DE68924000T2 - Process for refining the magnetic areas of electrical steels. - Google Patents

Description

This invention relates to a method for improving the core loss by refining the magnetic domain wall spacing of electrical steels, in particular of electrical steel sheets or strips.

In electrical applications, e.g. power transformers, distribution transformers, generators, etc., a grain-oriented silicon steel is usually used. The most important property is the ability of the steel to allow cyclic reversals of the applied magnetic field with only limited energy loss. A reduction in this loss, referred to as "core loss", is desirable.

In the manufacture of grain-oriented silicon steels, it is known that the Goss secondary recrystallization structure, expressed in Miller indices (110) [001], leads to improved magnetic properties, in particular improved permeability and core loss, compared to non-oriented silicon steels. The Goss structure refers to the body-centered cubic lattice with the grain or crystal oriented in the cube-on-edge position. The structure or grain orientation of this type has a cube edge parallel to the rolling direction and in the rolling plane, with the (110) plane representing the sheet plane. It is known that steels oriented in this way are characterized by a relatively high permeability in the rolling direction and a relatively low permeability at a right angle to it.

In the production of a grain-oriented silicon steel are typical steps of providing a melt of silicon magnitude of 2 to 4.5%, pouring the melt, hot rolling, cold rolling the steel to a final gauge of, for example, 7 to 14 mils (0.178 to 0.356 mm), typically 7 or 9 mils (0.178 or 0.229 mm), with intermediate annealing in two or more cold rolling operations, decarburizing the steel, applying a primer of a refractory oxide, e.g. a magnesium oxide coating, to the steel and finally tempering the steel at elevated temperatures to induce the desired secondary recrystallization and performing a purification treatment to remove impurities such as nitrogen and sulfur. The development of the cube-on-edge orientation depends on the mechanism of secondary recrystallization, whereby during recrystallization, secondary cube-on-edge oriented grains preferentially grow at the expense of primary grains of other and undesirable orientation.

The terms "sheet" and "strip" are used interchangeably here and in the following and have the same meaning unless otherwise stated.

It is also known that, based on the experiments of numerous researchers or authors, cube-on-edge grain oriented silicon steels generally fall into two basic categories, namely, first, regular or common grain oriented silicon steels and, second, high permeability grain oriented silicon steels. A regular grain oriented silicon steel is generally characterized by permeabilities of less than 1850 at 10 oersteds (796 A/m) with a core loss of more than 0.400 watts per pound (WPP) (0.88 watts/kg) at 1.5 Tesla at 60 Hertz for a nominal 9 mil (0.229 mm) thick material. High permeability grain oriented silicon steels are characterized by higher permeabilities. The latter may be the result of alloy changes alone or in conjunction with with process changes. For example, high permeability silicon steels may contain nitrides, sulfides and/or borides which contribute to the precipitates and inclusions of the inhibition system which itself affects the properties of the finished steel product. Furthermore, such high permeability silicon steels are generally cold worked to a final dimension, with a final severe cold work on the order of more than 80% being carried out to facilitate grain orientation. Although such higher permeability materials are desired, such materials tend to produce larger magnetic domains than a typical material. In general, larger domains are detrimental to core loss.

It is known that one of the ways to reduce the domain size and thereby the core loss value of electrical steels is to treat the steel in some way to introduce local stresses into the steel surface. Such practices may be generally referred to as "domain quenching by scribing" and are carried out after the final high temperature tempering treatment. If the steel is scribed after the final structural tempering, a localized stress state is introduced into the structurally tempered steel such that the domain wall spacing is reduced. These disturbances are typically relatively narrow, straight lines or scribe marks, generally spaced at regular intervals. The scribed lines are essentially transverse to the rolling direction and are typically made on only one side of the steel.

When these electrical steels are processed into transformers, the steel inevitably suffers from some deterioration in core loss quality due to cutting, bending and the construction of the cores during manufacture.

All such measures introduce undesirable stresses into the material. During manufacture in the production of stacked core transformers, and particularly in the power transformers of the United States of America, the deterioration in core loss quality due to the fabrication process is not so severe that stress relief annealing (SRA) is required to restore normal properties. For such end-use purposes, there is a need for a flat, domain-annealed silicon steel that does not require stress relief annealing. In other words, the scribed steel used for this purpose need not have a heat-resistant domain temper.

However, during fabrication in the manufacture of most distribution transformers in the United States, the steel strip is trimmed and subjected to a variety of bending and forming operations. The latter create more machining stresses in the steel than in the case of power transformers. In such cases, it is necessary and common for manufacturers to subject the product to stress relief annealing (SRA) to relieve such stresses. However, it has been found that during stress relief annealing, the valuable effect on core loss due to certain scribing techniques, such as mechanical and thermal scribing, is lost. For such end-use applications, it is necessary and desirable that the product be treated with a heat resistant domain (HRDR) to maintain the improvements in core loss values due to scribing.

