Background of the Invention

This invention relates to superconducting states and processes and products useful in achieving superior superconducting properties and results.

Of the two commercially available and useful bulk superconductors, NbTi and Nb ₃ Sn, NbTi is used in more applications because it is more ductile while Nb ₃ Sn has superior superconducting properties but, in comparison to NbTi, its useful forms have generally had less desirable mechanical properties.

The absence of good ductility characteristics of Nb ₃ S type superconductor structures has been substantially overcome by the so called "bronze" process, whereby the niobium and Sn are processed as separate metallurgical phases in a composite. A reaction heat treatment is applied to the finished wire. During this treatment, tin from the copper-tin bronze alloy reacts with the niobium to form the A-15 crystal structure of ₃ Sn at temperatures below 930°C. Copper is required as another element to stabilize or catalyze the formation of b ₃ Sn over more tin-rich niobium phases.

Typical structures and methods for their preparation are shown in U.S. Patents 4,262,412 and 4,414,428 to McDonald, the descriptions of which are incorporated herein by reference in their entirety.

So called modified jelly roll (MJR) as described in the foregoing patents, is incorporated herein by reference and is used in the processes for preparing the products of this invention.

While variations of many of the bronze processes exist including those which comprise alloys with various

other elements such as Ga, Ti, Mg ¹ ' ² , and the like, and the niobium filament can be further alloyed with small amounts of Ti, Ta, Zr, Hf, Sn, Fe, and V, the overall current density in such structures is affected, i.e., diluted, by the amount of tin-containing bronze that is required to convert substantially all the niobium to Nb ₃ Sn.

Increased ductility in wire drawing processes is achieved by separating the tin and copper as distinct phases which can be heavily worked without the annealing normally necessary with bronze after working.

In addition to the improved ductility, the separation of tin and copper allows the volume fraction of niobium to be increased while maintaining an adequate volume of tin for complete reaction. Thus, the Nb content of the non-copper volume of the wire, and subsequently Nb ₃ Sn, is less diluted by the non- dispersable metallic content.

There can be further improvements where, as in the above-described M R process, the tin, copper and niobium materials required to form the Nb ₃ Sn are surrounded with a layer of barrier material which is selected from materials which effectively prevents diffusion of tin from the tin core into the copper matrix of the wire remotely located from the niobium and the subsequently formed Nb ₃ Sn filaments. Wire thus formed contains both copper and non-copper areas, after

^l -I.W. Wu, et al, "The Influence of Magnesium addition to the bronze on the critical current of bronze - processed multifilamentary Nb ₃ S ", IEEE, Trans Mag-19, pp. 1437-1440, 1983.

² K. Togano et al, "Effects of Magnesium addition to the Cu-Sn Matrix in the Composite-Processed Nb ₃ Sn Superconductor", Jour. of Less-Common Metals, 68, (1979), 15-22.

reaction. Where the tin has reacted, substantially completely, both with the niobium and all the copper within or inside the barrier layer, the Nb ₃ Sn filaments are surrounded by bronze. While magnesium has been used heretofore in the production of niobium-containing products by the bronze process, that use causes substantial unwanted mechanical and physical anomalies resulting in a loss of ductility and poor workability when fabrication is attempted as well as problems with the higher temperatures required for processing.

As previously described, the most commonly used Nb ₃ Sn wire product is made by the bronze process.

In many applications, magnetization, by definition, is a concern. A superconducting body has significant magnetization because of its ability to pin fluxoids (which is also why it can carry high currents) . The magnetization at any field will be directly related to J _c (the critical current density) . Magnetization (M) is also directly related to the superconducting area and hence the filament width.

M J _c d where d = filament diameter While it would appear obvious that M can be reduced by decreasing d, the reality is different. As the diameter is decreased, so is the spacing between filaments. During the heat treatment, closely spaced filaments may "bridge" together creating an effective filament diameter which is much larger (20-100 times) than an individual filament. This undesirable result is compounded by the expansion of the filament on conversion from Nb to Nb ₃ Sn. From measurements of M and J _c a d _e (effective filament diameter) can be calculated.

One object of the present invention is to increase and enhance the critical current of Nb ₃ Sn wire

or rods while reducing the amount of bridging between filament elements.

