Technical Background There are strong commercial reasons driving the development of superconducting materials having high superconducting parameters, such as critical current density, Jc, critical temperature, Tc, and upper critical field, He2. These materials find applications in, for example, Nuclear Magnetic Resonance (NMR) and Nuclear Magnetic Imaging (NMI) magnets and also in cryogen-free magnets.

The microstructional requirements for such materials are very demanding and many approaches and many alloys or compounds have been developed to address the problems of improving materials performance, including mechanical and electrical performance for example, and ease and cost of fabrication.

An important superconductor material is Nb3Al which forms an A15 superconducting phase and, by way of illustration, methods for fabricating this material include the following.

The fabrication processes can be considered in three groups : low-temperature, high temperature and

transformation processing. In each case, an A15 Nb3Al strand is processed by first making the final-size strand, with the constituents subdivided, and then heat-treating it to form the A15 phase.

Low Temperature Processing Low temperature (<1OOOC) processes ensure that the grain size of Nb3Al does not become too coarse because the Nb/Al constituents directly react with diffusion to suppress Nb3Al grain growth. But, at low temperatures, there is a deviation from A15 stoichiometry, thus affecting high-field properties, especial J,. Low temperature processes include the following.

Jelly-roll (JR) -Alternate foils of Nb and Al are wound onto a copper rod and inserted into holes drilled in a copper matrix before drawing to form the final-size strand.

Rod-in-tube, (RIT) -An alloy rod is inserted into a Nb tube and drawn down. A triple stacking operation gives the desired lOOnm core diameter to match that of the Al layer.

Clad chip extrusion, (CCE) -A three-layered clad foil, of Al/Nb/Al, is cut into square chips, and then filled into a can in order to be extruded.

Powder metallurgy, (PM) -A mixture of hydride- dehydride Nb powder and Al powder are put in a copper tube can so that they can be extruded and drawn into a

monofilament wire. A bundle of these wires will give an Al layer thickness of lOOnm.

Despite differences in cross sectional area, Jc versus magnetic field curves are very similar for all of these processes. JR may have a slight advantage for producing long piece strands.

High Temperature Processing At low temperatures the Nb3Al phase will not be completely stoichiometric. However, at high temperatures (>1800C), a diffusion reaction of Nb/Al composites with laser or electron-beam irradiation allows stoichiometric A15 phase formation. Annealing at ~700C improves long range order, and so better Te and H, 2. Unfortunately, high temperatures will cause very coarse grains in the conductor, thereby destroying low field properties.

Transformation Processing This process, discovered in the last few years, involves a combination of rapid quenching and annealing.

The Nb/Al composite is quenched from 1900C to form a bcc supersaturated solid solution of Nb (Al) ss and then transformation annealed below 1000C. This process will produce Nb3Al that is highly stoichiometric and has a fine grain structure, and therefore the Jc will be high at both low and high fields. The most common method of transformation processing is known as the rapid-heating, quenching and transformation method (RHQT), which produces lengths of conductors a few hundred metres long.

Summary of the Invention The invention provides a method and an apparatus for fabricating materials as defined in the appended independent claims. Preferred or advantageous features of the invention are set out in dependent sub-claims.

The invention uses a process for extracting metals and alloys from solid compounds by direct electrochemical reduction, or electrodecomposition, in molten salt, known as the Fray-Farthing-Chen Cambridge process (FFC), as one of a series of steps to fabricate a material. The FFC process is described in the present applicant's earlier International patent application PCT/GB99/01781 which is incorporated herein by reference. The FFC process allows -the treatment of a solid material, which may be a compound between a metal (or semi-metal) and a substance (such as an anionic species), or a solid solution of the substance in the metal, by electrodecomposition in a molten salt to remove the substance from the solid material. On completion of the process, the solid material has been converted to the metal. Alternatively, if the solid material comprises more than one metal, being for example a mixture of metal compounds, or a mixture of a metal and a metal compound, or comprises a solid solution of metal compounds, then on completion of the process an alloy or intermetallic compound of the metals is formed.

