Nb₃Al Multifilamentary Superconductors

The RHQT process

We have developed a new fabrication technique for JR Nb₃Al conductors by exploiting the transformation from supersaturated bcc-solid-solution (Nb(Al)_ss).  The rapid-heating, quenching and transformation (RHQT) method is, from the practical point view, the most appropriate to prepare the transformation-processed Nb₃Al conductors, since a several-hundred meter piece-length and multifilamentary structure is available.  In the RHQT method, the JR and RIT processed Nb/Al multifilamentary wires, several hundred meters in length, have been reel-to-reel Ohmic-heated to a very high temperature (~1900°C), quenched into a molten gallium bath at ~50°C and transformation annealed to form the Nb₃Al phase via bcc Nb(Al)ss (Fig. 3).  Inside a vacuum chamber Nb/Al wire is fed from a pay-off spool, around a Cu contact wheel, into liquid Ga and onto a take-up spool.  Current is passed through the section of the wire between the Cu wheel and the Ga.  Molten gallium acts as an electrode for ohmic heating and a coolant for quenching simultaneously.  As shown in Fig. 3, Nb(Al)ss filaments are embedded in a pure Nb matrix.  In contrast to the conventional JR conductor, the starting subelement is made by rolling Nb and Al foils around a Nb rod (not a Cu rod).  A bundle of such rods was extruded in Cu sheath that was also removed by etching.  This is because the Cu sheath would melt and react with the Nb/Al composites and form unwanted ternary Nb-Al-Cu compounds during ohmic heating (~1900°C).  Thus the strand has hardly included the Cu stabilizer as a basic constituent composite so far.  Incorporation of the stabilizer has been achieved with several methods as described latter.  The PM and CCE Nb/Al composites are also likely to be useful for this process.

Fig. 3  

Reel-to-reel Ohmic-heating and rapid-quenching apparatus to form RHQT Nb₃Al multifilamentary conductors.  The typical cross-sectional image of as-quenched JR Nb/Nb(Al)_ss composite is given in the figure.　

Microstructures

Fig. 4 shows x-ray diffraction patterns for as-quenched and transformation-annealed JR conductors.  The diffraction patterns of the as-quenched conductor indicate the existence of two kinds of bcc phases, where the lattice parameters are 0.3311 nm and 0.3284 nm, assigned to pure Nb and supersaturated-solid-solution Nb(Al)ss, respectively.  After transformation annealing at 800°C for 10 h, the diffraction peaks of Nb(Al)ss disappear and those of A15 phase appear instead.  The Nb2Al phase does not appear, since such a massive transformation requires no long-range diffusion of constituent elements so as to transform Nb(Al)ss to the Nb₃Al phase without a change of composition.  Thus, the A15 Nb₃Al transformed from Nb(Al)ss has a high stoichiometry.  It is noted that the diffraction peaks at 2θ=27.3° and 48.3° (corresponding plane spacing: 0.326 and 0.188 nm) appear only at the early stage of transformation.  These peaks can be identified as {100} and {111} planes, respectively, for bcc Nb(Al)ss, which are commonly forbidden reflections: a pole figure of a reflection of 2θ=27.3° or 48.3° is completely the same with that of the {200} or {222} plane for bcc Nb(Al)ss.  Appearance of such forbidden reflections indicates an occurrence of sort of pioneer phenomena; the location of Al atom in the bcc lattice, just prior to the massive transformation, is orderly rather than in a random manner at the as-quenched state.  

Fig. 5 shows metallurgical microstructures of the JR conductor after transformation.  Grain boundary "marks" at two length scales are observed for the transversely fractured cross-section of the transformed Nb₃Al conductors.  The large one (several micrometers) corresponds to the old bcc grain boundaries of the as-quenched material, and the quite smaller one corresponds to the A15 grains (several tens of nanometers) which form from these when they are transformed.  Such a multiple structure can be interpreted from analogy with the prior austenite grain boundary in which ferrite is precipitated.  The crystallographic orientation relation of A15 Nb₃Al precipitates in the bcc matrix is given by the simple correspondence [001]A15║[111]BCC; (100)A15║(110)BCC, where the closest-packed planes and the direction of the precipitate are parallel to those of the matrix.  A pole figure of the (100) plane of the A15 phase is almost the same as that of the (110) plane of bcc Nb(Al)ss [52].  Thus, the misorientation of the neighboring transformed A15 grains seems to be quite small inside a prior bcc grain.  Any secondary phases were not observed in the A15 phase filaments.

