Load controlled rolling of superconducting tape The present invention relates to methods for obtaining an increased density of a material. More specifically, the invention relates to a method of critical current density of a superconducting material, especially of a Bi-based high temperature oxide superconducting ma- terial or precursor powder by load-controlled rolling, this increasing the density of the material. Such a method is especially useful when producing a Bi-based ox- ide superconducting tape in which a plurality of fila- ments of the oxide superconducter is surrounded by a metal matrix, the method providing a superconducting tape with improved critical current density. Presently pre- ferred superconducting materials for use in a method ac- cording to the present invention are materials belonging to the (BiPb)-Sr-Ca-Cu-0 group of materials.

Background of the invention A known manufacturing method for superconducting conduc- tors is the oxide-powder in tube method (referred to as Powder-In-Tube or PIT technology, in which an oxide pre- cursor powder is introduced into a metal tuoe, typically a silver or silver alloy tube, which by diameter reduc- tion is manufactured into a wire by means of mechanical deformation, such as drawing, swaging or rolling. The wire is subsequently cut into a plurality of sections and arranged in an outer metal tube, typically a silver or silver alloy tube, which is subjected to further mechani- cal deformation tc form a multifilamentary wire.

The multifilamentary wire is then rolled or pressed into a tape, which is subsequently intered to bring the oxide powder into the superconducting state and to increase the

critical current density of ~he tape. By tape is mean~ a length, which has a width greater than its thickness.

In order to increase the critical current density, * : he sintering process is repeated with intermediate mechani- cal deformation between the sintering steps to improve the texture and density of the superconducting material.

The intermediate mechanical deformation can be achieved by pressing or rolling. When the pressing or rolling de- formation is performed, cracking may be caused in the su- perconducting core. These cracks are to a certain degree healed in the subsequent heat treatment ; however, cracks not healed in the heat treatment will reduce he critical current density.

Therefore, there will be a trade-off between a higher fi- nal density and a more severe crack formation. Generally speaking, pressing tends to cause less cracking and a higher effective pressure can thus be used to increase the density of the oxide core while high pressure during rolling will decrease the critical current density due to crack formation (G. Grasso, Jeremie, and R.

Supercond. Sci. Technol., 8, 827, 1995).

However, the simple pressing method is not suitable to produce superconducting tapes in long lengths. One way to adapt the pressing method for production of long length tapes is the continuous periodic pressing in which a spe- cial profile die is used for pressing (F. Marti, G.

Grass, Y. B. Huang, M. Dalle, G. Witz, R. Passerini, E.

Giannini, E. Bellingeri, E. Walker, R. Flukiger, Progress in Improving Jc of Long (2223) Tapes by Periodic Pressing, Advances in Superconductivity XI/2, Koshizuka Tajima eds, Springer Tokyo, 939, 1995). But rolling is the preferred method for long length tapes due to its high productivity. Using rolls of a larger diameter can

alleviate the crack formation in the transverse direction of the tape and improve the critical current density (T.

Hikata and K. Sato, Patent 5, 246, 917, 1993). An eccen- tric rolling method-as also been developed 10 simulate the pressing condition (L. Kopera, P. Kovac, and I.

Husek, New rolling technique for texturing of-_ (2223)/ Ag tapes, Superconc. Sci. Technol., 11, 433,-998). In addition to the deformation method, the proper deforma- tion parameters are essential for high critical current density.

In the above-referrea disclosure by Grasso et a. it is further described a maximum critical current can be achieved by a rolling process in which a given optimum pressure is exerted on the tape during a number of defor- mation steps between the heat treatments.

Summary of the invention Accordingly, an objective of the present invention is to provide improved methods suitable for use in the produc- tion of an oxide superconducting length of tape, in order to obtain a high critical current density in a rolling process.

In a first aspect, a method for obtaining a-increased critical current density of a superconducting material is provided, the method comprising at least two intermediate rolling steps of subjecting the length of conductor to a rolling process in which a predetermined rolling load is set, each intermediate rolling step taking place between two heat treatment processes, wherein wherein least one inter- mediate rolling process takes place at a lower load than the previous intermediate rolling process.