The product of the FFC process is typically porous and, in the method of the present invention, is then

infiltrated with an element, metal or alloy, typically as a liquid, to form a material which can be used or further processed to fabricate a product.

In a preferred embodiment, the invention may be particularly efficacious for fabricating superconductors.

For example, if the FFC process is performed on a preform comprising a mixture of powdered Nb20S and Tio2, a porous sample of NbTi alloy is produced. This can then be infiltrated with molten Al to form a material which can be further processed, for example by deformation and heat treatment, to form a high-performance superconductor, advantageously at lower cost than for conventional methods. In an alternative embodiment, the FFC process is performed on a preform comprising a mixture of powdered Nb and Sn oxides. A Nb3Sn superconductor can then be fabricated.

Thus, the invention may advantageously provide a method having four steps as follows for fabricating Nb-based superconductors: 1) electrochemical reduction of the Nb-based compounds, 2) infiltration by Al-based alloys (or any other elements or alloys to form intermetallics or artificial pinning centres [APCs]), 3) deformation, and 4) reactive formation of an intermetallic layer followed by insulation processing.

Of course, the method of the invention envisages any suitable starting material or materials and not only Nb and Al. In addition it should be noted that the invention relates particularly to steps 1 and 2 of the list above and that steps 3 and 4 may be replaced by any appropriate superconductor fabrication techniques.

It should also be noted that the invention is not limited to the field of superconductor fabrication but relates primarily to the technique of infiltrating a porous material formed by the FFC process. The FFC process is very flexible and can produce a wide range of metals, semi-metals, alloys and intermetallic compounds, including materials which are difficult to fabricate in other ways. The additional novel step of infiltrating a product of the FFC process, which is typically porous, with a metal or other material may advantageously allow the fabrication of a wide variety of novel and useful materials compositions and microstructures.

The infiltration step may be carried out ex-situ or, preferably, in-situ. The FFC process can produce a porous alloy or intermetallic immersed in a molten salt. In the in-situ process, the molten salt is contained in a bath which also contains the molten material for infiltration.

The infiltration material will usually be denser than the salt, in which case the salt will float on the infiltration material. After the FFC process is completed, the porous sample can then move directly from

the salt into the infiltration material, which can displace the molten salt and infiltrate the porous sample.

If the infiltration material is less dense than the salt then it will float on the salt but as long as there is contact at an interface between the two, the porous sample can move directly from the salt to the infiltration material, advantageously avoiding contact with any other substances.

In an alternative implementation of in-situ infiltration, the porous sample may be immersed in the infiltration material by moving the interface between the salt and the infiltration medium rather than by moving the porous sample. For example, after completion of the FFC process the bath containing the salt may be flooded with infiltration material to displace the salt, or where the bath contains both salt and infiltration material, the bath may be moved, rather than the porous sample.

In ex-situ infiltration the molten infiltration material is held in a separate bath from the molten salt and the porous sample moves from one bath to the other for infiltration. If this is done in an oxidising atmosphere, disadvantageous oxidation of the porous sample may occur.

An inert atmosphere may be used to alleviate this problem but nevertheless contamination of the porous sample may be more likely than with the in-situ method.

In an alternative implementation of the ex-situ method, following completion of the FFC process the porous sample is withdrawn from the molten salt and allowed to

cool, preferably in an inert atmosphere or in vacuum above the molten salt. As the sample is withdrawn, much or all of the salt within pores in the sample is retained and then solidifies. The sample is then transferred to a pool of molten infiltration material, where it is immersed and the salt melts and is displaced by the infiltration material to infiltrate the sample. This implementation has the advantage that the solidified salt in the pores of the sample during transfer to the immersion material helps to protect the sample surface from contamination or oxidation.

