3
Technological Challenges and Opportunities for Developing Higher Fields

The design and construction of magnets that operate at high field is an art form that requires the balancing of many conflicting requirements. This chapter starts with a discussion of the nature and magnitude of the challenges that must be met and concludes with a description of the opportunities that now exist for improving magnet performance. A common thread in this discussion is the distinction between structural problems that limit what can be done and intrinsic material properties that often have not yet been exceeded, suggesting that it should be possible to build magnets that operate at fields significantly higher than any available today.

WHAT IS THE CHALLENGE?

Four different types of electromagnets are used to generate high fields: (1) resistive DC magnets, (2) (resistive) pulsed-field magnets, (3) superconducting DC magnets, and (4) hybrid magnets, combining both resistive and superconducting elements. The maximum fields obtainable from magnets of all four types have increased significantly over the past several decades, but the increases have come in increments of 10-20 percent rather than in large jumps (such as by factors of 10 or more), and it is important to understand why this is likely to remain the case (Figure 3.1 shows how record field strengths have increased with time).

The reason advances in maximum field strength are hard to obtain is that, everything else being equal, both the energy stored in the field of a magnet and the



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 69
Opportunities in High Magnetic Field Science 3 Technological Challenges and Opportunities for Developing Higher Fields The design and construction of magnets that operate at high field is an art form that requires the balancing of many conflicting requirements. This chapter starts with a discussion of the nature and magnitude of the challenges that must be met and concludes with a description of the opportunities that now exist for improving magnet performance. A common thread in this discussion is the distinction between structural problems that limit what can be done and intrinsic material properties that often have not yet been exceeded, suggesting that it should be possible to build magnets that operate at fields significantly higher than any available today. WHAT IS THE CHALLENGE? Four different types of electromagnets are used to generate high fields: (1) resistive DC magnets, (2) (resistive) pulsed-field magnets, (3) superconducting DC magnets, and (4) hybrid magnets, combining both resistive and superconducting elements. The maximum fields obtainable from magnets of all four types have increased significantly over the past several decades, but the increases have come in increments of 10-20 percent rather than in large jumps (such as by factors of 10 or more), and it is important to understand why this is likely to remain the case (Figure 3.1 shows how record field strengths have increased with time). The reason advances in maximum field strength are hard to obtain is that, everything else being equal, both the energy stored in the field of a magnet and the

OCR for page 69
Opportunities in High Magnetic Field Science FIGURE 3.1 Available maximum field strength for nondestructive magnets at facilities around the world as a function of year. The maximum fields delivered by different types of magnets are indicated separately. Figure courtesy of G. Boebinger, National High Magnetic Field Laboratory. stresses it experiences increase as the square of the field strength. Thus, the higher the field at which a magnet operates, the more complicated the stress and energy management techniques that must be used to ensure its physical integrity and the safety of those who work with it. All aspects of magnet design are affected, including the choice of materials for electrical conductors and structural components,

OCR for page 69
Opportunities in High Magnetic Field Science the electrical insulation systems employed, and so on. In the case of superconducting magnets, field stability and quench protection pose additional issues. The magnitudes of the energies and stresses that must be confronted are best illustrated with some numerical examples. The energy stored in the field of the highest-field magnet for high-resolution solution NMR now operating, which is the 21.6-T superconducting NMR magnet at the National Institute for Materials Science (Japan), is 33 MJ. This is roughly the kinetic energy of a 40-metric-ton railroad locomotive moving at 150 kilometers per hour, and the magnet is designed so that it will dissipate all of that energy harmlessly in the event the windings of its magnet become nonsuperconducting for any reason. Magnetic stresses are easiest to work out for long solenoidal magnets, because the mechanical forces generated by the interaction of the current in the windings of a solenoid with its own magnetic field are similar to an internal hydrostatic pressure. If the central magnetic field is 6 T, the equivalent pressure is 140 atm, which is about the working pressure of the typical gas cylinder. If the central field is 10 T, the internal pressure is about 400 atm, which is approximately the yield strength of annealed copper at room temperature. A 1-GHz NMR magnet (23.5 T) must withstand the equivalent of an internal pressure of over 2,000 atm and thus must be built of materials having yield strengths significantly greater than that of copper. Similarly large stresses must be dealt with in nonsolenoidal high-field magnets. The magnetic field of a high-field magnet is generated only by the current-carrying conductor in its coil, but the coil necessarily includes other components. For example, none of the materials used as conductors in high-field magnets today is strong enough to withstand the stresses generated, so all magnets include mechanical support systems, which contribute nothing to the field. Electrical insulation and a thermal management system are also required. The higher the field to be generated, the more non-current-carrying material a coil must include and the lower its bulk-average current density. However, the lower the bulk-average current density, the bigger (and more expensive) the magnet must be to deliver a specified field. Given these facts, it should not be surprising that the inner bores of magnets tend to fall as the maximum field at the windings increases (see Figure 3.2), which reduces the amount of field energy that must be managed. Further, as just explained, the overall winding pack current density also tends to fall as stored energy rises (see Figure 3.3). (Table 3.1 describes the magnets shown in Figures 3.2 and 3.3.) Other issues that designers must consider include joints for both resistive and superconducting magnets and AC loss, stability, and quench problems for superconducting magnets. Finally, it is generally the case that no single aspect of a high-field magnet design can be altered independent of the others, so even incremental improvements in magnet performance must be hard fought for.

