WO2025181781A1 - Method for preparing an austenitic fe-ni alloy strip, strip prepared thereby, and products produced from said strip - Google Patents

Abstract

The invention relates to a method for preparing an austenitic iron-nickel alloy strip, which method comprises: - preparing, by melting, casting and hot-forming, or by powder metallurgy, a semi-finished product of an alloy, the composition of which consists of the following, expressed as weight percentages: 30% ≤ Ni ≤ 60%; traces ≤ Co ≤ 10%; traces ≤ Mn ≤ 3%; traces ≤ Cu ≤ 10%; traces ≤ Mo + W + Cr + V ≤ 10%; traces ≤ Si ≤ 4%; traces ≤ C ≤ 500 ppm; traces ≤ Zr + Hf ≤ 500 ppm; 20 ppm ≤ S + Se + Te ≤ 60 ppm; 0.01% ≤ Nb + Ta + Hf + Al + Ti + B ≤ 0.5%; traces ≤ Al ≤ 0.02%; traces ≤ Ti ≤ 0.06%; traces ≤ B ≤ 0.06%; the remainder being Fe and impurities; - hot-rolling the semi-finished product at a temperature of 1000 to 1300°C until a strip metal with a thickness of 4 to 10 mm is obtained; - cold-rolling the strip metal to a final thickness with an overall reduction rate of at least 90%, in order to obtain a strip; - applying a final recrystallisation annealing, preferably static annealing, to the strip at a temperature of between 1000°C and the temperature at which abnormal grain growth starts minus 20°C, for 30 to 600 minutes, said final recrystallisation annealing taking place in a reducing atmosphere. The invention further relates to a strip made of an austenitic iron-nickel alloy prepared by the above method, and to the uses of said strip.

Description

Process for preparing an austenitic Fe-Ni alloy strip, strip thus prepared and products made from this strip

The present invention relates to textured and even “hypertextured” iron-nickel alloys (see below for the definition of these terms) cold rolled then optimally annealed, and more precisely Fe-Ni alloys with a {100}<001> texture, more commonly called “Cube texture”.

Industrial materials are generally polycrystalline, that is, made up of multiple crystals or grains. A material is said to be textured, or to have a texture, if the distribution of the crystallographic orientations of its grains is not random.

We recall that to define and represent a texture, we use two reference frames in a cold-rolled material: one reference frame linked to the sample (material) and another linked to the crystal (grain). By convention, we denote here (u,v,w) the vector of coordinates of value u, v and w; we denote according to the Miller indices (uvw) the plane perpendicular to the axis carried by the vector (u,v,w); we denote according to the Miller indices [uvw] the straight line carried by the vector (u,v,w).

The reference frame linked to the sample is defined by three mutually orthogonal directions which are: the rolling direction DL, the direction DT transverse to DL in the rolling plane, and the direction DN normal to the rolling plane.

The reference frame linked to the crystal with a cubic crystalline structure - that is to say having an elementary crystal mesh such as a perfect cube - is represented by the three edges of the elementary mesh (cube) and which are noted according to the Miller indices as the lines [100], [010] and [001], according to the same principle as the coordinates of the respective vectors (1,0,0), (0,1,0) and (0,0,1). In this notation, each of these three axes can represent one or the other of the 3 edges of the cube. The naming of Crystallographic Orientations (CO) and texture components follows the so-called “Miller indices” rule in the crystal frame: (hkl)[uvw] designates the crystallographic orientation of a grain or crystal such that the plane perpendicular to the vector (h,k,l) - also noted plane (hkl) - is parallel to the rolling plane, and such that the vector (u,v,w) - also noted straight line [uvw] - is parallel to the DL direction.

We denote by {hkl} the set of crystallographic planes equivalent to the plane (hkl), that is to say containing the same indices h, k and I with any places in the triplet and a positive or negative value: for example the family {hkl} contains in particular the planes (hkl), (-hkl) and (khi). We denote by <uvw> the set of crystallographic lines equivalent to the line [uvw], that is to say containing the same indices u, v and w with any places in the triplet and a positive or negative value: for example the family <uvw> contains in particular the lines [uvw], [-uvw] and [uwv]

In the case of Cube orientation {100}<001 >, one of the planes of the family {100} is parallel to the rolling plane (DL,DT) and one of the lines of the family <001 > is parallel to the rolling direction DL: for example the plane (010) of the crystal is parallel to the plane (DL, DT) and the direction [100] of the crystal is parallel to DL.

The non-random distribution of crystallographic orientations (which is reflected in the classical notion of “Crystallographic Orientation Distribution Function” or FDOC) is therefore by definition a texture: this texture can just as well be centered around a single particular Crystallographic Orientation (CO) (hidkidlid)[UidVidWid], thus forming the “texture component {hokolo}<uovowo> consisting of all the grains having a large or fairly large angular proximity with (hidkidlid)[UidVidWid], typically less than 15 or 20° from this reference or “ideal” CO: the inventors have chosen here a maximum misorientation limit of 10.0° between the ideal orientation (hidkidlid)[UidVidWid] and all the orientations (hokolo)[uovowo] of the grains composing the texture (component). The remaining grains not included in this texture component are therefore, with a good approximation, randomly distributed in terms of FDOC. But the texture can just as easily be multi-component with a residue of approximately randomly distributed grains. This is, for example, the case below of the texture comprising the components S, C and B.

A crystallographic orientation well known to metallurgists is the (010)[100] orientation (moreover crystallographically equivalent to (100)[001] or (001)[010] for example) for the numerous phenomena it allows (surface epitaxy, precision of chemical cutting, optimal magnetic performance of most ferromagnetic polycrystallines, rectangular hysteresis cycles of magnetic materials textured with this orientation, etc.). This (010)[100] orientation is commonly called the “Cube orientation”. The metallurgical production processes, hot and cold transformations alternating or ending with annealing, allow, after studies and optimization, at best, to obtain a significant proportion of grains having a crystallographic orientation more or less close to the Cube orientation, and we then call all of these oriented grains the “Cube texture component”.

In this case, the Cube (010)[100] orientation is the ideal orientation towards which the metallurgist tries to orient the greatest number of grains, and with the smallest possible disorientation of the grains (oriented (hokolo)[uovowo]) with respect to the ideal orientation (010)[100], This Cube orientation case therefore corresponds to hid =0, kid =1, d =0, uid =1, Vjd =0, wid =0.

The quantification of a texture includes on the one hand the surface fraction of grains whose disorientation with respect to the ideal orientation (010)[100] is less than a given critical angle, which the inventors have chosen to take equal to 10.0°, and on the other hand the average disorientation of these grains with respect to the ideal orientation (010)[100]. These parameters represent the acuity of the texture component.

To clarify the terms, "Cube (texture) component" means the set of grain crystallographic orientations (hokolo)[uovowo] which are obtained by rotation of at most 10.0° between the [uovowo] axis of each grain orientation (the closest to DL rolling direction) and the [100] axis parallel to the DL direction of the Cube orientation (010)[100]. A person skilled in the art calls this set of orientations described above "Cube texture" or "Cube texture component" indifferently because in the case of the invention the examples and counterexamples have at least 95.0% (by volume or material surface) of the grains whose crystallographic orientation (hokolo)[uovowo] is disoriented to less than 10.0° from (010)[100] (so-called ideal Cube orientation), or even at least 99.0%, and therefore less than 5.0% (on the surface of the metal) or even less than 1.0% of the metal corresponds to other very minor crystallographic orientations; thus, as a good approximation, we confuse the Cube component of the material texture and the Cube texture of the material.

In the following text, we will call "hyper-texture" a pronounced, single-component texture, whose surface or volume fraction of grains is at least 96.0%, and the average disorientation of the grains of the texture component with respect to the ideal orientation of less than 7.0°.

We can be even more precise:

- in the case of thin laminates in this invention, due to the high work hardening rates, the final material thicknesses are generally less than 0.2 mm and the final recrystallization into a highly textured Cube material is a volume process. Thus, the surface or volume characterization of the crystallographic orientations is equivalent.

- the disorientation coi between each grain i, oriented (hoikoiloi)[uoiVoiWoi], and the Cube orientation (010)[100] - constituting the orientation ideal to be obtained - is given by definition by the scalar product of the two vectors [uoiVoiWoi] and [100], which is worth uoi and also which defines a>j by cos(û)j) = u _0i / /u _0i ² + v ₀ j ² + w _0i ² , a>j being positive. The average misorientation co of grains misoriented within 10.0° of the Cube orientation is then given by a simple weighted arithmetic mean of the grain surfaces coi < 10.0° and counting a total number N of grains i with N>100 to have good statistics, preferably N>200.

A "twin orientation" is a crystallographic orientation deduced by a simple rotation, with a very precise rotation angle, from the orientation of the texture component considered. For example, for the Cube component {100}<001> the twin orientation is {122}<221 >.

The development of the Cube texture in materials with a face-centered cubic crystal structure with high and medium stacking fault energy, such as FeNi alloys, follows a number of steps, as described, for example, in US-B-6635,097, EP-A-1,828,425. After obtaining an ingot by a traditional metallurgical route or by powder metallurgy, the material is hot transformed by forging and/or rolling. The next step, essential for developing the texture, is to cold roll the hot-rolled material. The strain rate must be as high as possible (>70%) to develop a certain rolling texture composed of three main texture components, which are the ideal orientations S {123}<634>, C {112}<111> and B {110}<112>. The rolling texture composed of these three orientations is essential for the development of a pronounced {100}<001> Cube texture, after recrystallization annealing at high temperatures. This annealing is therefore the last step in the Cube texture development process.

In US-B-6635 097, it is shown that Ni-X alloys (with X = Cu, V, Al, W, Cr, Mo, ...) develop a Cube texture by following the following metallurgical range: formation of ingots by powder metallurgy, recrystallization heat treatment in order to develop grains of sizes less than or equal to 50 pm, cold rolling with a total reduction rate greater than 90%, the reduction rate per pass being less than 10%. The Cube texture quantified indirectly from the diffraction peaks obtained by X-rays has widths at half-maximums between 6° and 9°. FeNi alloys are, however, not concerned by this document.

In patent EP-A-1 828425 concerning Ni alloys not containing Fe, two ingot formation procedures are used: traditional metallurgy (from liquid metal) and powder metallurgy. The ingots undergo homogenization heat treatments and hot rolling at temperatures above 850°C. Cold rolling with a reduction ratio greater than 70% is then applied to obtain strips 80 μm thick. Samples annealed at 1150°C under a protective atmosphere develop a Cube texture greater than 90%. The technique used for texture quantification was not described in the patent.