In various embodiments of the invention, where the infiltration material wets the FFC product better than the molten salt, it may advantageously substantially entirely displace the salt from the porous FFC product. In general, molten salts wet metals relatively poorly and so, where the FFC product is metallic and the infiltration material is also metallic, the infiltration material will tend to wet the FFC product more strongly than the molten salt. In other embodiments of both the in-situ and ex-situ methods, provision may be made to enhance infiltration and to ensure that the infiltration material fills the pores in the porous sample, displacing the molten salt as much as possible. One method for this is to pump the molten salt out of the porous sample after or as it is immersed in the infiltration material. A second method, which may be combined with the first, is to

vibrate or agitate the porous sample or the infiltration material, for example by using an ultrasonic transducer.

Specific Embodiments and Best Mode of the Invention Specific embodiments of the invention will now be described by way of example, with reference to the drawings, in which; Figure 1 illustrates an electrolytic cell for carrying out the FFC process ; Figure 2 illustrates the infiltration and subsequent steps in a first method embodying the invention; Figure 3 is a micrograph of a sample of porous Nb alloy following the FFC process ; Figure 4 is a micrograph of a sample of Nb alloy following infiltration; Figure 5 is a micrograph of a Nb-Al-Ge (X) wire following mechanical reduction and diffusion treatment ; Figure 6 is a plot of AC susceptibility against temperature for Nb and NbTi rods embodying the invention; Figure 7 is a plot of AC susceptibility against temperature for reduced Nb205-SnO2 rods embodying the invention ; Figure 8 illustrates a cell for in-situ infiltration according to an embodiment of the invention ; and Figure 9 illustrates a second stage in the in-situ infiltration method using the cell of figure 8 ; and

Figure 10 is an element distribution plot for an infiltrated pellet of niobium oxide.

The electrochemical reduction route of the FFC process may advantageously be a much easier, quicker and cheaper way to extract many metals and alloys than established metallurgical routes. A schematic of such a process is presented in figure 1.

Figure 1 shows an apparatus for making the binary alloy NbTi. It comprises a cell 2 containing molten CaCl2 4. A graphite anode 6 and a rod-shaped preform 8 of mixed Nb205 and Tir2 are immersed in the salt. The preform is supported on and electronically connected to a Kanthal wire 10. The preform is made by mixing powdered Nb20s and TiO2 in the desired proportion, slip casting and optionally partially sintering the mixture.

Other techniques may also be applied to the fabrication of the preform. For example a polymer binder may be added to improve the slip-casting process and the polymer then burned off. Alternatively a prefabricated polymer matrix may be used to make the preform. In this case the polymer matrix is infiltrated with metal oxide powders and then the polymer is burned off. These techniques advantageously form porous preforms, which can be easily treated by electrodecomposition in the FFC process.

In the FFC process, the preform of mixed Nb20s, Trio2 powders is made the cathode in the molten CaCl2, whose

cation can form a more stable oxide, CaO, than Nb205 and Ti02. The oxygen in the Nb20s and Tir2 mixture is thus ionised and dissolves in the salt, leaving Niobium- Titanium alloy metal of the desired composition behind.

The extraction of Nb, NbTi, and Nb3Sn metals and alloys from oxides using this process has been carried out on a laboratory scale, as has the extraction of many other metals and alloys from their compounds.

After electrochemical reduction, the final product of the FFC process in the embodiment is a porous, rod-shaped, metallic sponge of NbTi alloy, as shown in figure 3.

Rapid oxidation of the Nb-based porous rod normally takes place after its removal from the chloride bath and may have a detrimental effect on its surface quality and any subsequent processing. Therefore a different approach is proposed to provide better infiltration conditions, using either in-situ or ex-situ infiltration as described below.

Porositv The degree of porosity of the final percolative network of the Nb-based alloy sponge'depends on the density of the oxide preform and on the initial preparation technique of the prefabricated oxide. For example, preforms which are sintered show a significant shrinkage (increased density) and greatly increased strength in comparison with those prepared by slip casting only.

Materials Considerations in Superconductor Fabrication Because the metallic product of the electrochemical reduction (for NbTi or other materials) is soft and porous, without structural defects or secondary phases, it can not be regarded as a high critical current density, Jct superconducting material.