OCR for page 69
Opportunities in High Magnetic Field Science FIGURE 3.2 The stresses in solenoid scale with the square of the magnetic field, limiting the highest field magnets to small bore sizes and sample volumes. See Table 3.1 for legend. Figure courtesy of J.R. Miller, National High Magnetic Field Laboratory. FIGURE 3.3 Overall winding pack current density decreases as the stored energy increases because increased amounts of structure, insulation, and thermal management are required to limit stresses and protect the magnet. See Table 3.1 for legend. Figure courtesy of J.R. Miller, National High Magnetic Field Laboratory.

OCR for page 69
Opportunities in High Magnetic Field Science TABLE 3.1 Description and Location of Magnets in Figures 3.2 and 3.3. Item Location Description/Intended Use T1a Japan DC solenoid/conductor testing T2a Japan DC solenoid/conductor testing T4a Japan DC solenoid/conductor testing T5a Japan DC split solenoid/conductor testing T6a Japan DC solenoid/conductor testing T11a Japan DC pancake module/superconducting motors and energy storage (SMES) development T13a Japan DC solenoid/SMES development T14a Japan Pulsed pancake module/SMES development NRIM 40 T HOb Tsukuba, Japan DC solenoid/research facility NHMFL 45 T HO Tallahassee, Florida DC solenoid/research facility FENIXc Livermore, California DC split solenoid/conductor testing ITER CS model coild Japan Atomic Energy Research Institute Pulsed solenoid/technology demonstration NRIM 21 Te Tsukuba, Japan DC solenoid/research facility NRIM 900 MHzf Tsukuba, Japan DC solenoid/research facility NOTE: Includes a historical sample of foreign and domestic magnets. NRIM, National Research Institute for Metals; NHMFL, National High Magnetic Field Laboratory. aO. Tsukamoto, S. Torii, T. Takao, N. Amemiya, S. Fukui, T. Hoshino, A. Ishiyama, A.Ninomiya, H. Yamaguchi, and T. Satow, Recent technical trends of superconducting magnets in Japan., IEEE Transactions on Applied Superconductivity 9(2), 547 (1999). bH. Morita and S. Ito, Development of a 40 T Class Hybrid Magnet, High Magnetic Fields: Applications, Generation, and Materials, H. Schneider-Muntau, ed., Singapore, World Scientific, Singapore, 1997. cD.S. Slack, R.E. Patrick, and J.R. Miller, FENIX: A test facility for ITER and other new superconducting magnets, IEEE Transactions on Magnetics 27(2), 1835 (1991). dH. Tsuji, JAERI, Private communication, April 19, 2000. eR. Hirose, T. Kamikado, O. Ozaki, M. Yoshikawa, T. Hase, M. Shimida, and Y. Kawate, 21.7 T superconducting magnet using (Nb,Ti)3Sn conductor with 14%-Sin Bronze, Proceedings of the 15th International Conference on Magnet Technology, Beijing, China, October 20-24, 1997. fT. Kiyoshi, A. Sato, H. Wada, S. Hayashi, M. Shimada, and Y. Kawate, Development of 1 GHz superconducting NMR magnet at TML/NRIM, IEEE Transactions Applied Superconductivity 9(2), 559 (1999) and T. Kiyoshi, NRIM, private communication, March 6, 2000. SOURCE: J.R. Miller, NHMFL. RESISTIVE DC MAGNETS Resistive DC electromagnets have been in use since the first half of the 19th century, and they can be as simple as a solenoid made of insulated copper wire. In addition to the force and energy problems described above, high-field resistive magnets present two unique challenges. First, they consume electric power in large quantities and hence need large DC power supplies. Second, most of the power they consume is converted into heat; thermal destruction is a serious problem.

