Int. J. Appl. Ceram. Technol.

, 4 [3] 250259 (2007)

Ceramic Product Development and Commercialization

Overview of MgB2 Superconductor Applications

Michael Tomsic,* Matthew Rindeisch, Jinji Yue, Kevin McFadden, and John Phillips
Hyper Tech Research Inc., Columbus, Ohio 43212

Michael D. Sumption, Mohit Bhatia, Scot Bohnenstiehl, and Edward W. Collings

LASM, Department of Material Science and Engineering, The Ohio State University, Columbus, Ohio 43210

Since 2001, when magnesium diboride (MgB2) was rst reported to have a transition temperature of 39 K, conductor development has progressed to where MgB2 superconductor wire in kilometer-long piece-lengths has been demonstrated in coil form. Now that the wire is available commercially, work has started on demonstrating a MgB2 wire in superconducting devices. This article discusses the progress on MgB2 conductor and coil development, and the potential for MgB2 superconductors in a variety of commercial applications: magnetic resonance imaging, fault current limiters, transformers, motors, generators, adiabatic demagnetization refrigerators, magnetic separation, magnetic levitation, superconducting magnetic energy storage, and high-energy physics applications.

Magnesium Diboride (MgB2) Wire Development at Hyper Tech Research Inc. (Hyper Tech)
Strand Design and Manufacturing Process

Hyper Tech currently offers a commercial MgB2 superconductor wire designated as an 1811 Nb/Cu/ Monel multilament strand. This designation, in the order as it is listed, corresponds to the number of monolaments in the strand, followed by the chemical barrier, monolament sheath, and multilament sheath materials. Figure 1 shows a typical wire cross section of the commercial wire currently fabricated at Hyper Tech.

*tomsic@voyager.net r 2007 The American Ceramic Society

Hyper Tech uses a patented process for manufacturing MgB2 superconductors; this process is called the continuous tube lling and forming (CTFF) process. During the CTFF process, powder is dispensed onto a strip of metal as it is formed into a tube. The CTFF process results in an overlap-closed tube lled with powder in continuous lengths; fabrication of precursor wire billets continuously is the fundamental advantage of the CTFF process. Hyper Tech has used the ex situ technique (lling the tube with already formed MgB2 powder), but primarily uses the in situ technique for fabricating MgB2 superconductor wire. The in situ technique involves direct lling of a metallic tube with elemental magnesium and boron powder and subsequent drawing, followed by heat treatment, during which the elements react to form MgB2. The in situ process offers the advantages of simplicity of fabrication, lower

www.ceramics.org/ACT

MgB2 Superconductor Applications

Superconductor powder

251

CuNi alloy

Copper

Niobium

Fig. 1.

Cross section of a typical 1811 multilament MgB2 wire.

reaction temperatures, and a more manageable way for adding dopants or other additives into the wire.1 The lower reaction temperature is particularly important as it helps to minimize the possibility of powderbarrier reactions. The strand fabricated by the CTFF process must not only contain the powder but must also be chemically compatible in the heat-treatment range where the powder is being reacted; the barrier material is typically pure niobium but Ni, Fe, and Ti can alternatively be used. In general, the Nb barrier is enclosed in an outer sheath, or sheaths, to aid the wire drawing and eventually to provide electrical stabilization. For most of the wire made at Hyper Tech, the overlap-closed niobium strand is inserted into a seamless copper (or copper alloy) tube and is drawn to a predetermined size, creating a monolament MgB2 strand. The monolament strands are then restacked into another seamless tube; the diameter and length of the restack tube will determine the nal piece length of the wire. Billets are presently being sized to produce continuous wire lengths of 5 km for a typical 0.8 mm diameter wire. Experimental work will continue to take the 0.8 mm conductor piece length up to 30 km, the typical continuously fabricated piece length for NbTi superconductor wires. The second seamless tube, or multilament outer sheath, is typically a nickelcopper alloy such as Monelt (Huntington Alloys, Huntington, WV) but can also be