Such semi-finished product has to be upgraded by introduction of, for example, Al or Sn alloys for the reactive diffusion formation of the intermetallic phase in the case of A15 intermetallic superconductors, and/or by introduction of Artificial Pinning Centres (APC) as in the case of NbTi.

Infiltration of a porous sponge of an alloy such as Nb-Ti-X or Nb-Al-Ge-Z by Al or Al alloy material, where X and Z represent additional metals, has its advantages because for example NbTiTa ternary alloys have recently been explored as a high-field avenue for APC materials.

The infiltration route is adopted in order to obtain a relatively ductile material that can be mechanically deformed into a fine wire, containing an interconnected network of superconducting filaments of the order of 1 micrometre in diameter. Additional advantages of the infiltration process are due to the fact that the composition of the infiltrant (Al-Ge or Al-Si alloys and eutectics) can be readily controlled. The Al-Ge system forms a low melting point eutectic (424C) at the composition of 70% Al-30% Ge. The inherent brittleness of

the Al-Ge eutectic when solidified requires particular attention to the temperature of the infiltrating bath, the temperature difference between the porous niobium or alloy rod and the bath, and the rate of infiltration. The -formability of the Al-Ge eutectic, entrapped and solidified within the pore volume of the rod, during form rolling or wire drawing can be significantly improved through the application of superplasticity principles.

Superplastic behaviour requires a fine duplex microstructure that is stable at the deformation temperature.

The invention may thus advantageously provide new techniques that allow manufacture of complex compositional superconducting alloy rods by ex-situ and in-situ infiltration processes.

Ex-situ infiltration The ex-situ infiltration process is conducted in vacuum or an inert gas container in such a manner that as many as possible of the pores in the sample are filled as completely as possible with molten metal or alloy. This maximises the volume fraction of superconductor realisable from a given amount of infiltrated alloy. A schematic representation of an embodiment of the process following the fabrication of the FFC process alloy sponge is shown in figure 2, which shows the steps of infiltration 50, here in a Sn/Ga/Al infiltrant bath 52, cladding 54, mechanical reduction 56 and diffusion processing 58 in a

furnace 60. Figure 2 relates to (Nb, X) 3 (Sn, Al, Z) wire processing, using porous (Nb, X) rod fabricated using the FFC process, for example.

In the embodiment of figure 1 described above, a NbTi alloy was formed by direct electrochemical reduction to form an alloy sponge, or rod-shaped sample. Its microstructure is shown in cross-section in figure 3. For infiltration, the Nb-based alloy rod is immersed in a bath of molten Sn or Al-based alloy maintained at a temperature above melting. Lower temperatures are preferred in order to prevent the extensive, very often rapid, formation of brittle intermediate phases which could impair the ductility of the composite infiltrated material. The microstructure of the infiltrated alloy sponge is shown in figure 4.

Taking into account that the final products of reducing Nb-oxide-based oxide mixtures in the FFC process typically have pore sizes in the range of 2-20ism, special care should be taken to ensure that the Nb-based sponge' surface is as clean and pure as possible to enable complete infiltration of the porous rod, efficient wetting by the infiltrating metal or alloy such as Sn, Al etc. and finally minimisation of superconductor filament damage caused by the formation of hard Nb20s (or even of more complex insulating compounds) on the sponge surface during removal of the metallic Nb-based rod from the chloride bath. In some circumstances the oxygen content in the Nb can reach 2%-3at. %, which is about the solubility limit at

the extrusion temperature used later in processing.

Oxygen adsorbed at the surface of the particles in the sponge may also diffuse into the Nb, increasing its microhardness to 3500 MNm~2. The plastic deformation of the composite may then not be uniform because of severe solution hardening of the Nb, mainly. due to interstitial oxygen. A successful co-deformation of Nb and Sn particles requires sufficient reduction of the oxygen content in the Nb. If the oxygen content in the Nb is reduced to O. lat% the microhardness of the Nb matrix drops to-120OMNm, which is about the value of the surrounding Cu matrix in which the Nb material is typically subsequently encased and which is used for cryostability.