OCR for page 69
Opportunities in High Magnetic Field Science Present Status The most powerful resistive DC magnets operating today are magnets of a modified Bitter design that generate fields up to 33 T (see Figure 3.4). The field strengths of these magnets are limited primarily by power availability and by cooling issues, but also to some degree by the mechanical strength of materials in the assembly. For economic reasons, high-field DC magnets are found only at large FIGURE 3.4 Typical plate from a (resistive) Bitter magnet. The small slots are for the circulation of cooling water. The ring of large holes is for axial prestress and accommodates structural tie rods. Image courtesy of National High Magnetic Field Laboratory.

OCR for page 69
Opportunities in High Magnetic Field Science facilities that have the financial means to build and operate the associated power supplies and cooling systems, as well as meet the other operating and maintenance costs of these magnets, which are substantial. To control energy costs, resistive DC magnets are energized only when experiments are conducted with them and thus are routinely cycled between their energized and unenergized states. This cycling leads to fatigue, limiting their operating lifetimes, which are typically 1,000 to 2,000 hours. It is reasonable to anticipate that resistive DC magnets that produce fields significantly higher than 33 T will eventually be built. Before this happens, however, new materials will have to be developed that combine high mechanical strength and low resistivity, and new structural designs may be required. Outlook for Resistive DC Magnets As described above, the difficulty of supplying sufficient power and cooling to a resistive electromagnet severely limits the development of higher-field resistive DC magnets. The physical strengths of the materials used within the magnet are also near their limits. Nevertheless, resistive DC magnets will continue to be important because record steady-state fields can be achieved by pairing resistive magnets with superconducting magnets. The Tallahassee site of the NHMFL is where most of the work on resistive magents is now being done in the United States. The objective of this work is to improve the field strength of the resistive outsert of NHMFL’s 45-T hybrid magnet (see the section “Pulsed Magnets”). PULSED MAGNETS Pulsed magnets have long played an important role in high-field science, primarily because they can generate much stronger magnetic fields than DC magnets. However, they do so for only short periods of time (10 to 100 ms), and they tend to have very small bores (10-20 mm). Pulsed magnets are resistive magnets that take advantage of the capacity of materials and power systems to withstand extreme conditions transiently. Figure 3.5 is a plot of the peak magnetic field produced by the pulsed-field magnets now operating around the world versus the duration of the field pulse generated (see Appendix B for more information). By providing access to fields far higher than those accessible using steady-state magnets, pulsed magnets offer opportunities for new discoveries that can take research in new directions; detailed follow-up with steady-state fields is often required, however. The 60-T long-pulse magnet at NHMFL at Los Alamos National Laboratory (LANL) is a good example of a device of this type (Figure 3.6 shows its (massive) power supply). It produced field pulses having a strength of 60 T for 100 ms in a

OCR for page 69
Opportunities in High Magnetic Field Science FIGURE 3.5 The field strength and pulse duration times of representative nondestructive pulsed magnets from facilities around the world (see Appendix B). 32-mm bore before failing prematurely in 2000. Pulsed magnets are often pre-cooled to cryogenic temperatures using liquid nitrogen and then energized with a pulse of electric current obtained from a large capacitor bank. Pulse length is limited by the heat capacity of the magnet and by the adiabatic temperature rise of its windings. The generation of high pulsed fields requires advanced materials for conductors and careful attention to stress management. The ideal conductor would combine high mechanical strength with high electrical conductivity and high heat capacity, properties that are unhappily in conflict, so compromises must be made. Whatever the conductor used, magnets are usually built up in nested layers, each designed to withstand its own radial force by hoop tension and each often reinforced structurally by shells, bands, or fibers of nonconducting or low conductivity material. Design parameters for each layer, such as conductor size, material, current density, and reinforcement, are all varied to optimize magnet performance. Pulsed magnets that operate in the 50- to 60-T range are relatively common, but only a few facilities have, or plan to build, magnets that produce fields that are

OCR for page 69
Opportunities in High Magnetic Field Science much higher, because 60 T is the field strength at which the mechanical strength of the materials used becomes limiting. Nevertheless, pulsed magnets have been built that achieve 70 T, but their pulse lifetimes are extremely short. As this report is being written, two institutions (NHMFL at Los Alamos and the High Field Laboratory at Dresden, Germany) are developing 100-T, multishot pulsed magnets. One approach is to build a magnet consisting of a large outer coil that produces a long pulse with a peak field about 55 T and a small inner coil that generates a much shorter 45-T pulse in synchrony. Fields substantially in excess of 100 T can be produced by pulsed methods. They are obtained by discharging large amounts of electric current through single-turn coils. The resulting magnetic field pulse ends when the coil vaporizes. Because the plasma so created is driven outwards by magnetic forces, the experimental sample may survive. FIGURE 3.6 The 1,430-MVA electric generator used by the NHMFL-LANL for pulsing very high magnetic field resistive magnets. Photo by K.N. Roark, Los Alamos National Laboratory.