It has been suggested in the known patent literature that impurities or "intruders" may be effective in improving magnetic domain wall spacing or grain oriented silicon steel. US Patent No. 3,990,923 to Takashina et al., dated November 9, 1976, describes that a primary recrystallized silicon steel can be subjected to a chemical treatment to control or inhibit the growth of secondary recrystallization grains. From GB patent application 2 167 324A a method for dividing magnetic domains or grain oriented silicon steels to survive a (SRA) is known. In this method, the sheet is subjected to a voltage and an "intruder" is formed on the grain oriented sheet. The "intruder" consists of a component or structure different from the electrical sheet. The "intruder" is formed before or after the voltage is introduced. This is followed by annealing, for example in a reducing hydrogen atmosphere, to introduce the "intruders" into the steel body. Numerous metals and non-metals are specified as the intruder material.

Japanese Patent Document 61-133321A describes a method for producing an oriented electrical steel sheet with ultra-low core loss, in which a surface coating on the steel sheet is completely or partially removed to expose the underlying steel, a chemical coating is applied to the sheet, deformations are then formed at intervals, and the whole is heat treated to react the chemical coating with the exposed steel. This involves heat-resistant domain annealing. If desired, stress relief annealing can then be carried out to obtain a stress-free steel.

Japanese Patent Document 61-139 679A describes a method for coating a magnetic steel sheet that has been oriented to the end of its structure in the form of spaced-apart lines or spots with at least one compound selected from the group consisting of phosphoric acid, phosphates, boric acid, borates, sulfates, nitrates and silicates. During subsequent firing at 300 to 1200ºC, a penetrated or penetrated layer is formed while tempering the magnetic domains. Body that is different from that of steel.

Japanese Patent Document 61-284529A describes a method of removing the surface coatings of final-annealed magnetic steel sheets at intervals, applying zinc, zinc alloys and/or zincate alloys in specific coating weights, applying one or more metal(s) having a lower vapor pressure than zinc, and forming impregnated bodies different from the steel in composition or structure at intervals by heat treatment or coating with an insulating layer to temper the magnetic domains.

Japanese Patent Document 62-512202 describes a method for improving the core loss of silicon steel by removing the forsterite film formed after the final finish annealing and attaching another metal, e.g. copper, nickel or antimony, by heating.

There is now a need for a process for aging the magnetic domain wall spacing of grain oriented silicon steel having thereon a heat resistant primer such as forsterite. The process should be compatible with the usual treatment of regular and high permeability silicon steels and should make use of the heat insulating coating such as forsterite primer on the sheet to facilitate domain aging. Furthermore, the process should be operable with a variety of techniques including common techniques for removing the primer in predetermined patterns.

The present invention relates to a method according to the definition of claim 1. Preferred embodiments are specified in the dependent claims 2 to 14.

Generally speaking, the invention provides the person skilled in the art with a A method is provided for annealing the magnetic domain wall spacing of a grain oriented silicon steel sheet or strip having an insulating primer thereon. The method comprises removing portions of the primer to expose a line pattern of the underlying silicon steel and applying a metallic impurity to the silicon steel. The metallic impurity may be copper, tin, nickel, zinc or antimony or combinations or compounds thereof. The exposed steel is free of thermal and plastic stresses and independent of such stresses for effective domain annealing. Thereafter, the steel and the impurity thereon are annealed at a temperature of 760°C (1400°F) or more in a protective atmosphere to diffuse sufficient and controlled amounts of the impurity into the exposed steel with the objective of producing lines of permanent pores to produce a heat resistant domain anneal and reduced core loss in the stress free steel.

According to one embodiment of the invention, a barrier layer of phosphorus or silicate or combinations or compounds of the same is applied to the steel sheet to seal the primer before application of the metallic contamination.

The invention is explained in more detail using the attached figures. In detail:

Fig. 1 is a photomicrograph of a copper-filled indentation in the silicon steel primer according to the present invention;

Fig. 2 is a photograph at 150x magnification of an X-ray map of copper according to Fig. 1.

Fig. 3 is a microphotograph at 3000x magnification of a silicon steel after diffusion tempering, in which the porosity achievable according to the invention is visible;

Fig. 4 a photomicrograph at 30x magnification of antimony nodules on a steel surface;

Fig. 5 and 6 Micrographs of Ni-Sn strips in cross section and

Fig. 7 to 10 Photographs at 3000x magnification of X-ray maps of nickel, tin, phosphorus and iron in the Ni-Sn stripe in the steel.

Broadly stated, the inventive method consists in a method for enhancing the domain structure of a grain-oriented silicon steel sheet of relatively large grain sizes by controlled chemical surface contamination. The method starts from a final-textured, tempered silicon steel as the sheet raw material with an electrically and thermally insulating primer already applied. The primer is then locally removed using one of a number of techniques to expose the bare metal. The metal does not require any plastic strains or stresses of any kind. The exposed bare metal is then contaminated by other materials on the surfaces of the exposed metal pattern. Finally, the steel is annealed to diffuse or alloy the contaminant into the iron-silicon steel sheet product. The domain enhancing achieved is heat-resistant because it survives stress relief annealing.