copper-rich CuNi alloys. Work is continuing to develop a wire, with the outer tube being all Cu. The number of laments in the restack can vary and usually Cu stabilizer laments are located in the center of the multilament restacked wire. The 1811 designation for the commercial wire therefore refers to 18 MgB2 monolaments and one center copper lament. Hyper Tech has fabricated a variety of experimental strands for various application-related projects, including the following:  Nb barrier, all-copper matrix. The Monel outer restack sheath is replaced with pure 101 copper. The use of copper as the outer sheath improves the ductility and stability of the strand.  Oxide-dispersion-strengthened (ODS) copper, trade name Glidcopt (SCM Metal Products, Research Triangle Park, NC). The Monel outer restack sheath is replaced with ODS Cu in order to combine the benets of lower resistivity without sacricing the significant degree of strength needed for drawing.  High lament count. MgB2 superconductor wire with up to 61 total laments has been fabricated. The strand design was a Nb/Cu/Monel.  Very small diameter. Small-sized MgB2 superconductor strands have been fabricated, such as a 0.07 mm round monolament and a 0.117 mm round seven lament Nb/Cu/Monel MgB2 conductor. The size of a MgB2 lament in the case of the 0.117 mm seven lament-restack strand was 17 mm. The size of a MgB2 lament in the standard 1811 multilament wire at 0.8 mm is 76 mm.  Rectangular. MgB2 superconductor wire has been fabricated in a rectangular shape with a 0.5 mm 1.0 mm aspect ratio in various multilament strand designs.
Transport Current Properties

A wire in short sample form, in long length on spools, or wound on coils is heat treated to react the elemental magnesium (99%) and boron (99.9%, amorphous) to form MgB2 in the strand. Heat treatments are single step and are performed under an argon atmosphere. The heat-treatment soak temperature is normally 7001C, held for 2040 min. The Ohio State University (OSU) has characterized the majority of the MgB2 conductors manufactured by Hyper Tech.24

252

International Journal of Applied Ceramic TechnologyTomsic, et al.

Vol. 4, No. 3, 2007

200

Jc(A/cm2)

60 40

60,000 40,000

10,000 8,000 6,000 4,000

20

20,000 2,000 0 2 4 H (T) 6 8

10

Fig. 2. Typical Ic, Je, and Jc properties of the standard Hyper Tech multilament MgB2 strand. Data on the graph are limited to Jc4104 A/cm2.

Usually, two types of wire samples are used in transport current density measurements: short samples and 1-m ITER barrel coil samples. The Jc criterion for both sample types is 1 mV/cm. Four-point Jc measurements are made with background elds of up to 15 T applied transversely to the strand. The ITER barrel-type samples have a gauge length of 50 cm and are made with a windand-react protocol. ITER barrel measurements are taken at 4.2 K in liquid helium. Short samples are 3 cm in length with a gauge length of 5 mm and are measured at elevated temperatures. Figure 2 shows the typical critical current and current density measurements versus applied eld at various temperatures of the standard multilament MgB2 conductor manufactured at Hyper Tech. Typical Jc values are 1 105 A/cm2 at 5 T, 4.2 K and 2 105 A/cm2 at 1 T, 20 K. The Ic at 1 T and 20 K is around 200 A. The average ll factor in the standard multilament strand is 15%. Hyper Tech is working to increase ll factor to 30%, which would increase the 20 K, 1 T Ic to 400 A. Thus far, experimental multilament strands with a ll factor of 24% have been produced.
n-Value

Nano-SiC Doping

Work is ongoing to improve the in-eld performance of MgB2 wires.5,6 The University of Wollongong as well as Hyper Tech and OSU have shown that SiC nanoparticles can significantly improve the properties of wire made from the binary Mg1B compound.7,8 Short sample measurements by OSU have found enhancements in (m0Hirr) and the upper critical eld (Bc2) with a SiC-doped wire made by Hyper Tech. Depending on the heat-treatment schedule, SiC doping yielded Bc2s as high as 29.7 T and Birrs as high as 25.4 T. Measure35 30 25 n-value 20 15 10 5 0 2 3 4 5 B (T) 6 7 8
4.2K 6K 10K 15K 20K 25K

Along with transport current density measurements, the corresponding exponential n values for Hyper Techs MgB2 superconductor wire have been determined. The n values shown in the variable temperature graph (Fig. 3) are typical values for the standard 1811 multilament wire described previously.

Fig. 3. n value versus B for a multilament MgB2 strand; short sample, variable temperature measurements performed by the Ohio State University.