OCR for page 69
Opportunities in High Magnetic Field Science Not surprisingly, multishot pulsed magnets suffer from cycling-induced mechanical and electrical fatigue. To alleviate these problems, some designs allow for limited plastic deformation of the windings. Electrical insulating materials must withstand cycles of high voltages and mechanical stress. Discontinuities in windings, structures, and insulation systems create mechanical and electrical stress concentrations where failure can occur. They are usually found at transition regions in windings, joints between subcoils, and at coil terminals. Given the magnitude of the energy stored in these magnets when they are fully energized, their end-of-life failure modes are often sudden and highly destructive (see Figure 3.7). Protection of users and surrounding equipment is an important consideration for this class of magnets. Advantages and Disadvantages Pulsed magnets are in widespread use today, especially outside the United States. The rapid development of instrumentation in recent years, notably in high-speed electronics and optical detector arrays, has greatly increased the range of work that can be undertaken in pulsed fields. Some topics in semiconductor physics are particularly well suited to such measurements since, due to the low concentrations of free carriers, there are no particularly severe problems associated with sample heating. Additionally, advances in the speed and miniaturization of electronics have increased the range of experiments that can be usefully performed using pulsed-field magnets. However, in general, experiments at the limits of sensitivity and high resolution can only be performed using steady or slowly varying magnetic fields. As developments in technology make measurement instrumentation smaller and faster, though, pulsed fields will continue to offer a valuable alternative. In fact, 16 of the 28 foreign high-field facilities surveyed in Table B.1 in Appendix B provide only pulsed-field magnets for their users, and only 2 of them employ DC magnets exclusively. There are many reasons for this bias. Advantages Pulsed-field magnets are cheaper to build and operate than high-field DC-magnets. For fields in the 30-T range, where there is overlap with what can be obtained from resistive DC magnets, pulsed magnets are generally much cheaper to construct and operate largely because they require much less infrastructure (power supplies and cooling facilities).1 For facilities that operate many pulsed 1   The experience of C.C. Agosta and his students at Clark University provides an example, albeit extreme. They recently built a pulsed magnet that delivers a peak field of 50 T with a rise time of 12 ms. The complete system, including the power supply, cost less than $150,000.

OCR for page 69
Opportunities in High Magnetic Field Science FIGURE 3.7 The fatigue stresses in high-field pulsed magnets often lead to dramatic failures. Images courtesy of National High Magnetic Field Laboratory.

OCR for page 69
Opportunities in High Magnetic Field Science not well understood and thus may emerge as a better low-temperature, high-field conductor than Bi-2223. In the long term, both substances may be replaced by biaxially textured YBCO, which has a much more favorable Hirr(T) curve (see Figure 3.10), lower anisotropy, and higher Jc (see Figure 3.8) and does not require Ag coating to be an effective conductor. Materials Occasionally Used Many other materials have been found that will superconduct under certain conditions and that might be used for magnet construction—for example, ternary alloys such as Nb-Ti-Ta, which is being studied as a possibility for accelerator magnets. Rather than discussing them all here, we will mention only one, the binary compound Nb3Al. Nb3Al has a number of potential advantages over Nb3Sn, not least that the stoichiometric compound has a slightly higher Tc and an Hc2 that can reach 40 T. Its greatest shortcoming is that it does not easily form when it contains 25 wt% Al, such that when quenched from the high temperatures at which the stoichiometric composition is stable, it forms a disordered structure in which both Tc and Hc2 are lower. An important advantage of Nb3Al is that the strain sensitivity of its Jc, Tc, and Hc2 parameters is much less than that of Nb3Sn. There are two principal approaches to making Nb3Al conductors. One uses jelly rolls of Nb and Al that are extruded and drawn into wire before being reacted at 700 to 900°C to form the A15 phase. The conductor that results has a Tc of about 16 K and an Hc2 of 20 to 24 T, which are both significantly lower than for competing Nb3Sn conductors. Because the properties of Nb3Al are little affected by strain, conductors made of it have been used in prototype coils for large fusion magnets. The second approach, which was pioneered in Japan, is to rapidly heat Nb-Al precursors to about 2000°C and then quench them in a liquid Ga bath. This maintains a ductile Nb-Al solid solution phase that can be then selectively crystallized into an off-stoichiometric but still high Jc and high Hc2 phase at lower temperatures. Some examples of the high Jc values that can be obtained by such a process are shown in Figure 3.8. It is, in fact, possible to find a range of magnetic field above ~20 T where Nb3Al is superior to Nb3Sn, and Japanese researchers are pursuing the use of such conductors in high-field NMR magnets. A variety of problems remain to be resolved. Among them are the viability and costs of such a complex process and finding a means of applying sufficient stabilizing normal metal—e.g., Cu or Au—to the conductor.