Still another object is to control the growth of

Nb ₃ Sn grain size and insure the homogenous unimpeded distribution and diffusion of small grains by employing the conditions and structure which provide a homogenous distribution of Mg along the length of wire.

Yet another object is to produce, in situ at the locus of the niobium filament, finer grain growth of smaller grains of N ₃ Sn (after reaction) and Mg ₂ Sn (before reaction) to provide a wire structure capable of being worked into filamentous wire or rods for superconducting products.

A further objective is to provide a method of fabricating a wire structure capable of achieving the superconducting state while exhibiting a Nb ₃ Sn current density greater than about 2600 Amps/mm ² at 10 T.esla

4.2°K and 10~ ¹³ ohmmeters.

A still further object of the present invention is to reduce the occurrence of bridging between filaments to thereby maintain as small an effective diameter of filament as possible after fabrication and reaction.

These and other objects are all accomplished while still retaining the fabricability of the metallic products such as wire and rods by the addition of small amounts of magnesium to the tin core of the tin core MJR. We have unexpectedly found that magnesium content at the eutectic composition ( 98% Sn, ~2% Mg) is enough to produce a dramatic increase in current density (overall) of about 20-30% without further dilution of the Nb ₃ Sn current density, and produces a very fine 2(d) phase Mg ₂ SN and the desired homogeneous dispersion by the radial diffusion of the eutectic either inwardly or outwardly.

Summary of the Invention

A tin-magnesium eutectic alloy is provided in combination with a copper matrix in a superconductor billet, bundle or wire structure such that the formation of excessively large intermetallic compositions, deleterious to current carrying ability and formability of the composite are avoided. This invention also provides a process capable of substantially and homogeneously distributing the magnesium alloy along the length of wires, rods and other products. As will be further described herein during the reaction heat treatment to produce Nb ₃ Sn filaments, the magnesium is uniformly absorbed into the Nb ₃ Sn at its locus of formation, resulting in a refined grain structure for the Nb ₃ Sn and an improvement in the current carrying capability of the wire product over a wide range of applied magnetic field strengths.

Brief Description of the Drawings

Figures 1A, IB, 1C and ID are schematic representations of the stages of the modified jelly roll process described herein for producing an improved Nb ₃ Sn containing superconductor;

Figure 2 is a graphic representation of how the variables of applied field, temperature and current flow interrelate to the formation of the superconducting state;

Figure 3 is a graph of voltage across a sample as a function of applied current at a given magnetic field strength used to define the critical current of the sample;

Figure 4 is a graph showing the hysteresis loss

(integrated magnetization over + 3T cycle) against critical current (J _c ) at 10 Tesla for various materials;

Figure 5 is a graphical representation showing integration of magnetization as a function of applied

field for a production superconductor described herein showing the hysteresis loss for a single cycle;

Figure 6 is a graphical representation of magnetization as a function of applied field for the Nb3Sn superconductor of the present invention showing the hysteresis loss for a single cycle;

Figure 7 is a photomicrograph of the wire produced by the process shown schematically in Figure 1, before heat treatment; Figure 8 is a photomicrograph of the wire of

Figure 5, after heat treatment; and

Figure 9 is a partial cross-section of an alternative embodiment of Figure 1A.

Detailed Description of the Drawings Referring to Figures 1A, IB, 1C and ID, the schematic representation shown is of the results of the fabrication steps in what will hereinafter be termed the modified jelly roll (MJR) process. A solid tin alloy core 10 is provided, which, according to the invention described herein, is a tin-magnesium eutectic alloy material containing from between about 1% to about 3% by weight of magnesium. Spirally wrapped around the tin core are overlapping layers of an expanded metal 11 and layers of copper sheet 12. The expanded metal consists of niobium metal or various alloys of niobium metal including alloys with small amounts of titanium, tantalum, zirconium, hafnium, tin, iron, and vanadium.