This aspect of the invention is based on the realisation that an optimum rolling load should be determined for each intermediate rolling, this in contrast to for exam- ple Grasso et al. (se above) which discloses that an op- timum critical current is achieved by a number of defor- mation steps at the same pressure.

A problem with this known method is that the optimum pressure (or load) is determined after the tape has been finished, i. e. after a number of intermediate rolling steps and corresponding heat treatments. As heat treat- ment is a very slow process, which for each treatment may take from 2-6 weeks, it takes a very long time to deter- mine the optimum load to be used during the rolling steps.

Thus, in a further aspect of the invention, a method is provided for determining an optimum load for a single intermediate rolling, where a lower load will not give the necessary densification and a higher load will cause excess cracking.

More specifically, a method is provided comprising the steps of determining a first load range for which a mo- notonous, substantially linear increase in reduction ra- tio takes place with increasing load, and a second load range for which the monotonous, substantially linear in- crease in reduction ratio no longer takes place with in- creasing load, determining the transition point repre- sented by the first data point in the second load range which deviates from the monotonous, substantially linear increase in reduction ratio in the first load range, and determining an optimum load corresponding to the transi- tion point. Preferably the method comprises a further step of determining an optimum load range around the op- timum load.

Based on the above, an optimum load for an intermediate rolling can be determined and the production of the"in- termediate"tapes can commence without awaiting the test results for the finished superconducting tape. This is in contrast to the method disclosed by Grasso et al. in which only the finally achieved critical current is meas- ured, as measuring the"intermediate"critical current cannot be used for determining a single optimum inerme- diate rolling load, i. e. a tape with a high critical cur- rent after a first intermediate rolling does not neces- sary lead to a finished tape having the highest possible critical current.

Further, load has also shown to be a more suitable pa- rameter for process control than the traditionally used reduction control since the change of thickness, i. e. the reduction, is very small compared to the thickness of the tape. However, the change of load is large which allows for easy process setting.

In yet another aspect, the optimum load ranges for a first rolling and a subsequent second rolling are pro- vided for a superconducting lape. Especially, the pres- sure for the first rolling should be the highest with a reduced pressure for the subsequent intermediate rolling. In a preferred embodiment, the pressure for the first in- termediate rolling should be around 1, 1-1, 4 G ? and the pressure for the second intermediate rolling should be around 0. 5-0. 7 GP In the present application the term"superconducting"is used to denote both the precursor powder material and the subsequent actually superconducting material.

The invention will now be described in further detail with reference to the accompanying drawings, wherein : Figs. la-lh is a schematic illustration of the Powder- In-Tube (PIT) process, Fig. 2 shows a graphic representation of a firs reduc- tion ratio-to-rolling load relation for a given tape, Fig. 3 shows a graphic representation of a second reduc- tion ratio-to-rolling load relation for a given tape, Fig. 4 shows a first graphic representation of a critical current-to-rolling load relation for a given tape, and Fig. 5 shows a second graph representation of a criti- cal current-to-rolling load relation for a given _ape.

First the principle steps for process of producing a length of a superconducting length of tape will be de- scribed with reference to la through lh.

First a silver or silver alloy tube is filled wicn a su- perconducting precursor powder, typically a metal oxide (fig. la), and then the tube is drawn into a single fila- ment wire (fig. lb), which subsequently is cut into sec- tions and packed into a multifilamentary wire (ri g. lc).

The multifilamentary wire is then drawn to a small- diameter wire (fig. ld), which is subsequently rolled into a thin tape, in which each of the single powder con- taining filaments is transformed into a flattened struc- ture resembling the overall appearance of the rolled tape. Thereafter the tape is heat treated in a furnace to obtain the superconducting properties of the precursor and to heal the cracks caused by the rolling process (fig. lf). After the first heat treatment the tape is

subjected to an intermediate rolling (fig. _g) after which it is subject : a further heat treatmene fig. lh) to heal cracks former in the deformation process during the intermediate rolling. The last two steps correspond- ing to figs. lg and _^. may be repeated several __mes.