Hypertextured materials can be used as epitaxy substrates when there is compatibility between the lattice parameters of the two materials involved: the substrate and the epitaxially grown material. Some studies have thus been carried out on the possibility of depositing silicon on Cube textured FeNi substrates (document US-B 9 309 593) and on the deposition of superconducting materials on Ni5%W substrates in the manufacture of superconducting cables by the RABITS technique (Rolling Assisted Blaxially Textured Substrate Technique), see document US-B-6 156 376).

Also, to reduce the amount of bulk silicon used in the photovoltaic industry while increasing the efficiency of the panels, several laboratories are working on thin-film photovoltaic panels. This technology requires a substrate on which a thin layer of Si is deposited. The substrates currently used for this application are glass and ceramic substrates. However, photovoltaic cells developed using this technique still have very low efficiencies, due to substrate instability problems, problems with the interaction of silicon with the substrate, the low melting temperature of the substrates, and especially insufficient epitaxial growth of Si on conventional supports.

To overcome these problems, one solution is to develop an alternative metal substrate to glass and ceramic substrates. The metal substrate must be thin, non-brittle, flexible, and must have a high melting temperature and structural characteristics favorable to oriented or epitaxial growth of silicon thin films. Some authors have shown that FeNi alloy substrates with a preferentially Cube {100}<001> texture are good candidates for this application (document US-B-9,309,592).

In another field, the materials that are used in current transport are mainly Cu and Al. However, these materials cause current losses by Joule effect along the electrical circuit, because of their electrical resistance. The use of superconducting materials for current transport would eliminate the losses caused by the resistance of the conductors and optimize current electrical networks. However, it is difficult, if not impossible, to manufacture electrical cables only from superconducting materials, because of their mechanical properties, as they are hard and fragile. To circumvent this problem, several laboratories and industrialists are working on RABITS technology which consists of depositing by epitaxy a superconducting layer (YBa2Cu3O?-5 noted YBaCuO) on a metallic substrate of Ni5%at.W textured Cube (document US-A- 6 156 376). The use of FeNi is not known for this purpose.

The Cube {100}<001> texture also has the advantage of having a <100> direction of easy magnetization parallel to the DL and DT directions, in a FeNi material that has undergone rolling deformation. Due to this preferential orientation, the magnetic properties are significantly superior in both the rolling direction and the transverse direction. Alloys that develop this type of texture are therefore very useful for applications in which magnetic fluxes must be controlled in one or two directions. The magnetic characterization of FeNi alloys with an intense Cube texture shows that they have a rectangular hysteresis cycle. As an example, we can cite the Cube textured Fe50%Ni magnetic alloys in low-noise current transformers where the “E” and “I” that make up the transformer by stacking are cut according to the DL and transverse DT rolling directions of easy magnetization (document PCT/IB2016/001409 in the name of the applicant).

In recent years, studies have been carried out with the aim of developing processes to obtain a Cube texture in face-centered cubic (FCC) materials with high and medium stacking fault energy. However, most of these studies are based on Ni5%at.W alloys used as substrates in the manufacture of superconducting cables, with the RABITS technique. However, some studies exist on FeNi alloys which are the subject of this patent.

It has been shown that after strong deformation, these alloys develop a texture composed of three main components B, S and C. The three corresponding ideal orientations all have an energy greater than the stored energy of the Cube orientation which is found in very small quantity in the deformed state (approximately 2%). During annealing, it is the Cube component which develops to the detriment of the original B, S and C components.

The strain rate plays a very important role in the development of the Cube texture. The higher it is, the better the Cube texture in intensity and misorientation after optimized recrystallization anneals. Other parameters such as annealing temperature, annealing atmosphere and alloying elements also have a significant impact on the development of an intense Cube texture.

To develop a Cube hypertexture, it is important to perform recrystallization annealing at high temperatures, as mentioned earlier. However, a harmful phenomenon called abnormal growth (AG), or recrystallization secondary, can be triggered if the temperature rises very high during annealing. The Cube texture already formed is degraded, or even eliminated, by the appearance of large grains with an orientation different from the Cube orientation in the microstructure.

In the literature, the role of alloying elements on the development of Cube texture is not well known. Some studies related to this subject are limited either by the difficulty in studying the individual and/or combined effect of the addition elements, or by an incomplete study, or finally by the unsuitable means of characterization.

It would be important to find ways to promote the development of a very strong Cube texture {100}<001> in FeNi alloys. This Cube hypertexture would be essential to solve epitaxy problems in the development of thin-film photovoltaic cells and the manufacture of superconducting cables, or to develop easy magnetization directions <100> essential for certain magnetic applications.

The aim of the invention is to obtain austenitic Fe-Ni type strips, which may contain other alloying elements, the microstructure of which is characterized by a Cube hypertexture, that is to say, in the intended application cases, by a sharp mono-component texture, in other words representing at least 99.0%, better still at least 99.5%, on the surface, of the grains of the material.

This texture is characterized by a tight distribution of grains with {hokolo}<uovowo> orientations having at most 10.0° of misorientation relative to the ideal Cube (010)[100] orientation (in which each of the <100> directions is parallel to one of the DL, DT, or DN directions). Relative to the ideal CUBE (010)[100] orientation, the grains have at most 10.0° of misorientation present, and furthermore collectively have an average misorientation less than or equal to 4.0°.

Summary of the invention

To this end, the invention relates to a process for preparing an austenitic iron-nickel alloy strip, characterized in that:

- a semi-finished product of an alloy is prepared by melting, casting and hot forming, or by powder metallurgy, the composition of which consists of, in weight percentages:

- 30% < Ni < 60%;

- traces < Co < 10%;

- traces < Mn < 3%;

- traces < Cu <10%; - traces < Mo + W + Cr + V <10%;

- traces < If < 4%;

- traces < C < 500 ppm, with, preferably 30 ppm < C, better 50 ppm < C, and, preferably, C < 200 ppm, better C < 150 ppm;

- traces < Zr + Hf < 500 ppm, preferably traces < Zr + Hf < 100 ppm;

- 20 ppm < S + Se + Te < 60 ppm;

- 0.01% < Nb + Ta + Hf + Al + Ti + B < 0.5%;

- traces < Al < 0.02%;

- traces < Ti < 0.06%;

- traces < B < 0.06%; the remainder being Fe and impurities resulting from the production;

- said semi-finished product is hot rolled at a temperature of 1000 to 1300°C, preferably 1100 to 1250°C, until a strip with a thickness (6LAC) of 4 to 10 mm, preferably 4.5 to 8 mm, is obtained;

- said strip is cold rolled, in one or more stages, to a final thickness (e _f ) with an overall reduction rate (TR) of at least 90%, preferably at least 95%, to obtain a strip; and

- a final recrystallization annealing of the strip is carried out, preferably a static annealing, at a temperature between, on the one hand, 1000°C, preferably 1020°C, better still 1030°C, and, on the other hand, a maximum normal growth temperature (TM), considered to be equal to the abnormal growth start temperature (TDCA) reduced by 20°C, for 30 to 600 min, said final recrystallization annealing taking place in a reducing atmosphere.

Cold rolling can be carried out in several stages, and at least one of these stages is carried out with a reduction rate of not more than 20%, preferably not more than 15%.

The recrystallization annealing of the strip can be carried out from room temperature with a temperature rise rate, up to the annealing temperature, of between 0.1 and 10°C/min.

After static recrystallization annealing of the strip, the rate of temperature reduction of the strip to room temperature may be greater than the rate of temperature rise of the strip during static recrystallization annealing.

Rolling and heat treatment parameters can be adjusted so that the average grain size is < 50 pm at a time between the end of hot rolling and a time when the strip thickness is at least 1 mm. An intermediate recrystallization annealing at a temperature between 600 and 1000°C, for 30 s to 10 hours, may be interposed between two of the cold rolling steps, and the cold rolling steps following said recrystallization annealing have a cumulative reduction rate of at least 90%, preferably at least 95%.

It is possible to insert, between at least two of the cold rolling stages, one or more intermediate restoration annealing operations carried out between 500 and 700°C, lasting 30 s to 24 hours, while avoiding recrystallization of the material, said intermediate restoration annealing operation(s) taking place after said possible intermediate recrystallization annealing operation.

An additional annealing can be carried out under argon, or nitrogen, or helium, or hydrogen + argon, or hydrogen + nitrogen, after the final texturing annealing under a reducing atmosphere and at a temperature 20 to 200°C higher, preferably 50 to 200°C, than the recrystallization annealing temperature, for between 30 min and 4 h.

Mechanical polishing and/or chemical pickling of the strip surface may be carried out at one or more intermediate stages between the end of hot rolling and obtaining the final thickness of the strip, preferably before the start of cold rolling or before any intermediate recrystallisation annealing.

Cutting of the strip may be carried out before said final recrystallization annealing, preferably by non-mechanical means.

The invention also relates to an austenitic iron-nickel alloy strip, characterized in that its composition consists of, in weight percentages:

- 30% < Ni < 60%;

- traces < Co < 10%;

- traces < Mn < 3%;

- traces < Or < 10%;

- traces < Mo + W + Gold + V < 10%;

- traces < If < 4%;

- traces < C < 500 ppm, with, preferably 30 ppm < C, better 50 ppm < C, and, preferably, C < 200 ppm, better C < 150 ppm;

- traces < Zr + Hf < 500 ppm, preferably traces < Zr + Hf < 100 ppm;

- 20 ppm < S + Se + Te < 60 ppm;

- 0.01% < Nb + Ta + Hf + Al + Ti + B < 0.5%;

- traces < Al < 0.02%;

- traces < Ti < 0.06%;

- traces < B <0.06%; the remainder being Fe and impurities resulting from the production; in that its fraction of Cube {100}<001> orientation grains is at least 99.0%, preferably at least 99.5%, the Cube {100}<001> orientation grains having a misorientation relative to the ideal Cube (100)[001] orientation of at most 10.0°, the direction considered for determining the misorientation being, for each of the grains, that among the <100> directions closest to the rolling direction (DL), in that the average misorientation (œ) of the Cube {100}<001> orientation grains, relative to the ideal Cube (100)[001] orientation, is less than or equal to 4.0°, and in that the average equivalent diameter of all the grains of the strip is between 40 and 200 pm.