Three alternative purification methods have been tested as follows: (1) because the oxygen solubility in Nb decreases with temperature, part of the oxygen can be precipitated in the form of oxides during an annealing treatment at 600-700C. This treatment has indeed proved to be beneficial but the deformability of the Nb particles may still be insufficient; (2) a reduction treatment of the porous Nb-based composite in an H2 or CH4 atmosphere may improve the deformability of the Nb; (3) an additional component Q can be added to the Nb205-Q powder mixture, which has a larger binding enthalpy for oxygen than Nb does (390 kJ/g-atom 0: interstitial solution). A variety of additives Q can be used as the skilled person would appreciate.

In summary therefore, ex-situ infiltration can be used to fabricate superconductors but care needs to be taken to avoid deleterious oxide formation, which may require additional processing steps and add to the complexity of the method.

A variant of ex-situ infiltration aims to address these concerns. In this embodiment, after completion of the FFC process the porous metal or alloy sponge is withdrawn from the molten salt using a control and positioning system until it is held in an inert atmosphere or in vacuo above or near the salt bath. If wetting of the metal by the salt is sufficient, the pores in the sponge remain filled with molten salt, and the sponge can be cooled to solidify the salt. The sponge surface is thus protected against contamination or oxidation by the presence of the solid salt and can be transferred to an infiltration bath without damage. On immersion in the infiltration bath, the salt melts and is displaced by the immersion material.

Molten salt can also be more effectively retained in the porous metal by lowering the temperature of the salt bath, prior to removal of the metal, to close to the melting point of the salt.

In-situ Infiltration Figures 8 and 9 illustrate the technique of in-situ infiltration. A cell 20 contains molten salt 22 floating above a molten metal alloy 24 held in an extended lower

portion 26 of the cell. As in the embodiment previously discussed for fabricating a NbTi based superconductor to carry out the FFC process, a Nb205/TiO2 preform 28 and a graphite anode 30 are immersed in the molten salt, which is CaCl2. The preform is supported on a tubular Kanthal support 32.

Following use of the FFC process to reduce the preform to a porous alloy sample or rod, which in this embodiment is a Nb-based alloy rod, in-situ infiltration of the porous rod is carried out by lowering it from the molten salt directly into the molten metal, (in the embodiment, molten Al) beneath, as shown in figure 9. The rod is lowered by a control and positioning system coupled to the Kanthal support. This in-situ process is advantageous because there is no direct contact of the Nb with oxygen before infiltration, and so the metal surface on infiltration is oxide free. During infiltration molten CaCl2 is displaced from the sample, or sponge, by the molten metal. With these materials, effective infiltration may be expected due to the better wetting of the Nb by the Al than by the CaCl2.

In further embodiments, various methods may be used to encourage the full metal infiltration of the porous Nb-based rod. One of these is to pump out or suck out the molten CaC12 from the rod. This can be achieved by pumping the salt through the tubular Kanthal support shown in figures 8 and 9. In the case of the hollow support and rod, sucking will be very effective and any excess of the

molten metal within the core of the sample or in the Kanthal support can be easily removed after infiltration and replaced with internal cryogenically stabilising composite such as Cu with a protective Ta diffusion barrier. In a second method an ultrasonic device mechanically coupled to the rod or its support, or to the bath of infiltration material, may also be used to accelerate the infiltration process.

In-situ infiltration of the Nb-based porous rod should advantageously minimise many of the negative effects related to the ex-situ process mentioned above, and in particular the risk of surface contamination of the sample before infiltration.

Post-Infiltration Processing After the infiltration process, rods would be machined to the desired shape and inserted in subsequent tubes 62 to serve as a diffusion barrier and for electrical and thermal stabilisation. This is the cladding step 54 of figure 2. Although an elevated temperature during the infiltration stage may produce some A15 phase, it may be desired to subject the infiltrated rod or tape to a substantial reduction in thickness by cold rolling 56 prior to the final diffusion formation 58 of the intermetallic A15 layers in the conductor.