Next the problem of obtaining a high densificat-on of the superconducting material yet avoiding the formation of excess cracking will be discussed with reference to fig.

2.

Fig. 2 shows the relation between the reduction ratio and the load (measured as the roll separating force during a first intermediate rolling. As can be seen frc-. the dia- gram, with an increase in load, there is a range of mo- notonous increase in-reduction ratio. Further,-n a rela- tively well-defined range, the increase is substantially linear, this being vr. e"linear"range. However, as is clear from the figure, the monotonous increase in reduc- tion ratio neither takes place at very low load levels, nor after a load of a certain height. As can also be seen in the figure, from loads of a given value the relation between the reduction ratio and the load become-very ir- regular. Experience has shown that within this range of load, excess cracking of the superconducting material takes place, however, corresponding to an increased re- duction ratio in the range of monotonous increase in re- duction ratio, increased densification takes place. The problem is therefore, to roll the length of tape corre- sponding to the maximum within the monotonously increas- ing range. In the past this has traditionally been achieved by determining the maximum reduction ratio and then to control the rolling process accordingly, i. e. if the optimal reduction rate had been found to be 8. 5--then this was the value according to which the rolling process was controlled.

However, as typical tape thickness is 0. 2 mm with _ variation of up to 0. 01 mm (or 5t), it is very difficu_ to control the optimum rolling condition with the reduc- tion ratio as a controlling parameter. This is also i-- lustrated by the figure, for example if the actual roll- ing process takes place at a reduction ratio of 9. 5f, as :, as compared to the set value of 8. 5t, then rolling takes _--. place far within the load range for which severe cracking may take place.

In contrast, load is a more suitable parameter for proc- ess control since the change of load is large as comparer with the change in reduction ratio, this both in absolute and relative ~ers, this allowing for easy process con- trol.

In the following a method is described by which the pre- sent invention can be turned into practise. For a given specimen, i. e. a tape to be rolled, tests are performed by which the reduction ratio (in) is plotted against the load (in : con) for a number of different load values, a data point representing a reduction ratio for a given load. The object of these tests are to determine a number of data points for the load range for which monotonous, substantially linear increase in reduction ratio takes place, as well to determine a number of data points for the load range for which the relationship between reduc- tion ratio and load becomes irregular or the reduction ration no longer increases with increased load. The num- ber of data points needed to achieve a useful graphic representation, i. e. a curve, of the reduction ratio ver- sus load relation may vary, however, as such a curve-- typical when materials are tested in deformation tests, it would be a routine experiment for the skilled person to achieve a"typical"curve as shown in fig. 2.

From the graphic representation a range can be approxi- mately outlined by two parallel lines corresponding to a range of monotonous, linear increase in re- duction ratio against load, the range enclosing the test values within the linear range. Again, the typical test values will not represent a perfect linear relationship, but in practise the skilled person will readily identify those values on the basis of which the range should be determined, i. e. obviously erroneous data points should be neglected. The load for the first data point that de- viates from the"linear"behaviour, i. e. falls outside the range between the drawn lines, is defined as the transition point. The optimum rolling load is, according to the invention, around the transition point as sche- matically shown by the rectangular range in fig. 2.

In the shown embodiment a range has been outlined by two parallel lines with a separation of about 1% in reduction ratio. The optimum load determined from fig. 2 would be approximately 3. 0 tons. The range for the optimum rolling load around the transition point is here within-30 to +30@ of the transition point load value, and preferably in the-10--to +10^-range cf the transition point load value. Although the T_ is lower for lower load, the scat- tering is less and this can be an advantage for some applications. In the present context the term"around" may also imply a range which just includes the optimum load, e. g. from 0 to +10, C.