The invention also relates to a substrate for a photovoltaic cell, characterized in that it was obtained from a strip of the previous type.

The invention also relates to a substrate for a superconducting cable, characterized in that it was obtained from a strip of the previous type.

The invention also relates to an electric current transformer core element, characterized in that it was obtained from a strip of the previous type.

As will be understood, the invention consists first of all in optimizing the composition of the austenitic FeNi alloy used, by fixing contents of certain well-defined alloying elements. In particular, the contents of Ni, S and certain other elements of comparable effects, and of Nb and certain other elements of comparable effects, must be controlled and maintained within precise limits.

It also consists, once this particular alloy is prepared, in treating it in a precise way taking into account, during a recrystallization heat treatment, the TDCA temperature at which, for the alloy considered, the abnormal growth of the crystals begins, and destroys the Cube texture. This temperature must possibly be determined by experience, for example in conditions that will be detailed, and the composition of the alloy is adjusted so that TDCA is as high as possible, so that maximum growth of the Cube texture can be obtained in a short time.

Also sought is a sufficient average grain size to reduce the average crystal disorientation, a surface not oxidized by the final heat treatment (e.g. for epitaxial application) and a final annealing temperature not exceeding 1100°C to avoid surface hollowing by thermal boundaries. These alloys have good magnetic properties (in particular, they have a low Hc coercive field).

Finally, the aim is to produce this hypertextured material without it suffering from brittleness or damage when cold or hot during its hot or cold transformation to the final thickness. Brief description of the figures

The invention will be better understood in light of the following description, given solely as a non-limiting example and with reference to the following appended figures:

Figure 1 shows the typical evolution of the grain size of an alloy of given composition, as a function of the temperature during a recrystallization heat treatment;

Figure 2 shows how the abnormal growth start temperatures TDCA of the grains are measured, and the maximum growth temperature TM acceptable for carrying out recrystallization annealing according to the invention.

Detailed description

As mentioned, the temperature at which the abnormal growth process (A.C.) begins will be called TDCA. The TDCA is therefore slightly higher than the highest normal growth temperature TM, normal growth being that which allows the Cube texture component and only this to develop in a given alloy. The microstructure transformation phenomena can be summarized as in Figure 1.

Figure 1 schematically represents, for an alloy of composition Fe- 50%Ni by weight, the way in which the microstructure, in particular the grain size (which is, in the case considered, initially less than 1 pm), is transformed during a recrystallization annealing, as a function of the increase in temperature. Between approximately 600 and 700°C, we are in the region where primary recrystallization takes place, the grain size increasing to approximately 5 pm. Between approximately 700 and 1000°C, normal grain growth takes place, during which the grain size increases to approximately twenty pm and the Cube texture becomes more pronounced (NB: this recrystallization/normal growth transition around 700°C would also be observed in a similar manner for all the other compositions concerned by the invention). Then, beyond about 1000°C (i.e. beyond the TDCA of the alloy considered, which depends not only on the composition of the alloy, but also on the cold deformation rate TR undergone during cold rolling) abnormal growth takes place, during which the grain size increases sharply and the Cube texture is progressively destroyed. It is also known that the Cube texture develops more easily in austenitic Fe-Ni alloys when the overall cold deformation rate TR is high. It must typically be at least 90%, better at least 95%, between the final hot rolling thickness 6LAC and the final cold rolling thickness ef.

Preferably, no recrystallization annealing should occur between hot rolling and the last cold rolling step, otherwise there is a significant risk of not obtaining the reduction rate (work hardening) sufficient to allow the development of the very pronounced Cube texture. But at least one intermediate annealing without recrystallization, carried out between two cold rolling steps, and/or between hot rolling and the first cold rolling, can be favorable to the development of the Cube texture.

Carrying out an intermediate annealing with recrystallization is, however, not completely to be excluded, provided that it is part of a process which provides, after this intermediate annealing with recrystallization, a very high cumulative reduction rate of the following cold rolling stages, of at least 90%, preferably at least 95%. This therefore requires, in general, that this intermediate annealing with recrystallization be carried out at a very early stage of the cold rolling, that is to say when the product is still at a thickness close to that which it had at the end of the hot rolling.

One or more restoration anneals inserted between two cold rolling stages is/are, on the other hand, favorable to the intensification of the Cube texture, especially when elements lowering the stacking fault energy, such as Mo, are present in the steel. What is necessary is that, if an intermediate recrystallization anneal is also carried out, this takes place before the first intermediate restoration anneal, otherwise the restoration anneal is useless.

As is conventional, the formula for calculating the overall reduction ratio TR in cold rolling, regardless of the number of cold rolling stages used to achieve it, is:

Overall TR (in %) = 100.(ei_Ac- ef)/ei_Ac. where ei_Ac and ef are respectively the thicknesses at the end of hot rolling and at the end of cold rolling.

In the usual industrial hot transformation processes for flat materials, 6LACs of between 2 and 8 mm, preferably 2 to 5 mm, are commonly obtained. The application of an overall reduction rate TR of 90% allows the emergence, after a final annealing at around 950-1000°C, of a very predominant Cube texture for a final thickness ef of 0.2 to 0.8 mm. Thus, we can see on pole figures that in the recrystallized state, from 77% reduction rate, the poles of the Cube texture {100}<001> become clearly visible. For a reduction of 77% there are other poles marked, that is to say other texture components, so the Cube texture component is far from being ultra-majority. At 95% reduction, the Cube component is this time very majority, but there still remains a small minority twin component and it is clear that we are still far from 99% on the surface of Cube grains.

With 90-95% overall TR reduction rate it is well known that the metal, although very predominantly in Cube texture, on the one hand contains at least a few % (surface or volume) of crystallographic orientations very different from the Cube texture, such as the twin orientation, on the other hand presents a Cube texture component of average disorientation well above 4°, typically of the order of 10°.

It is recalled that the person skilled in the art calls "disorientation" coi of a crystallographic orientation (O.C.) (hi ki )[ui vi wi] belonging to a texture component {hkl}<uvw>, of ideal orientation (hokolo)[uovowo], the angular difference between the directions [U1V1W1] and [uovowo] (vectors whose coordinates refer to the crystallographic reference frame of each O.C., and which by definition are all parallel to DL in the reference frame of the sheet). The "average disorientation" of a texture component is called the arithmetic mean of the disorientations coi of each crystal of the component with respect to the ideal orientation.

For the calculation of the surface (or volume) fraction of the Cube orientation, a dispersion of 10.0° is used in order to show the positive role of certain addition elements on hypertexturing, the ideal being to obtain all the grains of a microstructure in this dispersion in order to approach the single crystal.

In this usual case of 90-95% overall TR reduction rate, the texture does not achieve the intended goal and is much too imperfect to satisfy many of the intended applications (superconductor substrates, magnetic circuits for magnetic amplifiers or voltage regulators, substrates for Si for photovoltaics, etc.).

It is known that the increase in the stored energy difference (driving force for recrystallization) between the Cube texture and the other texture components in the deformed state, develops the Cube texture component, the latter having the lowest energy. This increase is easily accessible by high cold rolling strain rates TR, i.e. much higher than 90%, typically 98%, i.e. for strip thicknesses lower than 0.02-0.07 mm which is very low for cold rolled products, and especially for certain applications.

Indeed, few applications require such low thicknesses. On the contrary, silicon substrates for photovoltaics require greater thicknesses (0.2 mm for example) to ensure a certain rigidity. The same is true for substrates for superconductors. Only OLED substrates may be of interest. Cube-textured FeNi magnetic core transformers in cut pieces (E-shaped, I-shaped, C-shaped, etc.) typically require thicknesses of 0.2 mm. High overall TR reduction rates are therefore not sufficient to achieve adequate hypertexture at the mainly targeted thicknesses of 0.08 to 0.25 mm, corresponding to the majority of current applications.

Increasing the final annealing temperature always increases the sharpness and the volume (or surface) fraction of the Cube texture component. It can go from, typically, 90% to more than 98%, when high work hardening rates are applied. Two possibilities are then available: either annealing is carried out at a high temperature under a reducing gas such as H2; but then, as seen in Figure 1, abnormal growth appears from, typically, 1000°C in the example considered, which therefore significantly limits the range of improvement of the Cube texture component which would be achieved by the reduction of the average disorientation and the consumption of twins (therefore by the increase of the volume or surface fraction of the Cube texture) and other crystallographic orientations; or annealing is carried out at high temperature under neutral gas, such as Ar, and then the normal grain growth of the Cube component can be continued up to 1100 or 1200°C without the appearance of abnormal growth; but in this case, and despite all the careful precautions (purification of the neutral gas, use of a "getter" to capture oxygen in a preferential manner compared to the treated product, etc.) that can be taken to make the annealing atmosphere free of oxygen or water vapor, the surface of the metal oxidizes significantly at high temperature, making the surface of the material unsuitable for certain preferred uses of the material; and even the intergranular oxidation of the material can make it unsuitable for these same preferred uses, or for other uses which require high magnetic properties (Hc, p, magnetic losses).

The final annealing temperature increase is therefore not, at least on its own, adequate for most applications.

The inventors discovered, unexpectedly, that the coincidence on Fe-Ni or austenitic Nickel-based alloys:

- a high reduction rate during cold rolling TR overall (typically > 90%, preferably > 95%), preferably without intermediate annealing for partial or total recrystallization; - the presence of several tens of ppm of S + Se + Te, namely between 20 and 60 ppm (beyond 60 ppm, industrially the hot transformation of alloy ingots would not be possible without difficulties), as well as the presence of several tens to hundreds of ppm of Nb (10 to 5000 ppm), possibly accompanied by certain controlled quantities of residual elements Ta, Al Ti, Hf, B;

- the tolerance of only a very limited presence of Zr + Hf, namely less than 500 pm (preferably < 100 ppm); allows to obtain:

- at least 99.0% surface fraction, preferably at least 99.5%, of Cube orientation grains, namely as belonging to a {100}<001> texture component with less than 10.0° of misorientation relative to the ideal (010)[100] orientation;

- at most 4.0° of average disorientation œ (average of the distribution of angles between the <100> axes of the crystals of the distribution and the DL direction, which is also the [100] axis of the ideal orientation CUBE (010[010]) of the grains of the Cube texture component {100}<001>;

- an abnormal growth start temperature of at least 1030°C, in order to make the production of such hypertextured materials compatible with the adjustment uncertainties and temperature non-uniformities of industrial furnaces, while ensuring that sufficient development of the Cube texture component is achieved at a very high level (fraction > 99.5%, œ < 4.0°).