OCR for page 69
Opportunities in High Magnetic Field Science Emerging Superconducting Materials While the normal and superconducting states of HTS materials remain incompletely understood, work on their practical application continues to advance. In fact, in many applications, HTS materials exhibit surprisingly conventional behavior. The property of HTS materials that has presented the greatest barrier to their practical use is the intrinsic weakness of the links at grain boundaries. This weakness makes the supercurrent density at grain boundaries less than within grains, reducing the overall supercurrent in conductors made of polycrystals. This problem can be alleviated by using epitaxial films in which grains are aligned in the plane of the substrate. Indeed, HTS epitaxial films are used in the wireless communications filter systems deployed today, with a few thousand such systems having been sold. For applications requiring a conductor (e.g., magnets, power transmission, energy storage, motors, and generators), tapes and wires of Bi-based HTS conductors are used, and the performance of these conductors has steadily improved. More recently, attention has focused on the development of tapes that consist of a thick film of YBCO deposited on ribbon substrates in which the grains have been aligned by various means so that grain boundary misorientation angles are small. Finally, discoveries about the superconducting properties of MgB2 have catapulted it to the fore. These useful developments notwithstanding, the industrial application of HTS materials is hampered by other physical issues, one being the need for expensive cooling systems to maintain them in the superconducting state and another being their brittleness. Bi-2212 Bi2Sr2CaCu2O8-x is a two-CuO2-layer version of Bi-2223. It shares many of the features of the latter compound but has enough unique advantages so that it is still produced commercially in small quantities. One of its advantages is that it can be made as a round wire with high Jc. It appears that considerable self-alignment of Bi-2212 grains occurs during the partial Bi-powder precursor melt step in the manufacturing sequence, which is required to obtain a high Jc. Indeed, the Jc values of Bi-2212 round wires considerably exceed those of Bi-2223 wires. As Figure 3.8 shows, the field dependence of Jc for this material is very weak up to 30 T at 4.2 K. Indeed the very high field limits of this material, which are only poorly understood, are probably well above 50 T at 4 K. This is a direct consequence of the excellent connectivity of Bi-2212 conductors. A typical round wire conductor is shown in Figure 3.9c. Some manufacturing problems remain to be worked out. For example, the temperature window for optimum processing is very tight, perhaps as little as 5°C

OCR for page 69
Opportunities in High Magnetic Field Science at 890°C. This makes the wind and react process that has proven so successful for Nb3Sn magnets more difficult to execute for Bi-2212. Also, like all conductor materials except Nb-Ti, Bi-2212 is brittle and, even when sheathed in Ag or Ag alloy, not very strong. The 25-T superconducting magnet made by an Oxford/NHMFL team in 2003, which was mentioned earlier, demonstrates the application of Bi-2212 to small-bore solenoids. This technology is likely to make it possible to construct superconducting magnets that operate at 30 T. However, the large electronic anisotropy and strong flux creep of Bi-2212 make it unlikely to be a useful, persistent-mode material and may limit its use to driven magnets or to solid-state NMR. Builders of HEP magnets are also interested in Bi-2212 because the round shape of the wire makes it possible to cable 20-30 such wires into multikiloampere conductors suitable for accelerator dipole or quadrupole magnets. YBCO-Coated Conductors One key incentive for the study of YBCO-coated conductors is their exceptionally high Jc value, as shown in Figure 3.8. These values reflect the fact that biaxially textured conductors can be made of YBCO, in which there is almost no grain-boundary obstruction of current. The inherently low electronic anisotropy of YBCO (from 5 to 7) produces excellent flux pinning, giving it strong advantages in superconductor current density over any other HTS compound, including the Bi-based conductors. Prototypes of tape conductors made of this material are now in advanced development by several companies worldwide, suggesting that effective commercialization is imminent (see Figure 3.11). Complicating any evaluation of the prospects for YBCO-coated conductors is the fact that coated conductors are fundamentally different from any other conductors used for magnet construction (compare Figures 3.8 and 3.10). In contrast to all earlier conductors, which are made by conventional metalworking processes such as extrusion and wire drawing, a coated conductor is made by sequential deposition of a multilayer oxide buffer, which is interposed between the Ni-alloy substrate and the YBCO. Purpose-built production lines of a new kind will therefore be needed. An additional issue with coated conductors is that only about 1 percent of the cross section of such wires is superconducting. Thus, overall conductor current densities are heavily diluted compared with competing conductors, where the superconductor fraction is typically 25 to 50 percent. However, the development of coated conductors is proceeding rapidly, such that prototype lengths have overall current densities that are competitive with the Bi-2223 conductors that they are designed to replace, even with such small fill factors. It is clearly encouraging that a very high current density, very high Hc2 conductor is possible