Indeed, the linear range could also be defined using a straight curve fitted corresponding to the data points in the linear range. In the same way a (non-linear) curve could be fitted for the non-linear range. The transition point could then be determined using, for example, the standard deviation value calculated for the linear range,

for example as the point where the non-linear curve devi- ates from the calculated standard deviation value. n ~he past, when more than one intermediate rolling were used in the manufacturing of a superconducting tare,-he first and the subsequent rolling procedures were per- formed with the same set value for the reduction ratio (or load as in Grasso et al. However, the present inven- tors have found that the relation between the reduction ratio and the load (measured as the roll separating force) during a second (or subsequent) intermediate roll- ing is very different from ehe proceeding rolling To illustrate this, reference is made to fig. 3 depicting the relation between the reduction ratio and the load during the second intermediate rolling, i. e. after treat- ment in the furnace, of the same specimen used for deter- mining the relation shown _n fig. 2. In this case, the first data point outside the"linear"range is far away from the parallel lines, however, it is readily recog- nised by the skilled person that more data points are needed in this range to determine a more precise load value for the transition point. The optimum loao deter- mined from fig. 3 would be approximately 1. 25 tons. What is essential is the recognition that the optimum load for the second intermediate rolling is much smaller than the corresponding value during ehe first intermediate roll- ing.

Example 1 Reference is now made to fig. 4. A superconducting pre- cursor powder was filled into a silver tube of 20 mm in outer diameter and 17 mm in inner diame- ter, which was drawn into a wire of 2. 2 mm in outer di- ameter. Then the tube was cut into 37 sections of equal

length and the sections were filled into a silver alloy tube of 20 mm in ^_~^r diameter and 17 mm in inner diame- ter, which in turn was drawn into 1. 03 mm in outer diame- ter and then rolled into a tape of 0. 22 mm in thickness.

Thereafter, the tape was heat treated at 800 to 840°C and then gradually cooled. Then the tape was subjected to a first intermediate rolling with an optimised load of 3 tons after which the tape was subjected to a second heat treatment within the same temperature range. After the second heat treatment the tape was cut into a number of lengths which were ~~en subjected to a second intermedi- ate rolling in which rhe load for different lengths were gradually changed from 0 to 3 tons ; more specifically, the length were rolled at loads of 0, 0. 5, 1. 0, 2. 0 and 3. 0 tons. Thereafter, the samples were subjected to a third heat treatment. The critical currents (in Amperes) for the samples ã~ ~ 7 K were measured and are shown in fig. 4. The trend suggests that a sample treated in the optimised load range will give the highest critical cur- rents and less scattering, i. e. when treated at a load value in the range from 1. 0 to 1. 5 ton as illustrated in fig. 4 and as also found in fig. 3.

Example 2 Reference is now made to fig. 5. A tape was prepared in a similar way as in Example 1. The following three differ- ent loads were used for the first and second intermediate rolling.

(A) 1. 0 ton (B) 3. 0 tons (C) 5. 0 tons

Therefore we have nine combinations.--, AB, AC, BA, bol, BC, CA, CB, C, where the first and second letters repre- sent the loads for the first and second intermediate rolling respectively. Again the optimum condition B gave the highest critical current,-. e. a load for the first intermediate rolling of 3. 0 tons as illustrated in fig. 2, and a load for the second intermediate rolling c 1. 0 ton as illustrated in fig. 3. The average rolling pressure can be calculated by where P is the load, R'is the roll radius after adjust- ment for the elastic flattening, #h is the absolute re- duction in thickness, and b the average width of the tape. Corresponding to the load ranges illustrated in figs. 2 and 3, the corresponding pressure for the firs' : intermediate rolling should be around 1. 1-1. 4 GPa, and the pressure for the second intermediate rolling should be around 0. 5-0. 7 GPa.

When the optimised load range has been determined for a specific sample under specific condition, the correspond- ing pressure range can be calculated. However, it has also been found that this calculated pressure can be used to calculate a new optimised load range, for example when the width of the tape or the diameter of the rollers are changed, this saving a new determination of the optimised load range.