With such a surprising discovery, based on the presence, with limited and well-chosen contents, of very minor elements (S, Nb, Ti, Al, B) in very hardened austenitic Fe-Ni alloys, it is now possible on an industrial scale to develop, during a final annealing in a closed vessel (also called "static annealing") around 1000°C, a Cube hypertexture with almost 100% Cube grains having a very low average misorientation œ, and which is therefore particularly suitable for applications of epitaxy on substrate, low apparent magnetostriction or even rectangular hysteresis cycle.

Concerning the magnetic properties, as just mentioned, one of the targeted applications is the rectangular hysteresis cycle, which applies to a magnetic amplifier or a low-noise transformer, for example. In this case, another good control (this time indirect) of the quality of the Cube texture is the examination of the values of the coercive field Hc and the remanent field Br. The closer the Cube texture approaches 100%, and the more it is weakly disoriented, the higher Br is and the lower Hc. The closer we get to a "good" texturizing annealing, that is to say very close to TM and avoiding the the beginning of abnormal growth, the lower the average disorientation of the Cube texture by enlargement of the Cube grains, and the lower Hc. Under these conditions, the lower the magnetic losses will be (transformer, amplifier).

Note that the Hc value is strongly dependent on the TG grain size. The more TG increases, the more Hc decreases, so abnormal growth, in particular, significantly reduces Hc. It should therefore not be said that we systematically seek the lowest possible Hc. In reality, we seek the lowest Hc with 100% Cube grains: this is what normal growth of only Cube grains allows up to around 1000-1020°C, overlapping with a reduction in the average texture misorientation. The average texture misorientation is, moreover, another factor reducing Hc.

We will first detail the reasons which led the inventors to choose the alloys which are one of the elements of the invention.

All contents are given in % by weight.

When we speak of "traces" of a given element, we must understand that this element may not be added voluntarily and that it may be absent, or that its possible presence, if it is detected by the means of analysis (rightly or wrongly depending on the precision of the instruments), then only results from the fusion of the raw materials used during the production of the alloy and from possible contamination of the alloy by the environment during production (atmosphere, refractories of the production containers, slag, etc.).

The alloys used in the invention are austenitic materials: Fe-Ni alloys or Nickel-based alloys, of a general type known to provide a strong Cube texture component after strong work hardening (at least 90%, typically) and annealing at a fairly high temperature, as long as abnormal growth does not occur. The alloys of the invention therefore contain at least 30% Ni for the material to be austenitic. Beyond 60% Ni, texturing provides, however, much less advantage in terms of usage properties, because the electromagnetic constants (magnetocrystalline and magnetostriction) of high-Ni FeNi become low anyway and hypertexturing is no longer of interest. And, moreover, this low constants are obtained at the cost of a high and fluctuating material cost (due to Ni), which the invention makes it possible to overcome. The Ni content is therefore limited to the 30-60% range.

Co does not change the austenitic behavior of Fe-Ni, nor its stacking fault energy. Co can increase the saturation magnetization for alloys with 30 to 35% Ni, but is much more expensive than Ni: its presence beyond traces resulting from the elaboration is not essential, and its possible addition is limited to 10%, without this hindering the Cube hypertexturing capacity. Mn can be added up to 3% without significantly reducing the stacking fault energy or the saturation magnetization J _sa t. Beyond 3% Mn, Cube texturizing ability may degrade, and Mn only further dilutes the magnetic moments of Fe and Ni. It may be present only in trace amounts.

Cu is known to increase the saturation magnetization J _sa t of Fe-30 alloys containing 35% Ni while it does not modify the stacking fault energy at all. Therefore, it allows to maintain the same Cube hypertexturing capacity. On the other hand, it dilutes the magnetism of alloys containing more than 35% Ni and causes precipitation of Cu-rich phases, precipitation which becomes significantly magnetically degrading from 10% Cu. Therefore, the possible addition of Cu is limited to 10%, and Cu may only be present in trace amounts.

Mo, W, V and Cr significantly degrade the stacking fault energy of Ni-based alloys, hence their Cube hypertexturing capacity. Cr improves corrosion resistance. Mo, W, V are elements of the same price range as Ni. Mo, W, V and Cr degrade the saturation magnetization J _sa t (dilution of magnetism), but they increase the electrical resistivity (thus allowing the reduction of magnetic losses by induced currents). A good compromise between these favorable and unfavorable influences is to limit the total Mo + W + Cr + V to 10%, when at least one of these optional elements is present beyond traces resulting from the elaboration.

Si can be used as a deoxidizer during alloy production. However, it decreases J _sa t and increases electrical resistivity. Therefore, there is no need to add a significant amount of Si if the final product applications do not cause induced currents. The Si content in the alloy is limited to 4%, and it may be present only in trace amounts.

It is an effective deoxidizer during processing, if this takes place under vacuum as is preferable, but it can be replaced by Si or Al (for example) for this use, with the risk of creating oxidized inclusions which are undesirable for the mechanical properties of the final products and the ability of the metal to undergo very high reduction rates during cold rolling, and which will not always be easy to eliminate during processing. It can also cause magnetic aging by precipitation of carbides.

It may only be present in the final product in trace amounts, especially after decarburization. However, a minimum of 30 ppm, preferably 50 ppm, of residual C resulting from decarburization can be accepted. The minimum content of 50 ppm is preferable for practical and economic reasons, as it does not require excessively extensive, and therefore costly, decarburization of the liquid metal. Decarburize the liquid metal to a greater extent than less than 50 ppm (if this content or a higher content low is not already obtained without special action) would certainly not bring a gain in magnetic or metallurgical quality which would justify the additional costs required.

The C content should be limited to no more than 500 ppm, preferably no more than 200 ppm, and better still no more than 150 ppm, to avoid the precipitation of carbides.

It should be understood that for C, the preferred lower limits and the preferred upper limits are independent of each other. In other words, it is generally necessary that traces < C < 500 ppm, with, preferably 30 ppm < C, better 50 ppm < C; and, preferably, C < 200 ppm, better C < 150 ppm.

The presence of small amounts of Zr or Hf degrades the average disorientation œ of the Cube texture, of the order of 10 to 20% from 200 ppm of Zr + Hf. But Hf and Zr are also strong inhibitors, making it possible to significantly delay the start of the abnormal growth destructive to the Cube texture; as the elements Zr or Hf can be combined with elements such as S, it may be interesting to add a significant quantity of these elements depending on the desired temperature stability of the Cube texture. However, as we are looking for a high level of hypertexturation, we are careful to limit Zr + Hf to < 500 ppm, preferably at most 100 ppm.

S is notoriously known to seriously damage the metal during hot forming, typically between 700 and 1100°C. For contents of at most 60 ppm of S, this brittleness can be reduced to a level sufficient for industrial hot forming by very preferred additions of 0.3 to 0.5% of Mn (the precipitation of at least 90% of the S in MnS leaves only a few free S atoms to embrittle the metal at the grain boundaries). But from 60 ppm of S, even with an addition of Mn of this order, the free S at the grain boundaries becomes sufficient to industrially embrittle the metal between 700 and 1100°C, or even 1200°C, making production difficult during hot forming. The invention places the maximum tolerable S at 60 ppm.

However, the inventors have discovered that not only is a minimum of 20 ppm of S required to obtain Cube hypertexturing as required by the invention, namely a volume or surface fraction of Cube texture > 99.0%, better > 99.5%, and an average misorientation œ < 4.0°, and also that this necessary condition is not sufficient.

Furthermore, Se and Te have the same effects as S, and we must reason not on the S content taken in isolation, but on the total S + Se + Te. It must therefore be between 20 and 60 ppm. It is also necessary to add at least 0.01% Nb (and/or other elements with a similar effect, see below) to have both this hypertexturation and abnormal growth starting above 1030°C.

The inventors have, in fact, discovered that Nb hinders the start of abnormal growth, which then begins at more than 1030°C, which thus makes it possible to extend the growth of Cube texture grains over a wider temperature range, while further reducing the average disorientation œ.

However, Nb is not only a germ or grain growth inhibitor, but also a precipitate-forming element with, for example, C (forming NbC carbides), which degrades the Cube microstructure and magnetic properties if the Nb content is too high, i.e. above 0.5%.

However, Nb can be partially, or even totally, replaced in a comparably efficient manner by other elements: Ta, Hf, Al, Ti or B.

Consequently, it is considered that, according to the invention, it is the total of the contents of Nb, Ta, Hf, Al, Ti and B which must be between 0.01 and 0.5%.

Furthermore, the individual contents of Al, Ti and B must be sufficiently low, as indicated below.

Al is limited to 0.02% to avoid the formation of oxides, nitrides, NisAI intermetallics, and an Al-rich diffused layer on the surface of the metal.

Ti is limited to 0.06% to avoid the formation of nitrides, oxides, and NisTi intermetallics.

B is limited to 0.06% to avoid the formation of NB nitrides.

Nb and Ta are heavy atoms that hinder the movement of grain boundaries. Al, Ti, B, which segregate at grain boundaries quite easily, can easily transform into nitrides, or even oxides, and trap grain boundaries with strong misorientation.