OCR for page 69
Opportunities in High Magnetic Field Science FIGURE 3.11 Illustrative longitudinal cross-sectional architecture of a 165-µm thick, neutral-axis, YBCO-coated conductor showing the copper stabilizer, the solder joining the Cu to the YBCO, and the Ni5W-textured template (a nickel-tungsten alloy with tungsten at 0.5 percent). The actual YBCO superconductor layer is about 1 µm thick and thus represents less than 1 percent of the cross section, in comparison with competing Nb-base, Bi-base, and MgB2 conductors, in which 25 to 50 percent of the cross section is superconductor. Image courtesy of American Superconductor Corporation. with YBCO. As lengths longer than the present 10- to 100-m prototype lengths appear, it will become possible to make test coils that can be used to assess the true promise of YBCO-coated conductors for high-field magnets. Magnesium Diboride The most recent entry into this field is MgB2, an apparently simple binary compound that was discovered to be superconducting only in 2001. Its potential application space is indicated in Figure 3.10, which shows the transition temperature of MgB2 to be about twice that of Nb3Sn (35-39 K vs. 16-18 K). It can be alloyed so as to have a critical field superior to that of Nb3Sn at all temperatures. The superconducting characteristics of MgB2 are interesting not only because of its relatively high transition temperature but also because it has more than one superconducting energy gap, a state of affairs anticipated in BCS theory but never before

OCR for page 69
Opportunities in High Magnetic Field Science seen experimentally.4 MgB2 is a metal with a layered structure in which the boron atoms form hexagonal layers and the magnesium atoms are located between the boron layers, above and below the centers of the hexagons. The unusual properties of the material are thought to result from strong interactions between phonons and electron orbitals in the boron layer. But unlike previously known good superconductors, MgB2 does not have a high charge-carrier density, only simple s- and p-shell electrons are involved in its superconductivity, and it has a noncubic crystal structure. Other features make MgB2 even more remarkable: Only a few of the atomic-vibration modes are involved in its ability to superconduct, and only about half of its charge carriers are strongly affected during superconductivity. The remarkable properties of MgB2 open a new window in superconductivity for fundamental as well as applied research. Prospects for the applications of MgB2 are enhanced by the inherently low cost of Mg and B. Conductors can be made by placing either B + Mg or MgB2 powders into metal tubes that are then drawn out into wires such as that shown in Figure 3.9e using conventional metal fabrication technology. Wires up to 1 km long have already been made by two or three companies in Japan, and this success has stimulated the development of lower-field, higher-temperature magnets for MRI. Furthermore, the density of MgB2 is comparable to that of aluminum, which may lead to new applications where weight is an important consideration. Present conductors do not yet have the capability indicated by the thin films whose Hc2 data are shown in Figure 3.10, and the current densities achieved are not yet fully competitive with those of Nb-based conductors. Nevertheless it is clear that with proper alloying and development, MgB2 could outperform any Nb-based conductor. For these reasons, the HEP magnet community has started to explore options for using MgB2 in the next generation of accelerator magnets. Equally exciting, the critical temperature of MgB2, 39 K, would allow electronic circuits based on this material to operate at 20-25 K, achievable by a compact cryocooler, which gives this material a significant advantage over low-temperature superconductors. Compared with the high-temperature superconductors, MgB2 is simpler, cheaper, and more stable over time. Superconducting Magnet Design Superconducting magnet technology has advanced to the point where 20-T laboratory magnets are not uncommon, and 900-MHz NMR magnets are available commercially. As noted earlier, a 920-MHz NMR magnet (21.6 T) having a 54-mm bore has been put into operation at the National Institute for Materials 4   H.J. Choi, D. Roundy, H. Sun, M. Cohen, and S.G. Louie, The origin of the anomalous superconducting properties of MgB2, Nature 418, 758-760 (2002).