Je(A/cm2)

100 80

4.2K 6K 10K 15K 20K 25K 30K

40,000 200,000 20,000 100,000 80,000

Ic (A)

www.ceramics.org/ACT
1e+7

MgB2 Superconductor Applications

253

869 - Undoped 18+1 Nb/Cu/Monel (HT 700/20) 1011 - 30nm SiC-10% (HT 675/40)

37-lament conductor. Figure 6 shows the relationship between Jc and strain for a 37-lament wire.
Alternating Current (AC) Losses

1e+6

J , (A/cm )

1e+5

1e+4

1e+3

1e+2 0 2 4 6 8 H (T) 10 12 14

Fig. 4. Comparison between SiC-doped and undoped MgB2 wires at 4 K, Jc versus eld.

ments showed that Hyper Tech wires with SiC-doped strands increase the current density in higher elds. Figures 4 and 5 compare the current density of a MgB2 SiC-doped strand with an undoped one at 4 and 15 K.

Strain Tolerance

NIST has measured the irreversible strain limit eirr of several Hyper Tech-fabricated multilament strands using a stress-free cooling apparatus.9 NIST found that eirr increases with the number of laments in the wire, and therefore reduction of lament size improves the strain tolerance of the conductor. The strain tolerance limits reported are 0.37% for a 7-lament conductor, 0.40% for a 19-lament conductor, and 0.48% for
1e+6 626 -Undoped 18+1 Nb/Cu/Monel (HT 700/20)
1011 - 30nm SiC-10% (HT 625/3hr)

The most near-term application to generate wire volume to help drive the MgB2 wire price down is MRI applications. As wire price is driven down by high-volume requirements from the MRI industry, the potential for MgB2 to be used in electrical power applications increases. Most of these applications operate on AC, and the few applications that are DC powered involve some AC ripple, or ramping up and down. AC losses are an issue in power system design; the potential of MgB2 to become a good low AC loss superconductor operating in the 1530 K range will be an important advantage for system integrators. Fortunately, the MgB2 strand is such that the barrier and sheath material can be easily interchanged for fabricating a conductor with a more resistive matrix than the standard multilament wire to reduce external eld-induced eddy current losses. The dominant factor in external eld losses is the lament size, which is reliably in the 50 mm range at present; Hyper Tech is working on reducing the lament size further. Nevertheless, the lamentation and reduction of lament size is trivial as compared with YBCO-coated conductors. Twisting, exceedingly difcult with YBCO conductors, is simple and straightforward for MgB2. With respect to transport current losses, MgB2 has the traditional advantage of a wire as compared with a tapethe losses are proportional to the width or radius, and therefore can be quite low for MgB2.
Normal Zone Propagation

1e+5

1e+4

1e+3

1e+2 0 2 4 6 H (T) 8 10 12

Fig. 5. Comparison between SiC-doped and undoped MgB2 wires at 15 K, Jc versus eld.

As MgB2 superconductor wire development advances and more and more coils are fabricated, quench protection is becoming increasingly important in magnet design. For the protection of superconducting devices, a high normal zone propagation velocity (vnz) is preferred. The University of Twente studied the vnz of MgB2 superconductors from several manufacturers, including Hyper Tech.10 For a standard Hyper Tech multilament strand, Twente found a vnz of 15 50 cm/s at 5 K, 4.8 T. For comparison, the vnz of NbTi and Nb3Sn ranges from 1 to 100 m/s; the vnz of YBCO-coated conductors ranges from 0.2 to 1.0 cm/s. Twente found that vnz will decrease with lower n values,

JC (A/cm2)

254
175

International Journal of Applied Ceramic TechnologyTomsic, et al.

Vol. 4, No. 3, 2007

37 Filament Wire Critical Current Je (A/mm2) 150

B I' E' F' G' H' C' D' A' B' A C D

G H

37 filaments 0.48%
J

125

100 T = 4 K; B = 6 T Initial n-value ~ 29 75

Sample loaded Sample unloaded K' J'

19 filaments 0.40%

irr ~ 0.48 %

50 0 0.1 0.2 0.3 0.4 Applied Strain (%) 0.5 0.6

7 filaments 0.37%

Fig. 6.