The alloy according to the invention can be produced by the following succession of steps, with possible variants which will be indicated later:

- production and casting, continuously or in ingot, then hot transformation, of a bloom or an alloy slab, namely a thick semi-finished product (a few cm to a few tens of cm thick) of an alloy of composition conforming to what has been said; this semi-finished product can also be obtained by compacting and sintering, from powders of the different alloying elements, or from one or more pre-alloyed powders;

- hot rolling of the semi-finished product until a strip a few mm thick is obtained, after reheating to between 1000 and 1300°C, preferably between 1100 and 1250°C; the 6LAC thickness after hot rolling is between 4 and 10 mm, preferably between 4.5 and 8 mm; the reduction rate during this hot rolling is not important, the microstructure and texturing obtained being in any case completely modified by the subsequent operations;

- cold rolling of the hot-rolled semi-finished product between the thickness 6LAC and a final thickness ef less than or equal to 0.5 mm; this cold rolling can be carried out in one or, preferably, several stages; the overall reduction rate TR of the cold rolling between 6LAC and ef is at least 90%, better still at least 95%; for each of the stages taken in isolation, the reduction rate is preferably at most 20%, better still at most 15%, or even at most 10% (it is not excluded that certain stages meet these preferred criteria and that the others do not). Indeed, low reduction rates per pass are preferable to further reduce the shear component - and the related textures - which develop in the subsurface, and thus maximize the development of the Cube texture during the subsequent annealing; an intermediate recrystallization anneal can be inserted between two cold rolling stages at a temperature typically between 600 and 1000°C, for 30 s to 10 hours, typically 2 to 5 minutes in flow annealing and 1 hour to 5 hours in static annealing, but this is not always the most recommended operating method; in fact, the greater the work hardening before the final texturing annealing, the more intense the Cube texture, so inserting an intermediate recrystallization annealing tends to reduce the work hardening that would be favorable before the final annealing; however, if this intermediate recrystallization annealing is introduced in the early stages of cold rolling, this reduces the work hardening little before final annealing, while it allows the microstructure of the different castings to be "standardized" industrially before a cold rolling with strong work hardening, thus promoting the reproducibility of the results on an industrial scale; it is therefore acceptable to carry out such an intermediate recrystallization annealing between two cold rolling stages, if experience shows that its location in the process chain of operations and its precise conditions do not prevent the desired objectives in terms of texture of the final product from being achieved; preferably the parameters of the rolling and heat treatments undergone by the alloy are adjusted so as to obtain a microstructure with small grains, i.e. an average grain size < 50 pm, before the first cold rolling and/or in the early stages of cold rolling; typically this grain size is to be obtained at a time between the end of hot rolling (i.e. the time when the strip has a thickness of 6LAC) and a time when the strip has a thickness of at least 1 mm; it is also possible to insert, between two cold rolling stages, one or more intermediate restoration annealings (typically between 500 and 700°C, lasting 30 s to 24 hours depending on the type of annealing (static or on the flow) and depending on the annealing temperature, but absolutely avoiding the recrystallization of the material, at any thickness of the cold rolling; such an intermediate restoration annealing can be carried out between all the cold rolling stages, or only between some of them; but if there is an intermediate recrystallization annealing, the intermediate restoration annealing(s) must be carried out after it to be effective;

- preferably, mechanical polishing and/or chemical pickling of the surface of the strip, at one or more intermediate stages between the end of hot rolling and obtaining the final thickness of the cold-rolled strip, in particular and preferably before the start of cold rolling or before the intermediate recrystallization annealing, to remove the incrustations and oxides present on the surface of the strip; it should be understood that this polishing and this pickling only have their usual functions of improving the surface quality, and have no influence on the microstructural properties targeted for the final product;

- possible cutting of transformer parts, substrate, and any other application in the final metal strip resulting from cold rolling, preferably by non-mechanical means (without contact of the part with a cutting or forming tool);

- final heat treatment, preferably in a closed container to allow the texture time to develop, this final heat treatment being carried out at a temperature which does not allow abnormal growth to start, therefore below TM, and in a reducing atmosphere, for example under hydrogen.

As it is the cutting zones that initiate the first places of abnormal growth, we preferably choose a cutting method that introduces the least possible local work hardening, such as water jet cutting, laser cutting, electroerosion cutting, rather than mechanical cutting by shearing, stamping, punching, pressing, etc. This further illustrates in particular the need to obtain a sufficiently high abnormal growth start temperature TDCA in the core zone of a recrystallized plate, because abnormal growth can begin in the cutting zones, depending on the cutting method, from 10 to 50°C below this TDCA temperature.

Thus, for example, an Fe-48%Ni alloy, cold worked to 98%, containing only 40 ppm of S and no Nb (nor elements comparable by their effects to S and Nb respectively), and having only residual contents of other elements (therefore not being in accordance with the invention), has a TDCA measured at the core of the plate (far from the cutting zones) of 1030°C, and yet it is observed that the abnormal growth begins at 1020°C on the laser cutting edges, 1010°C on the cutting edges by water jet, and 995°C or even less on the mechanical cutting edges. The average texture misorientations œ are then respectively 3.7° (plate core), 4.1° (laser cutting areas), 4.4° (water jet cutting areas) and 5.3° (mechanical cutting areas); we can therefore see that it is very difficult, if not impossible, to obtain industrial production of parts of various shapes by different cutting methods with an average Cube texture misorientation assuredly < 4.0° throughout the material.

We will now describe this last example in more detail.

A material of composition: is produced industrially in an arc furnace, under slag, for a few dozen tons of liquid metal, with under slag refining, before casting into ingots, then reheating, then forging to a thickness of 80 mm between 1100 and 1300°C. Then the forged ingot is hot rolled to a thickness of 4.6 mm, before mechanical polishing and direct multi-pass cold rolling without intermediate annealing to a final thickness of 0.2 mm. The overall TR reduction rate is therefore 95.7%.

For example, we make precisely 34 passes of successively 13; 12.5; 11.4; 12.9; 11.1; 12.5; 11.9; 10.8; 9.1; 10.0; 11.1; 8.3; 9.1; 8.0; 8.7; 8.3; 9.1; 7.1; 7.7; 8.3; 7.3; 7.8; 8.5; 7.0; 7.5; 8.1; 8.8; 8.1; 7.0; 5.7; 6.0; 6.4; 4.6 and 4.8% in order to limit the development of shear inherent in cold rolling, to a very low amplitude sub-surface zone of the metal so that the texture components linked to this shear do not interfere with the development of germination and then growth of the desired Cube orientation grains.

The composition of this casting is representative of the prior art (the alloy contains S and Nb only as residual impurities), but the alloy develops all the Cube texturing and abnormal growth mechanisms already discussed previously, and this material is therefore representative of the interaction between these mechanisms and the techniques for cutting magnetic parts, whether or not the composition is strictly in accordance with the invention. E-shaped or I-shaped sheets for forming cut-and-stack transformer cores are taken from the cold-rolled strip by various cutting means (shear, laser and water jet).

First, the legs of the E's are cut with shears to dimensions that allow them to pass through the refining furnaces, which are put under hydrogen atmosphere to prevent oxidation of the samples. The backs of the E and I have the same types of cuts, and only these are observed.

The heat treatments undergone after cutting are as follows: temperature rise of 300°C/h, maintenance for 3 hours at a temperature T(°C), then cooling to 200°C/h, with T = 950, 970, 980, 1000, 1020 and 1040°C depending on the tests.

For T = 1040°C, we observe the presence of very large grains over the entire sample, a sign that growth was abnormal everywhere.

For T = 1020°C, abnormal grain growth is observed near the edges of the sample, particularly near areas where cutting has been most complex (for example where holes or notches have been cut in the sample instead of making a straight cut over a long length).

For T = 1000°C, abnormal growth zones are observed mainly in the vicinity of the cutting zones with the most complex shapes.

For T = 980°C, abnormal growth trigger zones are observed only in the immediate vicinity of the most complex cutting zones.

For T = 970°C, a simple initiation of abnormal growth is observed at the level of complex cutting zones.

For T = 950°C, no abnormal growth is observed anywhere.

The use of non-contact cutting processes with a cutting tool therefore makes it possible to significantly raise the temperature from which we are sure not to observe an initiation of abnormal growth in the cutting areas: this temperature is 980°C for water jet cutting, which has the disadvantage of not guaranteeing excellent cutting precision; this temperature is 1000°C in the case of laser cutting, the precision of which is, moreover, satisfactory.

Cutting methods that are less aggressive than a conventional cutting tool, in that they cause little or no local work hardening, therefore make it possible to raise the annealing temperature in Cube texture alone, and thus to obtain a tight “cubic” texture across the entire sample.

The reason is that since the Cube texture comes from the hardening texture, if the hardening texture is locally modified by cutting, then, during the subsequent annealing, the resulting texture diverges more or less from the Cube texture. Worse, not only are some grains not Cube at all, but they are very disoriented relative to the matrix of Cube grains surrounding them (just a little after the cutting area) so they are very mobile, and therefore do not need to be heated to as high a temperature as Cube grains to move and degrade the Cube texture.

The average misorientation of the Cube texture component {100}<001> is calculated as explained below.

The industrial alloy not in accordance with the invention given as an example, work-hardened to 95.6% up to 0.2 mm thick, annealed under the best conditions at the limit of abnormal growth (1020°C in this example), makes it possible to obtain 98.8% of Cube grains (therefore below the limit of the invention which is 99.0%) and an average disorientation of the Cube texture component of 4.4°, greater than 4.0°. The industrial alloy tested is therefore not in accordance with the invention.

Conversely, an Fe-48%Ni alloy having the composition of casting 14 given in Tables 3a and 3b below, 98% cold worked, containing approximately (40 ppm of S, 350 ppm of Nb and 500 ppm of Ti and having only residual contents of other elements (and therefore having a composition conforming to that required by the invention) has a TDCA measured at the core of the plate (therefore far from the cutting zones) of 1050°C. It is noted that, when this alloy was produced under the same conditions as those of the reference example just cited, the abnormal growth begins at 1045°C on the laser cutting edges, 1035°C on the water jet cutting edges, and at less than 1025°C on the mechanical cutting edges, and the average misorientations œ are then, respectively, 3.9° (core of the plate), 3.9° (laser), 3.95° (water jet) and 4.0° (mechanical cutting).

We therefore see that it is possible, with alloys according to the invention, and for production conditions similar to those of the reference example, to obtain industrial production of parts of varied shape by different cutting methods with an average disorientation less than or equal to 4.0°. But this low average disorientation generalized to the entire cut strip is better ensured with cutting processes without contact with a cutting tool than with mechanical cutting processes.

Cutting the strip before the final recrystallization annealing to obtain parts with the dimensions, often small, for their future use, also allows this annealing to be carried out in a closed vessel (static annealing), in one or more enclosures each having a relatively small volume, and with good efficiency. If this closed vessel annealing must be carried out on the uncut strip, this is often in the form of a relatively large coil, requiring the use of an annealing enclosure of suitable dimensions. In addition, the fact that the strip is coiled can be a source of inhomogeneities in the treatment of the coil, requiring, to compensate for them, a longer treatment than would be theoretically necessary.

The invention will now be described in more detail, with supporting examples.