OCR for page 69
Opportunities in High Magnetic Field Science Science (NIMS), Tsukuba (Japan), and a 900-MHz, 63-mm bore magnet manufactured by Oxford Instruments is operating at Pacific Northwest National Laboratories (PNNL). The NHMFL in Tallahassee recently brought to full field its 104-mm bore, 21.15-T magnet, which is the largest bore 21.15-T magnet with persistent field in the world. The next big step is clearly going to be the development of 1-GHz (23.5 T) NMR magnets, but their size, complexity, and cost will be daunting. High-field superconducting magnets present many of the same engineering problems as resistive and pulsed magnets—namely, very high stored energy and stresses—but they also have problems all their own. The superconductivity of superconducting materials is limited by the strength of the magnetic field they experience in a manner that depends on their operating temperatures, and all the superconducting materials now in use, except Nb-Ti, are mechanically weak, or intolerant of strain. In addition, today’s superconducting magnets require liquid-helium cooling systems that must be integrated into the coil without compromising its electrical or structural integrity. Even more challenging is the requirement that the energy in the field of an energized superconducting magnet be safely and passively dissipated in the event the coil quenches (i.e., ceases to be superconducting). The energy, which can amount to many tens of megajoules, must be dissipated within the magnet or to external energy-dump circuits. Failure during quench can occur due to overvoltage because of a too-rapid collapse of the magnetic field or, at the opposite extreme, because the magnetic energy dissipates in the generally small normal-state regions of the coil. To design superconducting magnets to withstand these perils is a complex but essential undertaking. Some of the same engineering solutions to the structural problems of resistive magnets have been adopted for superconducting magnets. For example, magnets are often divided into radially independent subcoils to reduce the transmission of radial forces and to limit hoop stress. This strategy can also result in a more efficient use of materials because the coils can be “graded.” More expensive superconducting materials such as Nb3Sn can be used in the inner coils, while less expensive, more ductile conductors like Nb-Ti can be used in the outer coils. In addition, conductor size and number of turns can be varied in each subcoil to optimize the winding current density, and mechanical reinforcing structures can be incorporated within each subcoil as needed. These techniques are generally used for high-field NMR magnets because of their persistent-mode operation. Magnets intended to produce more transient fields often use internal heaters or external shunt resistors. Mechanical strain is an extremely important issue for superconductors, as already noted, and electrothermal stability is another critical factor in the design and operation of superconducting magnets. All the solid components of a superconducting coil have extremely small heat capacities at 4 K and below, which

OCR for page 69
Opportunities in High Magnetic Field Science means that even a slight disturbance, such as the microheating generated by a local stick-slip friction event, can cause a small portion of the conductor to go normal, which generates resistive heating and may cause the entire coil to quench. This behavior requires a mechanical design different from that for a pulsed magnet, where some plastic deformation of the conductor and coil can be allowed. Outlook for Superconducting Magnets Experience shows that Nb3Sn reaches its practical limit of usable current density in the 21- to 23-T range, a limit that many current superconducting magnets have now reached. Nb3Sn technology might be further developed to the point that 1-GHz NMR magnets can be made using it, but it is unlikely that it will go much higher than that because of the intrinsic limitation on Hc2. A new superconducting material will probably be needed. An A15 compound such as Nb3Al might be suitable; this material is being developed at a very modest pace and has yet to be produced in large quantities. However, it is widely believed that some form of one of the newer HTS superconductors or MgB2 will emerge as the conductor of choice. Although these materials are widely recognized for their high critical temperatures, the more relevant measure of their usefulness is their very high upper critical fields. Like Nb3Sn, however, these materials are all weak structurally. Stress limitation will therefore be an engineering issue, although the strain tolerance of these materials is high enough to permit winding at reasonable bending diameters. While the availability of new and improved superconducting materials offers hope that superconducting magnets can be built that deliver fields significantly greater than those available today, the challenges still to be met should not be underestimated. The intrinsic coupling of high magnetic field and high stored energy, leading to high stresses, will force magnet sizes to grow unless the yield strength and modulus of elasticity of materials used for their construction can be increased. Increases in magnet size reduce performance because the farther conductors are from the center of a magnet’s bore, the less they contribute to the magnetic field in the bore. HYBRID MAGNETS The highest DC magnetic fields available today are produced by hybrid magnets, which consist of an inner, water-cooled, DC resistive magnet surrounded by a superconducting outer magnet (see Figure 3.12). The 45-T (32-mm bore) magnet at the NHMFL, which is the most powerful DC magnet in the world, obtains 31 T from its resistive magnet and 14 T from its superconducting magnet. The outer diameter of the resistive magnet of course determines the bore size of the super-

OCR for page 69
Opportunities in High Magnetic Field Science FIGURE 3.12 View of the NHMFL 45-T hybrid magnet. Images courtesy of National High Magnetic Field Laboratory.