The expanded metal wrap 11 is separated from other expanded metal by the thin sheet of copper 12. Typically, at least the first layer of wrap in contact with the tin core 10 will be a copper sheet 12 followed radially outward by an expanded metal sheet or wrap 11 and then alternately followed by a further copper sheet and then expanded metal repeatedly. More preferably, as shown in Figure 9, several wraps of copper are wound

around the tin or tin alloy core before beginning the alternative niobium expanded metal and copper sheet construction shown at Figure 1A. The cylinder or log of Figure 1A is then given an outer wrap or wraps or cylindrical covering of a diffusion barrier 15 made from material such as vanadium, niobium, tantalum, or mixtures thereof, either physically mixed or alloyed. Preferably the tin alloy core and the annular copper wraps can occupy in the annular space around the tin core greater than about 12% of the space (area) contained within the diffusion barrier. The barrier containing structure B is then inserted into a cylinder of copper 16 to form a billet structure shown in Figure 1C. The billet is then drawn repeatedly to the structure shown in Figure ID. Typically, the external hexagonal shape is provided and the final rods are about one-half inch in diameter. The hexagonal shape provides for better packing of the rods into a copper cylinder 17. The cylinder shown at Figure IE is completely filled with rods from Figure ID and additional drawing is employed to reduce the entire structure into a wire of from one-twentieth of an inch to one one-hundredths of an inch in diameter, which is shown in the photomicrographs of Figure 7. Figure 8 shows a cross-section of the wire of Figure 7 after the step of heating the wire for at least 100 hours at a temperature of from between about 600°C to about 800°C. Preferably the heat treating process is initiated at 200°C to about 220°C for about 100 hours, followed thereafter by heating to about 340°C for about 48 hours and then heating to less than 800°C for 100 to 200 hours. More preferably, the final heating step is conducted at between about 600°C to about 700°C and, more preferably, at about 675°C for at least 100 hours.

Referring to Figure 2, a surface is depicted wherein a point on the surface defines the critical state where superconduction transition occurs. The

relationships describe a point where critical current density J _c , critical field H _c and critical temperature T _c and define the superconductivity boundary conditions for the sample being evaluated. Therefore, any set of conditions defining a point between the surface shown and the planes of the axes or origin describes a material in the superconducting state. Simply stated, the surface defines the following emperical relationships, i.e., the critical temperature (T _c ) will decrease with an increase in value of either current (J) or field (H) ; and critical current (J _c ) will decrease in value when either temperature (T) or field (H) increases; and critical field (H _c ) will decrease in value when either temperature (T) or current (J) increases. Referring to Figure 3, the graph shows a typical transport current measurement which is used to evaluate different materials by a conventional method as defined in ASTM Specification B714-82 "Standard Test Method for D-C Critical Current of Composite Superconductors", ASTM Standards, Philadelphia, ASTM, 1982, which is incorporated herein, in its entirety by reference.

Wires were drawn from a control billet and billets comprising embodiments of the present invention wherein the core rods were respectively:

(1) Sn/0.7 wt. % Cu control (1087)

(2) Sn/1.8 wt. % Mg (1133) and

(3) Sn/2.0 wt. % Mg (1202)

The wires were all drawn to 0.0307" and given the same heat treatment (optimum for 1087) . 220°C/175 hr + 340°C/96 hr + 650°C/200 hr and exhibited the following characteristics:

Unreacted Filament size -3.2 microns Diffusion barrier Vanadium Local Matrix/filament ratio -1.7:1 Filament Alloy Nb 0.8 wt% Ti; matrix pure copper

# MJR elements in wire 18 (except 1202, having 14 elements)

Fraction Non-copper in wire -46% (except 1202, having

35.5% non-copper)

For each of the samples, the volume fraction of non-copper content (volume inside barrier layer) was as follows:

TABLE I

(copper around the filament)

Critical current densities are obtained for each wire sample as follows:

TABLE II

Non-Copper Area (A/mm J_

5 Tes la 8 Tesla 10 Tesla

0.1 V/cm 10 13 0 ■ 10 ^" 13 0.1 10 10 13

1926 2059 1089 1161 758 795 840

2556 2689 1443 1509 978 1019 828 1897 2000 1056 1097 717 738 701

(predicted)

Where "n", in the formula V rn is a description of the sharpness of the curve, reflecting uniform filaments as shown in Figure 3 of sample current in a perpendicular field.