Fig. 4     

(a) X-ray diffraction patterns for simple RHQT (800°C for 5min, 10 h annealed) JR conductors.  At the early stage of transformation annealing, the {100} and {111} forbidden reflections of the bcc phase are clearly observed.  

(b) Pole figures of the reflections of 2q=27.3° and the {200} plane (2q=56.2°) for bcc Nb(Al)ss.  

Fig. 5 

 Metallurgical microstructures of RHQT JR Nb₃Al conductors. 

RHQ Condition

Fig. 6 shows a variation of Tc of the finally annealed samples at 800°C against the RHQ condition, and the phases identified with the x-ray diffraction and the mechanical properties for the as-quenched state.  If the achieved temperature at self-heating is much lower than that which is most appropriate to form bcc Nb(Al)ss, Al-rich compounds (Nb2Al (σ) and off-stoichiometric Nb₃Al) form.  Such a formation of Al-rich compounds makes the composite brittle, and causes breaking when this composite is transferred to a take-up spool.  When the self-heating condition is slightly lower than the optimum, a nearly-stoichiometric A15 phase forms directly in elevated temperatures during the RHQ process.  The resultant RHQ conductors do not show ductility enough to be wound on a coil with a bore of 50 mm.  Such a directly crystallized and disordered A15 phase shows, however, a high Tc exceeding 18.3 K after the second heat treatment at 800°C for ordering.  Bc2 (4.2K) exceeds 29 T, and Jc ’peak-effect’ occurs.  The solubility limit of Al seems to extend to more than 25% in the RHQ process, due to the high cooling rate.  The optimum self-heating seems to be achieved when the sample temperature reaches that for the maximum solubility limit.  The resultant composites only include the bcc Nb(Al)ss phase, and thus the ductility of the whole composite is ensured.  However, Tc and Bc2 (4.2K) are slightly lower, probably because of the stacking faults.  When the self-heating exceeds slightly the optimum value, Tc falls further to a local minimum.  In such a case, an amorphous-like Nb2Al phase is observed, which indicates that the achieved temperature lies between the solidus and liquidus lines.  Two-phase separation (partial melting) causes the solute redistribution of the Al-poor Nb(Al)ss and the Al-rich liquid which eventually allows the formation of amorphous-like Nb2Al.  The resultant Al-poor Nb(Al)ss transforms to off-stoichiometric Nb₃Al and hence degrades Tc.  However, further excess self-heating completely melts the filament and the subsequent rapid cooling allows the formation of the stoichiometric Nb(Al)ss again. 　

Fig. 6 

 A variation of T_c of the finally annealed samples at 800°C for 10 h against RHQ condition (normalized self-heating current), the x-ray diffraction identified phases and the mechanical properties for the as-quenched state.  

Stabilization

An important and urgent issue with regard to the practical use of RHQT Nb₃Al is the incorporation of stabilizers into Nb₃Al conductors.  The Cu sheaths assembled at extrusion (Fig. 2) must be removed by etching and are usually unavailable as a stabilizer, since the RHQT process includes a short heat treatment around 1900°C which is much higher than the melting points of Cu.  Simple electrical Cu plating after RHQ treatment could not stabilize the Nb₃Al wire sufficiently, because a stable Nb oxide layer is formed on the surface of Nb/Nb(Al)ss composites and prevents the thermal and electrical contact between Cu and Nb.  Thus, we have developed special technologies to incorporate the stabilizer into the RHQT-processed Nb₃Al conductors.  　

Cu-clad conductor

All the methods we have developed for the external incorporation of Cu into the strand surface are based on the ductile nature of the as-quenched JR Nb/Nb(Al)ss composite at room temperature: a compacted-strand cable consisting of Cu wires and JR Nb/Nb(Al)ss strands, a composite of JR Nb/Nb(Al)ss in a Cu tube which is co-drawn to be in one piece, etc.  In the co-drawing (round-to-round) deformation, we could not obtain good bonding between Cu and Nb, whenever the total reduction ratio was less than 90%.  In a so-called ’mechanical-cladding method’ an as-quenched composite is wrapped longitudinally with Cu sheets and groove rolled to ensure mechanical bonding.  Fig. 7 (a) shows the cross sectional structure of such a Cu-clad Nb/Nb(Al)ss composite.  The volume ratio of Cu to Nb/Nb(Al)ss strand (rCu)is 0.45.  The starting 1.26 mm diameter Nb/Nb(Al)ss wire is plastically deformed to an almost rectangular shape, with a total reduction of area, R.A., of 42 % at Cu cladding.  Fig. 7 (b) also shows the variation in Ic and Jc as a function of R.A. at cladding.  Jc was rather enhanced at the beginning of deformation of the Nb/Nb(Al)ss strand; Jc of the Cu-clad Nb/Nb(Al)ss composite is increased from 150 A/mm2 to 335 A/mm2 at 21 T by 42% RA.  Such an enhancement in Jc compensates well for the reduction in cross sectional area; Ic at 21 T is increased by 50% for 26% RA.  Bc2* estimated by the Kramer plot is also increased by 1 T.  Thus, mechanical cladding with an appropriate amount of deformation is an effective way to incorporate a large-volume-fraction Cu-stabilizer.  The shear deformation between Cu and Nb, caused by the round-to-rectangular deformation, should play an important role in breaking the Nb oxide layer.  In comparison with the co-drawing technique, the contact resistance between the Cu stabilizer and Nb/Nb(Al)ss is three orders of magnitude smaller.  Consequently, this method is only useful in making a rectangular stabilized Nb₃Al conductor.  