OCR for page 69
Opportunities in High Magnetic Field Science conducting magnet. Although the 14-T field produced by the superconducting coil of this magnet is significantly less than the peak field of the NIMS 920-MHz magnet mentioned earlier, the energy stored in its field is three times greater (115 MJ) because its bore is so much larger. The apparatus to support and sustain a hybrid magnet is also much larger; the Florida 45-T hybrid is more than 10 m tall. Design Challenges Unusual engineering is needed to solve the stress and cooling problems of a hybrid magnet because there is so much energy stored in its field. For these magnets, field homogeneity requirements are usually relatively modest. The superconductor in the 45-T hybrid is a cable-in-conduit conductor that consists of a multistrand, multistage superconducting cable encapsulated in a structural steel conduit (see Figure 3.13). The conductor is insulated and then wound into the coil form required. If Nb3Sn is the superconductor used, as is the case for two of the three coils in the superconducting portion of the 45-T magnet at Tallahasee, the reaction heat treatment must be performed after the windings are formed. Cooling is accomplished by forcing supercritical helium (or stagnant superfluid helium-2) through the interstices between cables in the conduit, which account for approximately 35 percent of its total volume. The advantages of this design include (1) very high electrothermal stability because of the high heat capacity of liquid helium and large heat transfer surface area to helium in the conduit, (2) integration of structural steel throughout the winding pack, and (3) high voltage integrity of the coil because of the complete coverage of conductor with insulation, which is achieved without compromising the cooling of the magnet. The use of a multilevel cable allows for conductor currents of up to several kiloamperes and results in a lower terminal voltage when energy is dumped to an external room-temperature resistor if the magnet quenches or is deenergized for other reasons. However, high conductor current also requires the use of vapor-cooled current leads, which increases either helium consumption and/or refrigeration requirements. The NHMFL magnet also takes advantage of conductor grading to increase efficiency by using conductors of three different sizes for the three subcoils. One of the main problems of hybrid magnets is protecting their superconducting components from the consequences of sudden resistive magnet shutdowns caused by failure of magnet integrity or of the power supply. The resulting rapid decrease in the resistive magnet field can produce AC losses in the superconducting windings large enough to make them quench. In addition, the large mutual coupling between the magnetic fields of the two coils generates a large Lorentz force, the sudden removal of which may cause mechanical displacement of the superconducting magnet, which could have dangerous mechanical or electrical consequences. Such

OCR for page 69
Opportunities in High Magnetic Field Science FIGURE 3.13 Cross section of one of the cable-in-conduit conductors used in the 45-T hybrid magnet at NHMFL. The stainless steel conduit provides structural support and the 35 percent void around the cable provides a helium cooling area. The overall winding pack current density is reduced by these elements because they cannot carry current to contribute to the magnetic field generated. Image courtesy of National High Magnetic Field Laboratory. unplanned and disastrous quenches have occurred in the past, and newer designs take this possibility explicitly into account. Outlook for Hybrid Magnets NHMFL-Tallahassee is proposing the construction of a series-connected hybrid magnet, the resistive and superconducting components of which will be operated electrically in series from the same power supply. The high inductance of the superconducting magnet should reduce the current ripple in the resistive magnet, thereby improving the temporal stability of the magnetic field. A second important advantage of this design is that no overcurrent or unbalanced forces

OCR for page 69
Opportunities in High Magnetic Field Science occur during a quench. In addition, the resistive coil will be designed so that it can be shimmed to higher field homogeneity than normal for a resistive magnet. The series-connected hybrid magnet is designed to produce 35 T (40-mm bore) with an inhomogeneity of less than 1 ppm in a 10-mm-diameter spherical volume. Hybrid magnets such as this one are intended to reduce the power consumed to produce the fields they generate rather than maximizing field strength. COORDINATION OF MAGNET DEVELOPMENT One striking characteristic of all of the sciences that use high magnetic fields is how constrained they all are by a single body of technological information—namely, magnet technology. This shared interest notwithstanding, each constituency has historically tended to develop the magnets it needs without much reference to the others. The reasons are several and obvious. The different communities have different missions and need magnets that differ correspondingly. In addition, they are supported by different funding agencies, each with its own perspective. Recently the HEP and fusion communities have begun to more closely coordinate their development of Nb3Sn wire, an important first step in a healthy direction. Given the intense interest of all magnet user communities in magnet technology and engineering, it would seem to make sense for them to mount a coordinated effort to advance magnet technology. One component of such a coordinated effort might be the construction of high-field instrumentation at NHMFL specifically for the engineering research necessary for the production of high-performance magnets. A second component would certainly be the development of new materials for magnet construction, and a third would be an exploration of other constraints on magnet design and performance. The final ingredient would be a framework to coordinate and integrate communication among the different communities. This framework might start with something as simple as topical conferences and could extend to a management structure for operating a joint program. It is clear from recent experience that magnets producing fields significantly higher than those available today are going to be either impossible or prohibitively expensive if current technology is used. Rather than supporting an all-out, brute-force effort to prevail using barely adequate technologies, it makes sense to find new approaches that will make it easier (and cheaper) to build the magnets needed. Essential to this enterprise will be the development of both resistive and superconducting materials with improved electrical, magnetic, and mechanical properties. This initiative could be a collaborative enterprise involving both publicly supported researchers and commercial enterprises, but however it is structured, it will certainly require substantial public support.