If it is not already known, by reference to previous experiments or models, the determination of the triggering temperature of the abnormal TDCA growth on the strip of given composition, produced and treated under given conditions, which is the subject of the treatment according to the invention, is carried out in a temperature gradient furnace of a type known in itself and which is used, in the laboratory, to determine how the microstructure of a metal sample evolves as a function of temperature.

The characterizations of textures and microstructures are made very precisely by the EBSD technique (in English Electron BackScattered Diffraction, i.e. analysis by diffraction of backscattered electrons). The magnetic measurements, for their part, are made using a standard coercimeter, according to the standard IEC 60404-7 edition 2.0 of January 2019 (“method for measuring the coercive field of magnetic materials in an open magnetic circuit. Measurement of Hc up to 160 kA/m by compensation method (H-coil) of the demagnetizing field of the material and sensitive field probes”).

It should be understood that the grain size taken into account for the determination of the average grain diameter, within the framework of the invention, is reduced to their equivalent diameter, that is to say to the diameter of circular grains which would give them an area equal to their real area if they are not circular on the observed surface. The average diameter is an arithmetic mean (unweighted) of the equivalent diameters.

To obtain strips with an intense Cube texture, the inventors proceeded according to the operations detailed below.

Preparation

Reference FeNi alloys and those according to the invention were produced by traditional metallurgy, namely by melting raw materials in an electric induction furnace under vacuum. The chemical composition of the alloys obtained, in weight percentages, is as follows:

- Ni = 48%;

- Co = traces;

- Mn = 0.3%;

- Cu = traces;

- Mo = traces; - W = traces;

- Cr = T races;

- V = traces;

- If = traces;

- C = 0.01%;

- traces < Ti < 0.05;

- traces < S < 0.006%;

- traces < Nb < 0.05%;

- traces < Al < 0.05%;

- traces < Zr < 0.05%;

- B = traces;

- Fe = the rest.

The precise compositions of each alloy produced, whether reference or according to the invention, will be given later.

The liquid metal from the furnace is cast into ingots measuring 200 x 40 x 40 mm and is cooled naturally.

Transformation

The ingots are then hot transformed by blooming and/or hot rolling until bars with thicknesses between 1 and 10 mm are formed.

The bars are then mechanically polished before cold rolling in order to remove surface incrustations and oxides. By cold rolling the bars, on rolling mills having very low roughness rolls, cold-rolled strips 1 (called "strips") with thicknesses between 10 and 500 μm are obtained. Cold rolling, in the example described, is carried out in multiple rolling passes, with low reduction rates per pass (less than 15% per pass). But it could be carried out with higher reduction rates, without this affecting the results obtained. Generally, a maximum reduction rate of 20% per pass, better still less than 15%, is a preferred variant of the invention. Experience shows that relatively low reduction rates for the various passes are favorable to the development of the desired Cube texture.

Determination (if unknown) of the temperature at which abnormal growth is triggered TDCA Figure 2 shows the principle of an experimental measurement method, at least approximately, of the temperature at which abnormal growth begins (TDCA) and the maximum normal growth temperature (TM) on a strip 1. The temperature TM is slightly lower than TDCA, and it is considered, within the framework of the invention, that TM = TDCA - 20°C.

In a temperature gradient furnace 2 which is a standard instrument in metallurgical laboratories, that is to say that the temperature varies increasingly along the length of the furnace between approximately 700°C (far left part of furnace 2 in Figure 2) and 1400°C (far right part of furnace 2 in Figure 2), a long strip of the strip 1 from the cold rolling is placed, in a clearly marked manner along the axis of the furnace, for a time long enough for the effects of the temperature gradient existing in the furnace to be found on the strip 1. After this heat treatment, the strip 1 is cooled, and the start of the abnormal growth at a precise position corresponding to a given temperature, which is determined according to a prior calibration giving the temperature-position relationship in the furnace, is observed with the naked eye, and, more precisely, with an optical microscope.

Figure 2 shows that in zone 4 between the furnace inlet and the point where the maximum growth temperature TM is reached, the crystals are well ordered and of relatively homogeneous sizes. In zone 5 between the point where the TDCA temperature is reached and the furnace outlet, the crystals grow in an increasingly disordered manner in size and orientation. Between these two zones 4, 5 there is an uncertainty zone 6 where the microstructure is no longer entirely consistent with the ordered microstructure of zone 4, but where the abnormal growth is not yet fully established. The width of this uncertainty zone 6 is of the order of 20°C, and corresponds to the difference between TDCA and TM.

The annealing is then preferably repeated in a constant temperature furnace at ± 10°C of the temperature determined by the gradient furnace, in order to confirm or refine the measurement of the TDCA value obtained in the gradient furnace. The maximum normal growth temperature TM is then determined from the TDCA measurement as determined by the experiments carried out in the constant temperature furnace, and by setting TM = TDCA - 20°C.

Recrystallization heat treatment

Recrystallization annealing is carried out on work-hardened strips, at temperatures at most equal to the maximum normal growth temperatures (TM) determined in the previous step, in furnaces under a controlled atmosphere so as to avoid oxidation of the strip (hydrogen, argon, argon-hydrogen mixture, high vacuum). The recrystallization annealing temperature must be less than or equal to TM but preferably close to it: in a 98% austenitic Fe-Ni alloy, without the addition of other alloying elements, the temperature TDCA for the start of abnormal growth is around 1000-1020°C, and the maximum normal growth temperature TM is 20°C below TDCA if the cutting method is not taken into account (see previously the explanations on the interaction between abnormal growth and cutting method). It is considered that at least 1000°C, preferably 1020°C, and even better 1030°C, should be reached as the Cube texturing annealing temperature (analogous to TM), in order to be able to sufficiently reduce the average misorientation, to increase the fraction of Cube texture grains to almost 100%, and also to reduce the coercive field. In practice, depending on the composition and microstructure after cold rolling, for the alloys according to the invention the recrystallization annealing takes place at a temperature of 900 to 1150°C.

Starting from room temperature, the temperature rise rate should preferably be as low as possible to promote the development of the intense Cube texture and avoid thermal twinning. It is typically between 0.1°C/min and 10°C/min. Preferably, it is of the order of 1°C/min. The holding time at the recrystallization temperature must be sufficiently long (between 30 minutes and 600 minutes) to grow the Cube grains to large sizes, reaching at least 20 pm. Finally, after annealing, the temperature reduction rate to room temperature is preferably higher than the rise rate to avoid a too slow passage below the Curie point creating induced magnetic order, which would disturb the magnetic properties.

Preparation of samples before characterization

Samples for magnetic characterizations:

The samples do not undergo any special preparation, other than mechanical cutting of the sample to the dimensions required by the measuring devices. The magnetic properties are directly measured on the strips after heat treatment.

Samples for characterization of microstructures and textures by EBSD: Microstructure and texture analysis by EBSD requires a clean and shiny surface condition, and therefore specific sample preparation. Thus, after cutting to the desired dimensions, the annealed samples are mechanically polished on SiC paper and with diamond felt. Mechanical polishing is followed by chemical or electrolytic polishing, in order to remove surface microdeformations caused by mechanical polishing. The samples are then rinsed in an ethanol bath and dried with compressed air before EBSD analysis.

Characterizations

Magnetic characterization:

Magnetic measurements are made in a standardized manner either in open circuit with a coercimeter (according to the IEC 404-7 standard already cited), or by a fluxmetric device comprising primary and secondary circuits, in a closed circuit, and are, in this case, deduced from the major hysteresis cycle measured (according to the IEC 404-6 standard - edition 3.1 of July 2021 “fluxmetric measurement method for magnetic properties”).

Characterization of microstructures and textures by EBSD - Surface fraction of grains with Cube orientation fcube and average misorientation œ:

EBSD characterizations allow crystal orientation maps to be obtained from which the orientation of the grains that make up the microstructure can be observed. From the EBSD data, the surface fraction fcube of Cube-oriented grains, the average misorientation œ of the grains with respect to the ideal Cube orientation, and the average grain size 0 are determined by quantitative analysis. The fraction fcube is determined with a maximum dispersion of 10.0° with respect to the ideal orientation, i.e. only grains that have a misorientation less than or equal to 10.0° with respect to the ideal Cube orientation are considered to be Cube-oriented grains or belonging to the Cube texture component {100}<001>.

In some studies, calculations of the surface fraction fcube of Cube-oriented grains are made with a dispersion of crystallographic orientations largely exceeding 15° compared to the ideal Cube (100)[001] orientation. This makes the conclusions of these studies questionable, because other texture components distinct from the Cube component - but close to it in terms of orientations crystallographic - can then be included in the quantification of the Cube texture.

The EBSD method for characterizing crystallographic orientations (allowing the selection and counting of grains having at most 10.0° of misorientation with the ideal Cube (010)[100] orientation, the calculation of the misorientation of each of these grains with this ideal Cube orientation, the calculation of the fraction of grains considered in Cube texture, the calculation of the average misorientation with respect to the Cube orientation) used in the invention is applied to the different materials studied by selecting different maximum misorientations between grain orientation and ideal orientation, in order to calculate different surface fractions of Cube texture component, and thus in order to study the angular distribution of Cube grains in the population of Cube grains. Cube grains were thus counted for maximum misorientations between Cube (010)[100] orientation and grain orientation of 15°, 10°, 7°, 5 or 4°.

In order to account for grains having reduced misorientation with the ideal Cube orientation, the inventors limited the maximum value of the misorientation coi of each grain taken into account (relative to the ideal orientation (010)[100]) to 10.0°.

In the calculation of the average misorientation œ, only the grains of the texture component considered (in our case: the texture component Cube {100}<001> encompassing all the grains having individually a maximum misorientation of 10° with respect to the ideal Cube (010)[001] orientation) are taken into account, and the value obtained is a weighted average of the misorientations of these grains with respect to the ideal Cube (010)[100] orientation.

The formula that defines œ is: coi being the misorientation of grain i in absolute value with respect to the ideal Cube orientation, and Si being the area of grain i with misorientation Wj. This misorientation Wj between the orientation of grain i and the ideal Cube orientation cannot be by definition greater than 10.0° since the inventors have selected this value as the relevant limit for counting the grains of the Cube texture component. The Cube texture component {100}<001> being in fact the set of grains whose <100> axis closest to DL is less than 10° from DL (because it is also the [100] axis of the ideal Cube orientation (010)[100]), the misorientations Wj are the angles between these <100> axes of misoriented grains and the DL axis, thus describing an angular distribution of positions of axes <100> with respect to DL, in a cone centered on DL of half angle at vertex 10°. This axial symmetry in cone-shaped distribution does not give meaning to a positive or negative value of the misorientation value Wj of each grain axis <100> with respect to DL. All calculated values of Wj are therefore arbitrarily counted as positive, such as those resulting for example from a scalar product of two weakly misoriented vectors. Since the values of oi are then all made positive, the average misorientation value co is necessarily also positive.