Fig. 7 

 External stabilization for RHQT JR Nb₃Al by mechanical cladding method:  (a) transversely overall cross-sectional image, (b) Jc and Ic against R.A. at Cu-cladding.  An appropriate deformation at cladding enhances Jc significantly.  

Internally stabilized conductor

Large-current CIC conductors for large-scale applications, however, require round-shape strands.  We have developed another stabilization technique for JR Nb₃Al multifilamentary round wires: internally including the stabilizer as a basic constituent of the strand.  In the case of such internally stabilized JR Nb₃Al wires, the reaction of the stabilizer with JR Nb/Al filaments on ohmic heating must be avoided; the stabilizer should be separated from the JR Nb/Al filaments by a diffusion barrier.  Otherwise, as shown in Fig. 8 (a) and (b), Cu-stabilizer filaments react with the JR Nb/Al filaments and form the unwanted ternary compound including Cu.  A bundle of seven hexagonal stabilizer rods surrounded by an array of 12 hexagonal Nb diffusion-barrier rods is stacked in the center region, together with 66 hexagonal JR Nb/Al rods.  Fig. 8 (c) and (e) show such as-drawn overall cross sectional structures of the JR Nb/Al composites internally stabilized with Cu and Ag, respectively.  The final thickness of the Nb diffusion barrier is designed to be equal to that of the JR Nb/Al filament in this case.  This Nb-filament barrier can prevent the stabilizer from reacting with the JR Nb/Al in the subsequent rapid heating and quenching, so that the ternary compounds Nb-Al-Cu are not formed.  The Cu stabilizer apparently attacks the Nb barrier on ohmic heating and Nb dissolves into the melted Cu stabilizer, solidifying as dendrites dispersed in the Cu stabilizer (Fig. 8 (d)), while the Ag-stabilizer does not react with the Nb-filaments barrier (Fig. 8 (f)).  As a result, the Ag internal stabilizer enables the Nb-barrier thickness to be reduced to much less than 5μm; this is in contrast to a value of 50 μm for a Cu stabilizer.

  According to this value, we have fabricated another internally Ag-stabilized JR Nb₃Al conductor.  The basic specification is the same as that of Fig. 8 (e), except that 12 Nb-barrier rods and 12 JR Nb/Al rods are further replaced with 24 Ag rods jacketed with Nb (Ag to Nb volume ratio: 0.7).  The Nb center rod inside JR filaments is also replaced with the Ag rod.  In such a case, the overall stabilizer/non-stabilizer ratio is increased up to 0.26. Less contamination in the Ag stabilizer has been confirmed by measuring and comparing the residual resistivity.  Despite of only a 0.06 volume fraction of the stabilizer, the overall ρr of the Ag-stabilized conductor is less than half that of the Cu-stabilized conductor at zero field.  It is thus estimated that the resistivity of the Ag stabilizer is much smaller than that of the Nb-dissolved Cu stabilizer.  Such a high-purity Ag stabilizer causes a strong magneto-resistance effect.  This does, however, not degrade the efficiency of the Ag stabilizer, since the stabilizer is required, in particular, at low magnetic field where Jc is quite high.  　

Fig. 8 

  Internal stabilization for RHQT JR Nb₃Al conductors: overall cross sections of as-drawn composites where (a) JR Nb/Al filaments are replaced with Cu stabilizer, JR Nb/Al filaments are separated from (c) Cu stabilizer or (d) Ag stabilizer by Nb-filament barrier, and (b), (d) and (f) magnified cross sections of rapidly heated and quenched composites, corresponding to (a), (c) and (e), respectively.  Stabilizer/non-stabilizer ratio is 0.065 both for Cu and Ag stabilizers.