Critical current versus applied axial strain for a 37 lament MgB2 wire.9

and will decrease by a factor of 25 when transitioning from 25 K (vs 5 K). MgB2 Coil Development Advancements in MgB2 wire manufacturing have led Hyper Tech into superconductor coil and magnet development. The development of long-length multilamentary wire has enabled the design and fabrication of a series of MgB2-wound solenoid and racetrack coils. The characterization of these coils allowed us to determine the superconductivity properties of MgB2 wire over a long length. Establishing excellent properties over length in early demonstration coils was followed by designing of more realistic coils using MgB2 superconductor wire for specific applications. Long-length characterization in coil form is equally important for the advancement of the MgB2 superconductor because most commercial applications will require many kilometers of wire.
Wind-and-React Solenoid Coils

treated in owing Argon in a stainless-steel retort in a furnace and was then impregnated with epoxy. The multilament strands were insulated with an s-glass braid. One of the larger wind-and-react coils fabricated by Hyper Tech was wound with 740 m of 0.83 mm strand, totaling 3463 turns around a 3.8 cm bore. This coil attained a eld of 3.9 T at 4 K, 3.0 T at 15 K, 2.4 T at 20 K, 1.8 T at 25 K, and 0.9 T at 30 K. Depicted in Figs 7 and 8 is a photograph of the coil and a graph showing

Hyper Tech fabricated several solenoid coils wound with long lengths of Nb/Cu/Monel-type multilament wire, using the wind-and-react approach. As coil fabrication techniques advanced, each successive coil was wound with increasing continuous lengths of wire. The magnet bore size was either 3.8 or 7.6 cm, and the coil height was 7.1 cm in the largest coil. Each coil was heat

Fig. 7.

Photograph of a 740 m wind-and-react coil.

www.ceramics.org/ACT
25 120 Critical Current, I , A 100 J , kA/cm J , kA/cm 80 60 40 20 0 5 0 20 15 10 80 60 40 20 0 0 10 20 Temperature, T, K 30 100

MgB2 Superconductor Applications

255

4 Bore Field, B , T

Fig. 10. Fifty-three centimeters react and wind coil wound with 820 m of an 1811 multilament MgB2 wire.

40

Fig. 8. Ic, Je, and Jc as a function of temperature for 740 m windand-react solenoid coil.11

the transport properties of the coil measured at various temperatures from 4.2 to 35 K. A load line was generated for the 740 m solenoid coil to compare the test results of the coil with short sample measurements. Using the coil geometry and Ic measurements, a eld distribution in the bore was calculated to nd the peak eld in the coil at the innermost winding.11 At an Ic of 84.5 A at 4.2 K, the peak eld in the coil was estimated to be 3.96 T. The short sample was a 1 m ITER barrel that was heat treated with the coil. The load line in Fig. 9 shows that the coil performed better than the short sample; the lower short sample value may be due to a variation in the properties in this particular strand.
React-and-Wind Solenoid Coils

in this coil was 53 cm. Over 820 m of an s-glass insulated Nb/Cu/Monel-type multilament strand with a total of 482 turns were wound on a copper coil former. A picture of the 53 cm coil is shown in Fig. 10. The insulated wire was heat treated on a stainless-steel spool to react the wire to form MgB2. The reacted wire was wound on the coil bobbin from the stainless-steel spool. Epoxy was applied during coil winding to impregnate the coil using a wet-wind-compatible epoxy. The coil generated an axial eld of 0.12 T at 20 K, which was the expected result.
Wind-and-React Racetrack Coils

Hyper Tech demonstrated a large solenoid coil fabricated with the react-and-wind approach; the bore size
400 Nb/Cu/monel short sample at 4.2K Solenoid Coil Load Line

300

Hyper Tech fabricated over 12 single-layer racetrack coils for a cryogenic rotor in an LH2-cooled, superconducting generator demonstration under a NASA contract. The racetrack coils were made using the windand-react approach. The wire was insulated with either a single-layer s-glass braid or a ceramic solgel coating, and the wire length varied between 30 and 80 m. The coils were epoxy impregnated after heat treatment. Figure 11 shows a picture of a racetrack coil. The best-performing coil reached 400 A at 4 K. This rotor coil generator project will lead to the fabrication of four rotor coils, with each coil being fabricated from one 600 m continuous length of a MgB2 superconductor wire. The superconducting generator demonstration will be a

Ic, A

200

100

0 0 2 4 B, T 6 8 10

Fig. 9. Coil load line and short sample (1 m barrel) Jc versus B curve for the 740 m wind-and-react solenoid coil.11

Fig. 11.

Typical MgB2 racetrack-type rotor coils.

256

International Journal of Applied Ceramic TechnologyTomsic, et al.

Vol. 4, No. 3, 2007

Fig. 12. The rst full size MgB2 rotor coil fabricated for a 2 MW generator demonstration. The rotor coil is shown installed on a test plate.