In EBSD, the surface of the sample is previously electropolished. Scanning the electron beam in lines parallel to the surface of the sample, each line consisting of repeated stops of the beam at regular intervals A (intervals smaller than the smallest of the grains) allows, at each stop, to characterize and store the crystallographic orientation (O.C.) of the crystal, and to create a two-dimensional mesh (X, Y) of the surface in O.C. In post-processing, the O.C. at the coordinate position (x,y) is compared to the Cube orientation (010)[100] - called ideal (see previously) - to determine its misorientation. Thus, in EBSD, the X,Y scanning of the electron beam makes it possible to collect on the mesh diffraction points at regular intervals A, Ni points which present the same disorientation coi with respect to the Cube (010)[100] orientation - called ideal - and these Ni points constituting a set of neighboring points.

Thus a grain / of surface Si is a set of pixels (or points) neighboring each other, of a minimal size (for example three pixels), of the same disorientation Wj with respect to the ideal Cube orientation {100}<001 >, and bordered by points of different disorientation(s), the absolute value of the difference(s) being greater than a threshold value, for example 1.0°. These points therefore correspond to other grains (the change of disorientation occurs at the grain boundary). If a grain i, thus defined, includes Ni pixels or points (defined by coi), then the surface Si of grain i is Si = Ni. A ²

Furthermore, it is important that the number of grains considered is sufficient so that the co value is derived from a grain statistic representative of the metal. Advantageously, the number of grains considered is greater than 100, preferably greater than 200.

The characteristics of the FeNi alloys according to the invention are as follows:

- Abnormal growth initiation temperature TDCA: 1040°C < TDCA;

- Fraction of grains with Cube orientation {100}<001>: 99.0% < fcube, better 99.5% < fcube,

- Average disorientation œ of the grains relative to the ideal Cube {100}<001> orientation: œ < 4.0°;

- Average grain size 0: 40 pm < 0 < 200 pm. These characteristics are obtained by means of the treatments described, but also of the composition of the alloy according to the invention which, as has been said, is maintained within a restricted range in which particular care has been taken with the presence, or absence, or near-absence of certain elements.

Due to the electromagnetic characteristics (magnetocrystalline anisotropy and true magnetostriction coefficients X o and Xm), the strong Cube or {100}<001> type texture and the fairly large grain size given previously, it is inherent to the material that its coercive field is of the order of 0.09 to 0.35 Oersted (i.e. approximately 8 to 28A/m in SI units).

Examples and counterexamples

Examples of implementation of the invention and reference examples will now be described.

First, we will examine the effects of the elements S, Nb, Al, Ti, Zr.

Tables 1a and 1b group together the compositions of different examples of Fe - 48% Ni alloys which were produced in the vacuum electric induction furnace by mixing and melting the different raw materials conventionally used for this purpose. When a single chemical element X varies from one material to another, the same production and liquid bath was used, alternating the addition of element X and the removal of a portion of liquid metal in a mini-ingot mold: in this way, all the other chemical elements in this same liquid bath remain identical for the same series of additions of the element and since the additions of element X are very small in quantity and make the dilution of the other elements by the addition of X negligible.

Tables 1a and 1b: Chemical compositions of Fe-48%Ni alloy castings

The resulting liquid metal is cast into conical ingots. The ingots are kept at 1100°C for 6 hours and are then hot rolled at 1100°C, to be formed into 4.5 mm thick strips.

These hot-rolled strips are then cold-rolled into 50 μm thick strips, with rolling rates per pass of less than 20%, after prior degreasing of the strip surfaces.

Then, the abnormal growth trigger temperatures (TDCA) and the maximum normal growth temperatures TM are determined on the strips 1 in a temperature gradient oven varying between 700°C and 1400°C, according to the procedure seen above.

The samples from the strips are annealed in a furnace under a pure hydrogen atmosphere with a temperature rise rate of 5°C/min from room temperature to TM, with a 240 minute hold at TM and a temperature drop, at the end of the annealing, faster than the temperature rise, in this case between 300 and 30,000°C/h depending on the mass and thickness of the cold-rolled and annealed sample, up to room temperature. After annealing, part of the samples is used, for example, for measuring the coercive field with a coercimeter, and the other part is used for texture and microstructure analyses by EBSD.

The coercive field measurement is performed on raw samples from annealing, whereas for EBSD measurements, a surface preparation is mandatory. This consists first of mechanical polishing of the samples with SiC papers (800, 1000, 1200, 2400, 4000, respectively) and then on diamond felts in decreasing order of diamond particle size (3 pm, 1 pm, 0.25 pm, respectively). Electrolytic polishing is then applied, in order to remove the surface work-hardened layer left by mechanical polishing. The samples are cleaned in ethanol and dried with compressed air. The clean and shiny samples are introduced into the SEM/FEG (Scanning Electron Microscope equipped with a field emission gun) for EBSD analysis.

From the EBSD data, we thus obtain the mapping of the crystallographic orientations and the parameters such as the surface fraction of the Cube orientation (fcube), the average disorientation of the grains with the ideal Cube orientation (œ) and the average grain size (0).

Table 2 presents the characteristics measured on the different alloys of Table 1, which were treated according to the procedure previously described.

Table 2: Characteristics of the alloys in tables 1a and 1b

With :

TDCA: Temperature triggering abnormal growth

TM: Maximum normal growth temperature fcube: Fraction of Cube grains in a dispersion of less than 10.0° co: Average misorientation (of all grains in the microstructure relative to the ideal Cube orientation)

0: Average grain size in the microstructure

Hc: Coercive field

The temperature at which abnormal growth begins is greater than or equal to 1020°C in all the castings. However, we note that Nb, Ti and Zr allow an increase in this temperature depending on their content in the alloy. At the TM temperature, where each casting develops its maximum Cube fraction, we show that all the castings have a Cube fraction greater than 99% (between 99.5 and 100%), except for casting 1. For the first four castings, which differ in the S content (3.5 ppm for casting 1, 17 to 50 ppm for castings 2 to 4), it is shown that by adding S, the fraction of Cube-oriented grains is increased, considerably from an addition of 17 ppm of sulfur. A further addition of 17 to 35 ppm further increases the Cube grain fraction quite significantly. Since, moreover, we see in examples 1 to 4 that it is necessary to exceed 17 ppm S to see a clear fall in the average disorientation below 4.0°, we deduce that at least 20 ppm S is required to combine the advantages of a very high Cube fraction and an average disorientation below 4.0°.

A very interesting and unexpected result is the beneficial effect of Nb on the development of the Cube texture at TM. On the three castings 5, 6 and 7 which contain respectively 190 ppm, 340 ppm and 480 ppm of Nb, the developed Cube fraction thus reaches 100%.

However, in some articles, Nb is described as an alloying element that is detrimental to the development of Cube texture.

The surprising result that these tests have made it possible to obtain on this point can still be understood, because in previous studies, the annealings were carried out at temperatures significantly lower than the optimal temperatures TM determined according to the invention. It is considered that, according to the invention, the annealing temperature must be between 1000°C and TM for this positive effect of Nb to be observed. Minimum temperatures of 1020°C, even better 1030°C are still preferable.

The average misorientations are less than or equal to 4.3° in all the castings. But it is noted that S and Nb, when they are simultaneously present in a significant way, considerably reduce the average misorientation (œ goes down to 3.4° for casting 4 and down to 3.2° for casting 7). Only castings 1, 2, 12 and 13 do not respect all the characteristics of the alloys of the invention (fcube s 99.0%, better > 99.5%, and co < 4.0°). These castings contain practically no (cast 1) or little (cast 2) of S (and no Se and Te which could have replaced S), and no Nb (hence the fact that cast 2, although it contains 20 ppm of S, does not lead to a sufficiently low average misorientation). They also do not contain Ti or Al which have been said to have effects comparable to those of Nb against abnormal growth. S and Nb (and/or elements with similar effects), in the prescribed contents, are therefore both essential to ensure the desired low average misorientation of at most 4.0°.

At the current stage of research, the inventors believe that S (as well as Se and Te) is the main element that contributes to obtaining a low average misorientation. However, it is still necessary that this low average misorientation is not degraded during annealing due to abnormal growth. The presence of Nb (and/or Ta, Hf, Al, Ti, B) makes it possible to inhibit this abnormal growth and to significantly broaden the temperature range of the normal growth of the Cube component alone. Normal growth occurs at the expense of the other residual orientations, and the inhibitors that have been mentioned make it possible to further reduce the risk that, if the annealing is carried out at a relatively high temperature and/or for a relatively high duration, an average misorientation may occur, which would take it beyond the 4.0° which constitutes its acceptable upper limit according to the invention. The higher the S content (staying within the prescribed 20-60 ppm range, for the reasons given), the greater the potential for improving average disorientation.

Zr, at contents greater than or equal to 0.02%, leads to an increase in the TDCA and therefore in the Cube orientation fraction, but does not allow a reduction in misorientations, as shown by the comparison of castings 12 and 13. The grain sizes are different depending on the castings. It follows from this table that temperature is not the only parameter that affects the increase in grain size. By comparing, for example, casting 3 (where TM = 1010°C) with casting 13 (with TM = 1060°C), we see that the alloy that has undergone annealing at a lower temperature nevertheless has a larger grain size, the difference between these alloys being the greater presence of Zr in alloy 13. Zr therefore slows down grain growth. Furthermore, as we have seen, these flows 12 and 13 are free of Nb, and the average disorientation œ of their grains is therefore not satisfactory.

The Hc values are between 0.09 Oe and 0.35 Oe (respectively between 7 and 28 A/m), which is consistent in this type of Fe-Ni50% material with the relatively small grain size of the tested materials, of the order of 50pm.

We will now examine the combined effects of the elements Nb + Al, Nb + Ti and

Nb + B Tables 3a and 3b show, in the same way as for Tables 1a and 1b, the compositions of different examples of alloys developed and tested for this purpose, according to the same procedure as for the examples in Tables 1a, 1b and 2.