TABLE III

Hysteresis Loss (0.78 mm)

(Non-Cu) (Non-Cu)

Lot# + 3T Cycle Fraction mJ/cm ³ Loss/Jc(10T) 1087 218 .463 471 0.592

1133 399 .4625 863 0.847

1202 75 .355 211 0.286

Observations of reacted MJR elements reveal that the bridged filaments are located closest to the core and hence to the tin source. Thus, it is likely that the hysteresis loss is controlled by these filaments closest to the core. The gradient in tin concentration will result in a gradient in Nb ₃ Sn current density. The magnesium is unlikely to result in a change in the effective filament size while it is expected to increase the critical current density of filaments incorporating it.

Comparing 1087 and 1133 where only the 1.8 wt% Mg is the difference, the following results are obtained:

Ic (1133) = 1.28 10T, 10~ ¹³ m

Ic (1087)

Hysteresis (1133) = 1.83 +3T cycle Hysteresis (1087) Assuming d _e ff is unchanged, then the loss data indicates that J _c in the bridged filaments has been increased by 1.83 times while the overall current density (average) has only been increased 1.28 times. Simple algebra then shows that about 34% of the filaments must contribute to the loss and the benefit of Mg is clear.

Sample 1202 was subjected to similar reaction conditions as 1133 and 1087 and better filament separation was achieved. The result was significantly lower (factor of 2 over 1087) [factor of 3 over 1133] + 3T hysteresis loss and magnetization.

However, for sample 1202, the J _c enhancement was not as great as expected though no control of the same geometry was run. It is believed that even though less Nb was employed, the increased amount of matrix around the internal filaments may have added significant pre-stress not normally encountered in the standard tin core MJR wires as described.

These results are graphically depicted in the Figures 5 and 6. Further analysis of the foregoing data can provide a better understanding of the improvement provided by the use of a tin-magnesium eutectic in any structural form. Keeping in mind that the formation of the Nb ₃ ≤n produces a volumetric expansion of approximately 1.35 times the volume of the original niobium, the following expression can be applied:

J _c (Non-Copper) = J _c (Nb ₃ S ) (minimum average)

% Nb in Non-Copper x 1.35

10 Tesla (10" ¹³ ) 4.2K

Wire %Nb in Non-Cu JJ _cc ((NNoonn--CCuu)) JJ _cc ((NNbb ₃₃ SSnn)) 1087 22.9 795 A/mm 2571 A/mm ²

1133 23.4 1019 3226

1202 21.1 738 2591

Taking the foregoing data, we can calculate directly Nb ₃ Sn area (maximum potential) :

Nb-Sn Area = wire area x non-copper fraction x ZNb filament in non-copper x 1.35

This provides us with an indication that A15 grown by the diffusion of the present invention using the critical current density of the Nb ₃ Sn at 10~ ¹³ m

criteria, 10 Tesla, 4.2 Kelvin will be as large or larger than a value of about 2600 A/mm ² .

While the foregoing description identified many variable that must be balanced in terms of useful metals and alloys for the various structural elements, including filament composition barrier layer composition core composition and the like, the described relationships in superconductor characteristics are still employed to evaluate the performance of the wire and provide the basis for deducing what changes can be made to produce predictable changes in specific performance and fabricability. It is contemplated, therefore, that the tin-magnesium eutectic core composition described herein can be employed with pure niobium or niobium alloyed with 0.05 to 10 Wt. % Ta; 0.05 to 5 wt. % Ti; 0.05 to 10 wt. % Hf; 0.05 to 10 Wt. % Hf plus 0.05 to 5 wt. % Ti; 0.05 to 7 wt. % Zr; 0.05 to 7 wt. % Zr plus 0.05 to 5 wt. % ti; 0.05 to 10 wt. % Ta plus 0.05 to 5 wt % Ti; and 0.05 to 7 wt. % Ta plus 0.05 to 7 wt. % Zr. Likewise, as previously described, the unexpected improvement in conductivity and current density where less niobium is used and the copper content is increased around the tin core, can be expected to have utility with a wide variety of compositions and structures. It is anticipated that the improvements observed are obtained when the annular area of copper around the tin core is greater than 12% of the area within the barrier layer.

This invention has been described with respect to specific embodiments thereof. It will be readily appreciated that the scope of the claims appended hereto is to be limited only by the applicable prior art and that a breadth of equivalents are possible where the described parameters for properties are produced.