This would include funding for general plasma science and for peer-reviewed individual investigator grants, each presently a small fraction of the total fusion energy science program (see Appendix B). Such funding would encourage new interchanges that enrich fusion science. To ensure that increasing institutional diversity is a continuing goal, the committee recommends that the breadth and flexibility of participation in the fusion energy science program should be a program metric. Several new centers, selected through a competitive peer-review process and devoted to exploring the frontiers of fusion science, are needed for both scientific and institutional reasons. Many of the issues in fusion science are now of sufficient complexity that they require closely interacting, critical-mass groups of scientists to make progress. For example, understanding the dynamics of plasma turbulence and transport requires the development of appropriate physical models; computational algorithms for treating disparate space- and timescales, as well as complex magnetic geometries; efficient programming on massively parallel computing platforms; and an understanding of nonlinear physics (energy cascades, intermittency, phase transitions, avalanches). No single scientist and no small collaboration of practicing scientists has the breadth of knowledge required to tackle such large and complex problems. The centers of excellence could create a new focus on scientific issues for the U.S. fusion program. A center could serve as a node for a distributed network of close collaborators or it could undertake scientific explorations of significant magnitude, or it could do both. The centers would combine the expertise and approaches of national laboratories on the one hand and universities on the other. They should have a number of programmatic and structural features that will let them play an appropriate role in addressing the critical problems of the field: A proposal for a center should have a plan to identify, pose, and answer scientific questions whose importance is widely recognized.

OCR for page 62
Page 81 One size cannot meet all scientific challenges. The committee envisions a center comparable in size to current NSF-sponsored centers (with operating costs of $1 million to $5 million per year), although the size should ultimately be determined by the proposal process. Some centers may need on-site experimental facilities and some may need only computing facilities and access to larger national computer centers. A team of between four and six coinvestigators with broad expertise and connections to other research groups and laboratories should form the core of the center's personnel. This team should be augmented by a similar number of temporary research staff (funded, at least in part, by the center) and an appropriate number of support staff. The center should enable links to various scientific disciplines, including physics, mathematics, and computer science, depending on the problem it is focusing on. It should have a plan for bringing practitioners of other disciplines, from other institutions, into the fusion community and should make the community's experimental resources more widely available. The institutions housing or participating in such centers should make a commitment to add faculty or permanent research staff, as appropriate, in plasma/fusion science and/or related areas. The centers should have a strong educational component, featuring outreach to local high schools, undergraduate research opportunities, and a graduate research program. Centers should sponsor multidisciplinary workshops and summer schools focused on their central problem that will bring together students and researchers from various fields and institutions. The workshops would aim to bring in new ideas and collaborators as well as to disseminate to other fields the results they are achieving as they address the fundamental problems of fusion science. Potential focus topics for centers include turbulence and transport, magnetic reconnection, energetic particle dynamics, and materials; other topics would emerge in a widely advertised proposal process. Topics such as these are of broad scientific interest in allied fields. To build another bridge to allied fields, the DOE should cooperate with NSF in establishing one or more centers addressing a topic of general interest in plasma science. The DOE/NSF centers should have as a central objective establishing collaborations with scientists possessing expertise of value to the plasma science and fusion research effort. An explicit goal of the centers should be to convey important scientific results to the broader scientific community as well as the fusion community. Even just an announcement of opportunity for fusion centers of excellence would signal to the broader scientific community the fusion community's intent to significantly bolster the scientific strength of the field. It would be highly desirable for other federal agencies, particularly the NSF, to collaborate in one or more fusion centers for reasons of disciplinary and institutional diversity as well as to obtain the benefits of interagency collaboration. However, DOE should play a lead role in these centers, not only for reasons of administrative clarity but also because its leadership will ensure that the technical capabilities of the fusion energy science community are made available to new participants. DOE leadership will also ensure that progress at the centers is communicated throughout the fusion community and translated into DOE program plans, to hasten the progress toward the fusion energy goal. The procedure for awarding grants for fusion centers of excellence could do much to remedy the isolation of the fusion science community by ensuring that the broader scientific community participates in the institution-building effort. The selection process for the centers should feature open, competitive peer review employing clear, science-based selection criteria, as outlined above. The committee believes this recommendation to be critical enough to the new science-based approach to fusion energy that ways should be found to fund a first center even in a level budget scenario. The success of the competition and the quality of the first center should guide the decision whether to launch

OCR for page 62
Page 82 a second or even a third center. In other programs, such centers have been effective mechanisms for strengthening the breadth and depth of a broad scientific area. In the committee's view, there is a very strong argument for expanding program funding to give fusion centers of excellence a strong and durable foundation. The National Science Foundation should play a role in extending the reach of fusion science and in sponsoring general plasma science. The mission of OFES, following the restructuring of the program in 1996, has been to establish the knowledge base in plasma physics required for fusion energy, with the result that a substantial number of plasma science issues are being explored within the fusion regime that also have applicability to allied fields such as astrophysics. For this reason, the committee believes that NSF should begin to play a larger role in the solution of these basic plasma science issues. The involvement of NSF could have an intellectual impact on basic plasma science similar to that which it has had on basic research in other scientific disciplines where mission agencies like DOE play the main funding role. NSF involvement would facilitate linkage to other fields and the involvement of new scientists in the program. Recently, NSF and DOE collaborated on a small but highly effective program to encourage small-laboratory plasma experiments and the theoretical exploration of topics in general plasma science. The large number of proposals submitted to this program is an indication of the need for it. The rationale for the expansion of research in general plasma science was well articulated in the NRC's Plasma Science: From Fundamental Research to Technological Applications, National Academy Press, Washington, D.C., 1995. The NSF/DOE plasma science initiative, if operated at a dollar level closer to that contemplated in the Plasma Science report (an additional $15 million per year for basic experiments in plasma science), can serve more than one important function: Stimulating research on broad issues in plasma science that have potential applications to fusion and Enhancing interagency cooperation and cultural exchange on the approaches used by the two agencies for defining program opportunities, disseminating information on research results to the scientific community, selecting awardees, and judging the outcomes of grants. The optimal process for this partnership, if there is sufficient funding (as requested in the Plasma Science report), would be an annual solicitation of requests for proposals (RFPs). In particular, this frequency would give new Ph.D.s the chance to enter the field and stay in it, since new Ph.D.s are produced by degree-granting institutions each year and new graduate students enter school each year. Another limitation of the ongoing NSF/DOE program in basic plasma science is the absence of any provision for modest experiments in the $1 million per year class. Historically, neither DOE nor NSF has funded plasma science experiments of this scale. For this reason, the committee recommends a cooperative NSF/DOE effort to broaden the scientific and institutional reach of fusion and plasma research to obtain valuable scientific results. Increased NSF funding and a stronger focus on fusion as well as plasma science within NSF would be required. As discussed in the preceding recommendation (recommendation 4 in the Executive Summary ), NSF could cosponsor one or more centers of excellence in fusion and plasma science.