Tables 3a and 3b: Compositions of castings 14-16 of Fe-48% Ni alloys

Table 4 shows, under the same conditions as Table 2, the characteristics developed by the different alloys in Tables 3a and 3b.

Table 4: Characteristics of the alloys in tables 3a and 3b

It is shown in Table 4 that when both Nb and either Ti, Al or B are present, the Cube texture developed is very sharp, as in the examples in Tables 1a and 1b which contain only Nb in addition to Ni, C, Mn and S (and other residual elements given in the composition tables). In particular, heat 16 which has the highest TDCA temperature due to the combination of its Nb and B contents, has 100% Cube-oriented grains with the largest grain size of all alloys 1-16, although its Nb content of 0.031% is lower than that of alloy 7. The average misorientations of the three heats are equivalent to each other and less than or equal to 4.0°, as sought by the invention. The coercive field changes greatly here, in a ratio of 4, while the texture is almost identical and the grain size only varies from 50 to 75 pm. Even more astonishing, the material at the highest Large grain size has by far the highest Hc. This is clear evidence of effects due to precipitation of compounds, such as probably Boron Nitride (NB), which did not hinder the growth of Cube grains, but trapped the magnetic domain walls.

Experience also shows that the effects of the development parameters are consistent with what a person skilled in the art can typically expect.

Moreover, as mentioned, alloy ingots can be obtained not by classical melting of massive raw materials, but by powder metallurgy.

Recrystallization annealing can be done by modifying the thermal cycle compared to the one given as an example. It is thus possible to eliminate the temperature rise at a regulated and slow speed, by making a "hot furnace" charging of the work-hardened strip, that is to say by charging the work-hardened strip directly into a furnace preheated to the annealing temperature.

Texture quantification can be done by other analysis techniques such as X-ray diffraction and neutron diffraction.

Experience also shows that an additional annealing after the final annealing, this additional annealing being carried out under Ar at a temperature T > TM, makes it possible to increase the Cube texture fraction without reaching abnormal growth. It turns out that the Ar atmosphere during this additional annealing prevents the development of textures other than the Cube texture. T must not, however, exceed 1300°C so as not to cause irremediable oxidation of the metal in depth.

Indeed, all the examples described so far included a single final texturing anneal carried out under hydrogen, therefore in a reducing atmosphere.

Experience shows that if the final annealing is applied directly under Argon, the volume fraction of the Cube component and the average disorientation of this component deteriorate somewhat compared to what would be obtained by carrying out a final annealing directly under a reducing atmosphere.

But experience also shows that if, after the final texturing annealing in a reducing atmosphere, an additional annealing is carried out in a non-oxidizing or slightly oxidizing atmosphere, for example under argon, helium, hydrogen + argon, or even nitrogen if the residual oxygen content and the dew point are strictly controlled, at a higher temperature, i.e. between 20 and 200°C higher than that of the final texturing annealing previously mentioned, for 30 min to 4 h, then the development of the Cube texture continues (reduction of the average disorientation, increase of the volume fraction) without abnormal growth. This phenomenon of blocking abnormal growth could be explained by a slight oxidation of the grain boundaries, which does not prevent most of the grain boundary movements (hence the improvement of the Cube texture) but which would nevertheless be sufficient to inhibit the development of abnormal growth seeds.

The invention makes it possible to obtain a material that resembles a single crystal because a majority of the grains have the same orientation to within a few degrees (less than 10.0° from the ideal Cube orientation) and have an average misorientation of at most 4.0°.

The strips according to the invention can, after having been optionally cut, be used in particular for the following applications.

These hypertextured Cube foils can be used as substrates in the manufacture of photovoltaic cells by replacing glass and ceramic substrates which are fragile and hard. The textured metal foils according to the invention allow good epitaxy of Si deposition and can thus increase the efficiency of photovoltaic cells.

These hypertextured Cube foils can be used as substrates in the manufacture of superconducting cables. They provide the appropriate mechanical properties to the superconducting material as a whole and optimize the current density of the cables.

These hypertextured Cube strips are also very advantageous for magnetic applications. All grains in the microstructure have easy magnetization directions <100> in both DL and DT directions. They can therefore be used to form low-noise current transformer cores and applications that use two orthogonal directions of magnetic flux passage (aeronautical transformers, etc.).

Tables 5a and 5b show, in the same way as for Tables 1a and 1b, the compositions of different examples of alloys produced and tested, according to the same procedure as for the examples in Tables 1a, 1b and 2.

Tables 5a and 5b: Compositions of castings 17-19 of Fe-48% Ni alloys

Table 6 shows, under the same conditions as Table 2, the characteristics developed by the different alloys in Tables 5a and 5b.

Table 6: Characteristics of the alloys in Tables 5a and 5b

Claims

1.- Process for preparing an austenitic iron-nickel alloy strip, characterized in that: - a semi-finished product of an alloy is prepared by melting, casting and hot forming, or by powder metallurgy, the composition of which consists of, in weight percentages: - 30% < Ni < 60%; - traces < Co < 10%; - traces < Mn < 3%; - traces < Cu < 10%; - traces < Mo + W + Cr + V < 10%; - traces < If < 4%; - traces < C < 500 ppm, with, preferably 30 ppm < C, better 50 ppm < C, and, preferably, C < 200 ppm, better C < 150 ppm; - traces < Zr + Hf < 500 ppm, preferably traces < Zr + Hf < 100 ppm; - 20 ppm < S + Se + Te < 60 ppm; - 0.01% < Nb + Ta + Hf + Al + Ti + B < 0.5%; - traces < Al < 0.02%; - traces < Ti < 0.06%; - traces < B < 0.06%; the remainder being Fe and impurities resulting from the production; - said semi-finished product is hot rolled at a temperature of 1000 to 1300°C, preferably 1100 to 1250°C, until a strip with a thickness (6LAC) of 4 to 10 mm, preferably 4.5 to 8 mm, is obtained; - said strip is cold rolled, in one or more stages, to a final thickness (e _f ) with an overall reduction rate (TR) of at least 90%, preferably at least 95%, to obtain a strip; and - a final recrystallization annealing of the strip is carried out, preferably a static annealing, at a temperature between, on the one hand, 1000°C, preferably 1020°C, better still 1030°C, and, on the other hand, the abnormal growth start temperature (TDCA) reduced by 20°C, for 30 to 600 min, said final recrystallization annealing taking place in a reducing atmosphere. 2.- Method according to claim 1, characterized in that the cold rolling is carried out in several stages, and in that at least one of these stages is carried out with a reduction rate of at most 20%, preferably at most 15%. 3.- Method according to claim 1 or 2, characterized in that the recrystallization annealing of the strip is carried out from room temperature with a temperature rise rate, up to the annealing temperature, of between 0.1 and 10°C/min. 4.- Method according to one of claims 1 to 3, characterized in that, after the static recrystallization annealing of the strip, the rate of temperature reduction of the strip to ambient temperature is greater than the rate of temperature rise of the strip during the static recrystallization annealing. 5. Method according to one of claims 1 to 4, characterized in that the rolling and heat treatment parameters are adjusted so that the average grain size is < 50 pm at a time between the end of hot rolling and a time when the thickness of the strip is at least 1 mm. 6.- Method according to one of claims 1 to 5, characterized in that an intermediate recrystallization annealing at a temperature between 600 and 1000°C is inserted between two of the cold rolling steps, for 30 s to 10 hours, and in that the cold rolling steps following said recrystallization annealing have a cumulative reduction rate of at least 90%, preferably at least 95%. 7.- Method according to one of claims 1 to 6, characterized in that one or more intermediate restoration annealing operations carried out between 500 and 700°C, lasting 30 s to 24 hours, are inserted between at least two of the cold rolling stages, avoiding recrystallization of the material, said intermediate restoration annealing operation(s) taking place after said possible intermediate recrystallization annealing operation. 8.- Method according to one of claims 1 to 7, characterized in that an additional annealing is carried out under argon, or nitrogen, or helium, or hydrogen + argon, or hydrogen + nitrogen, after the final texturing annealing under a reducing atmosphere and at a temperature 20 to 200°C higher, preferably 50 to 200°C, than the temperature of the recrystallization annealing, for between 30 min and 4 h. 9.- Method according to one of claims 1 to 8, characterized in that mechanical polishing and/or chemical pickling of the surface of the strip is carried out, at one or more intermediate stages between the end of hot rolling and obtaining the final thickness of the strip, preferably before the start of cold rolling or before the possible intermediate recrystallization annealing. 10.- Method according to one of claims 1 to 9, characterized in that a cutting of the strip is carried out before said final recrystallization annealing, preferably a cutting by non-mechanical means. 11.- Austenitic iron-nickel alloy strip, characterized in that its composition consists of, in weight percentages: - 30% < Ni < 60%; - traces < Co < 10%; - traces < Mn < 3%; - traces < Cu < 10%; - traces < Mo + W + Cr + V < 10%; - traces < If < 4%; - traces < C < 500 ppm, with, preferably 30 ppm < C, better 50 ppm < C, and, preferably, C < 200 ppm, better C < 150 ppm; - traces < Zr + Hf < 500 ppm, preferably traces < Zr + Hf < 100 ppm; - 20 ppm < S + Se + Te < 60 ppm; - 0.01% < Nb + Ta + Hf + Al + Ti + B < 0.5%; - traces < Al < 0.02%; - traces < Ti < 0.06%; - traces < B <0.06%; the remainder being Fe and impurities resulting from the production; in that its fraction of Cube {100}<001 > orientation grains is at least 99.0%, preferably at least 99.5%, the Cube {100}<001 > orientation grains having a misorientation relative to the ideal Cube (100)[001] orientation of at most 10.0°, the direction considered for determining the misorientation being, for each of the grains, that among the <100> directions closest to the rolling direction (DL), in that the average misorientation (œ) of the Cube {100}<001 > orientation grains, relative to the ideal Cube (100)[001] orientation, is less than or equal to 4.0°, and in that the average equivalent diameter of all the grains of the strip is between 40 and 200 pm. 12.- Substrate for photovoltaic cell, characterized in that it was obtained from a strip according to claim 11. 13.- Substrate for superconducting cable, characterized in that it has been obtained from a strip according to claim 11. 14.- Core element of an electric current transformer, characterized in that it was obtained from a strip according to claim 11.