2 MW generator operating at 20 K in liquid H2. The rst rotor coil has been fabricated, featuring a step prole, as shown in Fig. 12. Potential Applications for MgB2 Superconducting Wire
MRI

Currently, the largest commercial market for superconducting wire is MRI systems. An MgB2 superconducting wire has the potential to impact this industry.12 The degree of impact depends on the price performance ($/kA m) of the wire at a given temperature and background eld. The eld of an MRI system is typically described by the center bore eld; however, the wire can potentially see up to twice that of the center. For example, a 1.5 T MRI system might experience up to 3 T on the wire. The rst MRI-type applications will most likely be for systems operating in the 1027 K range in background elds on the wire from 0.5 to 3.0 T; these systems could be either an open or enclosed MRI system. By operating at temperatures higher than 4 K (liquid He, operating temperature of present MRI devices), there are potential savings in the construction of a vacuum cryostat, especially with the open MRI systems that typically require two separate cryostats. Because MgB2 superconductors have a higher temperature margin compared with NbTi and Nb3Sn superconductor wires, the MgB2 conductor can be constructed with

less copper (which is needed for stability), potentially increasing the overall coil engineering current density. Also, for smaller bore magnets (sizes smaller than fullbody MRI, o100 cm bore), less costly cryocoolers can potentially be used. Therefore, the mostly likely systems to be initially marketed using a MgB2 superconducting wire will be smaller bore extremity (arm and leg) MRI systems, and open full-body MRI systems. An increase in MgB2 wire usage beyond these initial 1027 K MRI system markets will depend on how low the price performance can be driven for a MgB2 wire operating at 4 K, especially at high elds. An MgB2 wire operating at 4 K has the potential to exceed the performance of NbTi wire in high elds and become equal to or greater than Nb3Sn wire performance at 4 K. For MgB2 to compete on a price performance basis with NbTi and Nb3Sn at 4 K, increases in eld performance and lower manufacturing costs need to be demonstrated. As discussed previously, dopants like nano-SiC are being added to increase Bc2 and work is being conducted worldwide to increase both pinning and connectivity in MgB2. Experimental work with MgB2 thin lms13,14 has indicated that an order of magnitude improvement in current density at a given magnetic eld should be possible. With regard to the $/m price of MgB2 wire, manufacturing processes are similar to that used for NbTi and Nb3Sn. A typical NbTi superconductor wire at 0.8 mm sells for around $1/m, and Nb3Sn sells for $46/m. The selling price for MgB2 wire on a per-meter basis should be less than NbTi wire based on manufacturing and material costs. Not only are the basic materials Mg and B less costly than Nb and Ti, the density of MgB2 is one-third that of NbTi and Nb3Sn. Thus, for the same kilogram of raw material, we obtain three times the wire length, meaning that the $/ m selling price of MgB2 wire should become much less than NbTi wire. If the engineering current density of the MgB2 wire can be increased to exceed the NbTi or Nb3Sn wire performance at 4 K in eld, then the combination of high superconductivity performance and low wire price should generate a much lower $/kA m than the present NbTi or Nb3Sn wire at a given magnetic eld. Therefore, for full-body MRI systems (1.5 and 3.0 T) and ultra-higheld (7 T plus) MRI and NMR, MgB2 wire could become the wire of choice based on basic wire economics.
Fault Current Limiters

Superconducting fault current limiters have been investigated and developed over the past 15 years. Their

www.ceramics.org/ACT

MgB2 Superconductor Applications

257

main advantages are a negligible inuence on an electrical network under normal operating conditions, practically instantaneous limitation, and automatic response without an external trigger. There are basically two types of fault current limiters: resistive and inductive. MgB2 is known to have a very sharp transition from a superconducting to a normal state. This sharp transition makes MgB2 ideal for the resistive-type fault current limiter. Moreover, the normal zone heat propagation of MgB2 is rapid compared with ceramic superconductors, which minimizes hot spots, and the conductor can be designed with a variety of highly resistive sheath materials. With its potential low cost and ability to operate at moderate temperatures (2030 K), MgB2 is very attractive as a cost-effective FCL element. Work has been reported1517 on the current limiting properties of different kinds of MgB2 wires at 50 Hz in the temperature range of 20 30 K. MgB2 wires showed good limiting properties characterized by a fast transition to the normal state during the rst half of the rst cycle and a limiting effect during the subsequent six cycles without damage to the wire samples. The temperature differences at the beginning of the transition in the rst cycle smoothed out during subsequent cycles; this was more evident for Custabilized multilament MgB2 wires. A stabilized twisted multilament wire is also a low AC loss superconductor, an important factor for fault current limiters that operate on AC during the normal (no fault) condition. The other type of fault current limiters is inductive in nature, which involve superconductor coils fabricated with considerably longer lengths of superconductor. MgB2 is also suitable for potential low-cost coils in the 2030 K operating range. MgB2 fault current limiters will be in competition from inductive FCL coils using YBCO-coated conductors.
Transformers