Microstructure Control An important aspect of superconductor fabrication is microstructure control. Use of the FFC process as a step

in fabrication enables an element of control through control of the particle size of the powder used to make the preform, densification of the preform through sintering, and the temperature and other electrolysis parameters at which the electrodecomposition is performed.

Advantages of the Method for Superconductor Fabrication Using the FFC process for reducing solid oxides to metals and alloys of pre-defined alloy composition and infiltrating these materials in our opinion opens new opportunities for the manufacture of the highest quality low temperature superconductors, not only the most saleable ones but also those best intermetallic ones which are difficult to manufacture eg A15 intermetallic conductors such as (Nb, X) 3 (Sn, Z) and Nb3 (Al, Ge) characterised by the highest Jc, BC2 ! Tc values may be produced for a fraction of the cost of the currently available superconductors. Examples of the measured performance of various superconductors embodying the invention are shown in figures 6 and 7. The ex-situ and in-situ infiltration processes can be applied to the niobium-titanium, niobium-tin and niobium-aluminium systems as described but in general these techniques can be used to infiltrate any metals or materials manufactured by the FFC process, whether for superconducting applications or other purposes.

Additional Example of Infiltration for Materials Fabrication The following specific example describes the aluminium infiltration of a porous, partially metallised, Nb pellet produced by the FFC process.

Preparation of the cell for in-situ infiltration About lOOg of aluminium shot was put into a cylindrical crucible made, in different tests, from either alumina or graphite (internal diameter and height were 50mm and 90mm respectively). The crucible was then filled to the rim with dry powders of CaCl2 and NaCl in the eutectic ratio, and placed in an Inconel tube reactor that was sealed, flushed with argon and heated in a furnace to 950°C. It was observed that both aluminium and the salt mixture had melted before the temperature reached 800°C, with the salt floating on top of the molten aluminium.

Thus, the cell comprised a layer of a molten eutectic mixture of CaCl2 and NaCl floating on a layer of molten Al.

Preparation of the Nb Pellet A cathode preform of Nb205 was prepared by pressing oxide powder into a small cylindrical pellet (10mm diameter, 10mm height, approximately 1.5g mass) which was then sintered at 1000°C for 2 hours. After sintering the preform gained a reasonable strength and had a porosity of about 40-50%, depending on parameters including starting material parameters and pressing pressure. A hole (1.5mm diameter) was drilled through the sintered pellet which

was then threaded onto Kanthal wire. This assembled cathode was placed in the molten eutectic mixture of CaCl2 and NaCl at 950°C. FFC electrodecomposition was carried out with a graphite rod anode at 3. 1V under argon for a relatively short time (between 1 and 10 hours) so that only the surface region of the pellet was reduced to metal while the central part remained in the oxide phase, i. e. the oxide pellet was partially metallised.

Infiltration of the Pellet After termination of the electrodecomposition, the temperature of the furnace was lowered to 690°C. The cathode was then immediately lowered into the molten aluminium underneath the molten salt. After a very brief infiltration/reaction time (a few seconds), the cathode was removed from the crucible, cooled first in the upper region of the Inconel reactor, and then removed from the reactor and further cooled in air. It was seen that the pellet was completely covered in aluminium.

The pellet was broken into two halves, and the cross section examined by SEM (scanning electron microscopy) and EDX (energy-dispersive x-ray analysis). It was observed that the pellet contained two different phases. The outer layer of the pellet was about 400 micrometres in thickness and relatively dense, but the central part was porous.

EDX analysis revealed that, as shown in Figure 10, the outer layer was composed mainly of niobium and aluminium with about 20at% oxygen, but the central part was of

niobium and calcium with 58at% oxygen (Nb20s contains 71at% oxygen). The calcium content was also much lower in the outer layer than in the central part. These results indicate that aluminium had infiltrated into the outer layer of the pellet, which had been metallised by electrodecomposition.