Jc and Strain Tolerance

The phase transformation enables the formation of Nb₃Al with compositional stoichiometry and fine grain structure.  Thus, an extremely high Jc is achieved over the whole range of magnetic fields, in particular, in high fields more than 20 T, as shown in Fig. 9.  The core Jc and non Cu Jc at 21 T, 4.2 K are 336 A/mm2 and 186 A/mm2, respectively, for a Cu-clad JR Nb₃Al conductor.

 The strain sensitivity of Jc of A15 phases was predicted to increase dramatically with increasing long-range atomic order parameter.  If this is the case, the resulting high stoichiometry would balance against the excellent strain tolerance.  Based on this prediction, we have actually evaluated the strain sensitivity of the RHQT RIT and JR Nb₃Al conductors in comparison with the low-temperature-processed JR Nb₃Al  and the ITER Nb₃Sn conductors at 12 T.  The Bc2* (εi) was larger by about 5 T than that of the low-temperature-processed Nb₃Al, which is known to be off-stoichiometric and of high strain tolerance.  The Bc2* degradation with –0.7 % intrinsic strain is 8 % for the RHQT RIT Nb₃Al, and almost comparable to the low-temperature processed Nb₃Al.  It is, thus, clear that the Bc2* degradation for the RHQT Nb₃Al is also much smaller than 30 % for the ITER Nb₃Sn.  In line with the prediction, the high stoichiometry achieved in the RHQT RIT and JR Nb₃Al conductors seems to have slightly increased the strain sensitivity of the Bc2*.  However, since Jc also strongly depends on Bc2* (εi), at a given magnetic-field (12 T) it looks less sensitive to strain if Bc2* (εi) is larger.  It should, moreover, be sensible from the practical standpoint to compare the strain sensitivity of Jc at 12 T between different conductor, since this is the operational field designed for fusion and accelerator magnets.  The Jc degradation with 0.7 % intrinsic compressive strain was only 20 % for both the transformed RIT and JR Nb₃Al conductors, which is almost the same in magnitude as that for the conventional JR Nb₃Al conductor which has a lower Bc2* by 5 T.  Consequently, the excellent strain-tolerance is compatible with substantially improved high-field performance for the transformed Nb₃Al conductors. 　

Fig. 9

 Comparison of (a) Non-Cu Jc versus B curve and (b) strain tolerance between the RHQT Nb₃Al conductor and conventional Nb₃Al and Nb₃Sn conductors.

Coil Performance

Specifications of a new solenoid coil, which was fabricated by a wind and react technique, using the above-mentioned externally-stabilized Cu-clad RHQT JR Nb₃Al conductor, are summarized in table I.  The Cu-clad RHQT JR Nb₃Al conductor was wound into the ‘new coil,’ of which the inner and outer diameters and height are 19.7, 40.8, and 49.7 mm, respectively, transformed at 800°C for 10 h, and impregnated with beeswax.  The conductor was insulated with Al2O3 fiber, which is mechanically stronger than the glass fiber after heat treatment.  Furthermore, a coefficient of expansion of Al2O3 is closer to that of a metal.  Fig. 10 shows the loading tests of the new coil at 4.2 K and 2.1 K.  The generated magnetic field was measured with a hall sensor set in a bore of the coil.  The hall sensor is calibrated with the back-up field.  The coil, while carrying a current of 179 A at 2.1 K in a superconducting back-up field of 21.2 T, generated an additional 1.3 T.  The coil current capacity increased a factor of six as compared to the previous coil.  The resulting total magnetic field of 22.5 T is the highest record ever for a metallic superconducting coil.  It is noted that the quench current of the coil is larger than the Ic of the short sample.  The Cu stabilizer may induce a compressive strain in the Nb₃Al phase, and such a strain may be released when loading the coil.  Degradation of coil performance has not been observed, although we have loaded the coil in each case until it quenched.  This indicates that the Cu-clad RHQT JR Nb₃Al conductor and its coil technology are performing well. 　

Fig. 10

  Load lines of the new coil wound with Cu-clad RHQT JR Nb₃Al conductor.  Overall J_c is defined as J_c/{(1+r_Nb)(1+r_Cu)}.

New Coil

Previous Coil

RHQT JR Nb₃Al Conductor

 Stabilizer

Cu-clad type

-

 Fraction of stabilizer (r_Cu)

0.45

0

 Piece length (m)

30

137

 Cross section of bare strand

1.61^Wx 0.71^H mm²

0.5 mm in diam.