impedance will improve voltage regulation and stability and increase real and reactive power availability to the power system. There are many challenges in building and demonstrating a superconducting transformer, but with regard to the wire the primary factor is producing a superconductor wire with a price performance in the $13/kA m range at the targeted eld and temperature. Several of the proposed transformer designs suggest that the eld on the wire can be as low as 0.150.20 T, but some designs could go up to 1.0 T. With a stabilized, multilamentary, twisted, resistive matrix, low AC loss MgB2 wire at 2030 K in elds of around 0.21.0 T, it appears that the targeted conductor cost can be met in a few years. The targeted power utility transformer for commercialization has been identied at 30 MVA or larger. If the other issues for superconducting transformers such as cryogenic insulation, high voltage terminations, and cryocooler refrigeration optimization can be addressed, there is the potential of commercializing a high-MVA transformer that is of a lower cost than a conventional transformer, based on low-cost MgB2 superconductor wire. The competition to a MgB2 wire will be a YBCO-coated conductor operating in the 6577 K range in liquid N2.
Motors and Generators

Superconducting transformers will have reduced size and weight and lower losses compared with transformers fabricated with conventional wire. In addition to these attributes, the superconducting transformer may significantly benet the entire electrical system. Benets come from reducing the short-circuit current in the system and lower transformer impedance. The superconducting wire has current-limiting capability. This can reduce the interrupting ratings of circuit breakers and in some cases, permit the use of mesh networks for a tightly coupled power system. The lower transformer

Superconducting motors and generators have several potential advantages. They can be power-dense, lightweight, small volume, highly efcient, and reliable. MgB2 can offer advantages in several of the motor and generator systems being demonstrated. One design is the superconducting homopolar motor being developed for the Navy; the present design uses NbTi superconductors operating at 4 K. As MgB2 current densities start to exceed those of NbTi at 4 K, MgB2 offers the potential of higher power densities, high temperature margins, and lighter weight coils for these superconducting homopolar motors. Other superconducting motor systems have been demonstrated using liquid neon and helium gas operating in the 2030 K range; MgB2 superconducting coils have generated magnetic elds in the 0.52.0 T eld range at these temperatures. For all-electric aircraft motors, the fuel of choice is liquid H2; liquid H2 will be available for the turbine engine on an aircraft. Presently, for NASA, Hyper Tech is building MgB2 superconducting rotor coils for a cryogenic 2 MW generator demonstration that will be cooled with liquid hydrogen at 20 K.

258

International Journal of Applied Ceramic TechnologyTomsic, et al.

Vol. 4, No. 3, 2007

High Energy Physics (HEP) Applications

The long-term goal of MgB2 in HEP applications is the production of a wire that will produce engineering current densities at least 1.2 105 A/cm2 at 4 K in elds of 1216 T.18 While MgB2 has the potential to reach this performance, MgB2 should be able to satisfy several nearterm accelerator-related requirements. Continually improving the properties should enable MgB2 to be used in the windings of undulator magnet installations and replacement wiggler magnets in accelerator applications. Some near-term, lower eld special applications include replacement light source bending magnets and solenoids for a muon collider. In synchrotron light sources, the benet of increased brilliance of the photon beam at higher harmonics follows the replacement of permanent-magnet undulator magnets with superconducting ones for elds o5 T. For such a modest eld in a high-radiation application, MgB2 superconductor with its large thermal margin seems to be an ideal candidate superconducting material.
Adiabatic Demagnetization Refrigerators (ADR)

available since the early 1980s. The primary application of superconducting magnetic separators is removing iron from kaolin clay. These systems typically operate in the 2 T range. Experimental systems have been built for various wastewater treatment demonstration projects. Higher eld systems in the 25 T range have been suggested for some magnetic separation systems. Because these systems are typically large diameter bore magnets, that is, 12 m, they require considerable superconductor wire. Therefore, a low-cost superconductor wire is needed. MgB2 superconductors offer the potential of a low-cost wire with a high temperature margin that can enable large conduction-cooled magnets operating in the 1025 K range.
Superconducting Magnetic Energy Storage (SMES) and Magnetic Levitated (MagLev) Trains

Future NASA instruments operating in space will feature detectors operating well below 1 K. The cooling system of choice is ADR.19 These devices produce cooling by manipulating the entropy of a paramagnetic material, commonly referred to as a salt pill. This process involves ramping up and down the magnetic eld seen by the pill while opening and closing a thermal switch between the pill and a heat sink. The heat is then pumped up a chain of stages each at successively higher temperatures, each stage requiring a superconducting magnet. Targeted temperatures are 4, 10, and 15 K. While presently NbTi is being used at 4 K, and Nb3Sn is being considered at 10 K, MgB2 is being considered for a 15 K operating temperature by NASA. As MgB2 performance exceeds that of NbTi at 4 K and Nb3Sn at 10 K, MgB2 could be considered for coils operating anywhere from 4 to 15 K or higher. This superconducting application is unique because a very small diameter wire is required. The desired superconductor wire size needs to be in the 0.0750.200 mm range with an Ic in the 330 A range, operating at the appropriate temperature (4, 10, 15 K) in 34 T magnetic elds.
Magnetic Separation

Both of these applications require the storage of energy in fairly large coils. The typical magnetic elds generated are in the 25 T range. The cost of the superconductor is important because a considerable amount of superconductor wire is required for these large coils. MgB2 superconductors have the potential to become a low-cost wire with a high temperature margin. If the desire is to operate at temperatures above 4 K with these large coils, MgB2 superconductors will be a good t. There is also the potential for hybrid coils operating in the 1025 K range using MgB2 as the outer coil and a YBCO-coated conductor as the inner coil in the higher eld region. Summary Significant progress has been made since 2001 on the development of long-length MgB2 superconductor wires. MgB2 superconductor wire performance is expected to increase rapidly over the coming years through improved pinning and connectivity. The price performance at 4 K operation could exceed both NbTi and Nb3Sn superconductor wire. At operating temperatures from 4 to 30 K, the MgB2 wire price performance can potentially enable several commercial applications. The primary competition to MgB2 in the 430 K temperature range will be the ceramic superconductors, BSCCO, and YBCO tape; the main question is whether BSCCO and YBCO will eventually be as cost effective as MgB2 for superconducting devices operating in this temperature range. Presently, the future is positive

Superconducting magnetic separators using NbTi operating at 4 K in liquid He have been commercially

www.ceramics.org/ACT

MgB2 Superconductor Applications

7.

259

for MgB2 superconductors, as there are several ongoing demonstration projects that are directed toward commercial applications. Acknowledgments The authors thank Dr. Shi Xue Dou from the Institute of Superconducting and Electronic Materials Department of the University of Wollongong for his assistance in the research and development of MgB2 superconductors presented in this paper, and Drs. Najib Cheggour and Jack Ekin at NIST for providing axial strain analysis. References
1. M. D. Sumption, M. Susner, M. Bhatia, M. Rindeisch, M. Tomsic, K. McFadden, and E. W. Collings. High Critical Current Density Multilamentary MgB2 Strands, Applied Superconductivity Conference, 5MB05, 2006. M. D. Sumption, M. Bhatia, M. Rindeisch, M. Tomsic, and E. W. Collings, Transport and Magnetic Jc of MgB2 Strands and Small Helical Coils, Appl. Phys. Lett., 86 102501-1102501-3 (2005). M. D. Sumption, M. Bhatia, X. Wu, M. Rindeisch, M. Tomsic, and E. W. Collings, Multilamentary, In-Situ Route, Cu-Stabilized MgB2 Strands, Supercond. Sci. Technol., 18 [5] 730734 (2005). M. D. Sumption, M. Bhatia, M. Rindeisch, M. Tomsic, and E. W. Collings, Transport Properties of Multilamentary, In Situ Route, Cu-Stabilized MgB2 Strands: One Meter Segments and the Jc(B, T) Dependence of Short Samples, Supercond. Sci. Technol., 19 155160 (2006). H. Kumakura, H. Kitaguchi, A. Matsumoto, and H. Hatakeyama, Upper Critical Fields of Powder-in-Tube-Processed MgB2/Fe Tape Conductors, Appl. Phys. Lett., 84 3669 (2004). R. Flukiger, H. L. Suo, N. Musolino, C. Beneduce, P. Toulemonde, and P. Lezza, Superconducting Properties of MgB2 Tapes and Wires, Physica C, 385 286305 (2003).