 Insulator

Al₂O₃ fiber

glass-fiber

 Filament diameter (μm)

90

55

 Number of filaments

84

36

 Nb/Nb₃Al ratio (r_Nb)

0.8

1.5

J_c (4.2 K)@21 T (A/mm²)

278.1

189

  I_c (4.2 K)@21 T (A)

123.5

14.8

Winding

 Inner diameter (mm)

19.7

20

Outer diameter (mm)

40.8

45.6

 Height (mm)

49.7

50

 Number of turns

311

1331

Coil  

 Coil constant (T/A)  

0.00736  

0.0278  

 Impregnation  

beeswax

epoxy resin

 Inductance (H)  

0.00106

0.0226

The TRUQ Process

As shown in Fig. 6, a directly crystallized A15 phase shows a higher Tc exceeding 18.3 K after the second heat treatment at 800°C for ordering, and a Bc2 (4.2K) exceeding 29 T.  The lower values of Tc and Bc2 of the RHQT Nb₃Al conductor may be attributed to stacking faults (shear structures or non-conservative antiphase boundaries) formed in the A15 phase.  The stacking faults cause local compositional deviation from stoichiometry with an excess of Al.  Recently, we have shown that the ordering of the Nb(Al)ss is responsible for the formation of stacking faults, and then developed a new transformation technology termed ‘TRans-formation-heat-based Up-Quenching’ or TRUQ.  TRUQ is characterized by the self-heating of the bcc phase by the transformation heat, which propagates through the whole length of a composite wire and transforms it to Nb₃Al, as shown in Fig. 11.  Annealing around 1,000ºC causes instantaneous transformation from Nb(Al)ss to Nb₃Al, accompanied by a remarkable heat generation equivalent to the change in enthalpy.  This heat started up self-heating and up-quenching the sample wire to some temperature, seemingly higher by a few hundred centigrade degrees than that of adjacent untransformed Nb(Al)ss.  Since a source of heat is limited in principal, the temperature is automatically turned down to the ambient one in a short time.  This TRUQ involves the propagation of the transformation interface in the form of a transformation wave which is sustained by the large enthalpy release in the transformation front.  A subsequent annealing at 800°C enhances the long-range ordering of the Nb₃Al phase and drastically improves the high field properties of Nb₃Al conductors: Bc2(4.2K) is 29 T and the Jc(4.2K) values are 1160 A/mm2 at 17 T and 260 A/mm2 at 23 T, respectively (Fig. 9).  These values are clearly much larger than those of the above-mentioned ordinary RHQT JR Nb₃Al conductor, and they indeed approach the specifications of the 1GHz NMR magnet.  An attempt shall be made at the TML/NRIM to apply the TRUQ process to a wind-and-react coil with a special furnace utilizing a fluidized bed process.  

Fig. 11

  Illustration of TRUQ: (a) the propagation of the transformation interface from the left to the right in a Nb/Nb(Al)_ss composite quickly pushed into 1000°C region in a gold-mirror furnace, (b) the enthalpy release during the transformation from bcc to A15 and (c) the temperature profile for the whole heat treatment process.　

New Process for Nb₃(AlGe) Conductor

(under construction) 　

Related Review Articles

• Takeuchi T, Nb₃Al wires appraoching to practical applications, Web21 July 15, 2003.
• Takeuchi T, B3.3.4 Processing of lLow T_c compounds: the compound Nb₃Al, HANDBOOK OF SUPERCONDUCTING MATERIALS, Ed. D Cardwell, University of Cambridge, UK; D Ginley, NREL, USA Dec 2002, IOP pp. 673-684.
• Takeuchi T, Nb₃Al superconductors, IEEE Trans. Appl. Superconductivity vol.12 (2002) pp.1088-1093.
• Takeuchi T, Nb₃Al conductors - rapid-heating, quenching and transformation process - IEEE Trans. Appl. Superconductivity vol.10 (2000) pp.1016-1021.
• Takeuchi T, 2000, Nb₃Al conductors for high field applications, Supercond. Sci. Tech. vol.13 pp. R101-R119.
• Takao Takeuchi, Characteristics of Nb₃Al multifilamentary superconductors prepared by phase transformation from a supersaturated solid solution, Teion Kogaku vol. 33 (1998) pp. 614-622 (in Japanese).
• Takao Takeuchi, Metallurgy of multifilamentary superconductors – Fabrication of new materials by designing structure and controlling diffusion reaction – Matelia vol. 35 (1996) pp. 220-224 (in Japanese).  　

Interesting or Useful New WWW sites