8.

9. 10.

11.

12.

13.

14.

2.

15.

3.

16.

4.

17.

5.

18.

6.

19.

M. D. Sumption, M. Bhatia, M. Rindeisch, M. Tomsic, S. Soltanian, S. X. Dou, and E. W. Collings, Large Upper Critical Field and Irreversibility Field in MgB2 Wires With SiC Additions, Appl. Phys. Lett., 86 092507-1 092507-3 (2005). S. X. Dou, A. V. Pan, S. Zhou, M. Ionescu, H. K. Liu, and P. R. Munroe, Substitution-Induced Pinning in MgB2 Superconductor Doped With SiC Nano-Particles, Supercond. Sci. Technol., 15 1587 (2002). N. Cheggour and J. W. Ekin, Electromechanical properties of MgB2 superconductors, (to be published). H. van Weeren, M. Dhalle, A. den Ouden, and H. ten Kate. Normal Zone Propagation in MgB2 Conductors, Applied Superconductivity Conference, 5MB01, 2006. M. D. Sumption, S. Bohnenstiehl, F. Buta, M. Majoros, S. Kawabata, M. Tomsic, M. Rindeisch, J. Phillips, J. Yue, S. Kawabata, and E. W. Collings. Wind and React and React and Wind MgB2 Solenoid, Racetrack and Pancake Coils, Applied Superconductivity Conference, 2LC01, 2006. Y. Iwasa, D. C. Larbalestier, M. Okada, R. Penco, M. D. Sumption, and X. Xi, A Round Table Discussion on MgB2 Toward a Wide Market or a Niche Production?A Summary, IEEE Trans. Appl. Supercond., 16 1457 (2006). C. B. Eom, M. K. Lee, J. H. Choi, L. J. Belenky, X. Song, L. D. Cooley, M. T. Naus, S. Patnaik, J. Jiang, M. Riakl, A. Polyanskii, A. Gurevich, X. Y. Cai, S. D. Bu, S. E. Babcock, E. E. Hellstrom, D. C. Larbalestier, N. Rogado, K. A. Regan, M. A. Hayward, T. He, J. S. Slusky, K. Inumaru, M. K. Hass, and R. J. Cava. High Critical Current Density and Enhanced Irreversibility Field in Superconducting MgB2 Thin Films, Nature, 441 558 (2001). W. N. Kang, E. M. Choi, H. J. Kim, H. J. Kim, and S. I. Lee, Growth of Superconducting MgB2 Thin Films Via Postannealing Techniques, Physica C, 385 2430 (2003). M. Majoros, L. Ye, A. M. Campbell, T. A. Coombs, A. V. Velichko, D. M. Astill, P. Sargent, M. Haslett, M. D. Sumption, E. W. Collings, M. Tomsic, and M. Husband, Fault Current Limiting Properties of MgB2 Superconducting Wires, IEEE Trans. Appl. Supercond., 17 (2007). Ye Lin, M. Majoros, A. M. Campbell, T. Coombs, S. Harrison, P. Sargent, M. Haslett, and M. Husband, MgB2 Sample Tests for Possible Applications of Superconducting Fault Current Limiters, IEEE Trans. Appl. Supercond., 17 (2007). M. Majoros, L. Ye, A. M. Campbell, T. A. Coombs, M. D. Sumption, and E. W. Collings, Modeling of Transport AC Losses in Resistive Fault Current Limiters Under no Fault Conditions, IEEE Trans. Appl. Supercond., 17 (2007). E. W. Collings, S. Kawabata, M. Tomsic, and M. D. Sumption, Magnesium Diboride Superconducting Strand for Accelerator and Light Source Applications, IEEE Trans. Appl. Supercond., 16 14451448 (2006). J. Tuttle, S. Pourrahimi, P. Shirron, E. Canavan, M. DiPirro, and S. Riall, A 10 K Magnet for Space-Flight ADRs, Cryogenics, 44 383388 (2004).