The starting material for the chemical lining process according to the present invention is the grain-oriented silicon steel tempered to the final structure with an insulating coating already applied thereon. Such an insulating coating can be the usual primer, also known as forsterite or rolled glass coating. Preferably, the just washed, grain-oriented silicon steels tempered to the final structure can be used. Such steels can be the regular or usual grain-oriented silicon steels or grain-oriented silicon steels with a high permeability. The special compositions of such Steels are not critical for the present invention, but rather they can be of conventional compositions. Steel melts that can be used in the present case initially have the following nominal composition: Steel Rest

Steel 1 is a high permeability grain-oriented silicon steel. Steel 2 is a conventional grain-oriented silicon steel. Steel 3 is a modified conventional grain-oriented silicon steel. In this case, all information regarding composition is given as weight percentages unless otherwise stated.

Steels 1, 2 and 3 were manufactured by casting, hot rolling, normalizing, cold rolling to a final gauge with intermediate tempering using 2 or more cold rolling steps, decarburizing, coating with MgO and final structure annealing to ensure the desired secondary recrystallization of a cube-on-edge orientation. After decarburizing the steel, a refractory oxide coating containing mainly magnesium oxide is applied prior to final structure annealing at elevated temperature. During the latter tempering or annealing, a reaction occurs on the steel surface to form a forsterite primer. Although the steel melts of steels 1, 2 and 3 initially had the nominal composition previously stated, C, N and S were reduced to trace amounts of less than about 0.001 wt.% after fine or final structure annealing.

According to the invention, it is important that parts of the coating or primer are removed to expose a line or stripe pattern in the underlying silicon steel How the coating or primer is removed is not critical to the present invention. However, unlike the prior art, the underlying steel need not be subjected to any mechanical, thermal or other stresses as a result of the removal of the coating or primer. In other words, the exposed steel must be free of any thermal and plastic stresses before any subsequent steps of applying the metallic contamination are carried out. An advantage of the present invention is that a variety of techniques can be used to remove selected portions of the primer. For example, mechanical or laser scribing can be used to form a controlled pattern of markings on the strip surface. The line or stripe pattern selected for the removed primer can be any conventional pattern used in known scribing techniques. Preferably, the patterning can consist of removal of the primer or coating in generally parallel lines substantially transverse to the rolling direction of the steel with conventional line widths and line spacing. Depending on whether the grain-oriented silicon steel is cube-on-edge, cube-on-face or other orientation, other patterns are also suitable.

According to the invention, the exposed silicon steel is plated or coated with selected metals or metal alloys. The metals are preferably selected such that their diffusion rate is lower or their diffusion speed is slower than that of iron in silicon steels. The metals and metal alloys suitable for the invention are referred to as "contaminants" or "diffusion materials". The term "contaminants" here and below refers to certain suitable metals and metal alloys for the inventive selective application to the exposed surfaces of the steel sheet. It has been found that a variety of metallic contaminants selected from copper, tin, nickel, zinc or antimony or combinations or compounds thereof can be used. The metallic contaminants can be applied as a coating to the silicon steel by a variety of conventional means, e.g. by electroless deposition or electrolytic plating. Because of the insulating nature of the primer, the metallic contaminant or metallic contaminant can only be applied to the selected line patterns or stripes which correspond to the pattern of primer removal. It is important that the insulating glass primer on the silicon steel allows selective deposition of the metallic contaminant in the predetermined or preselected pattern.

The silicon steel with selected portions of the primer removed and the metallic contaminant applied is then annealed in a protective atmosphere for a time and temperature to diffuse sufficient and controlled amounts of the contaminant into the exposed steel and thereby form permanent pores. The purpose of this action is to provide a heat-resistant domain temper and reduced core loss. The annealing provides a diffusion annealing with slight alloying of the metallic contaminant with the iron-silicon steel sheet to provide a heat-resistant domain temper. The annealing temperature ranges from about 1400ºF (760ºC) or more to 2100ºF (1150ºC). Preferably, temperatures range up to 1800ºF (982ºC), more preferably from about 1400 to 1700ºF (760-927ºC).

It is advisable to use an annealing temperature that is at least equal to or higher than the temperature normally used for stress relief annealing, so that the property effects achieved remain stable during any lower temperature downstream treatments, such as stress relief annealing (SRA). In other words, the improvements in core loss are the result of a heat-resistant domain temper. The duration of the annealing may be up to 20 hours. Preferably it is from 30 minutes to 5 hours at a temperature sufficient to ensure the magnetic domain temper. For practical reasons, the diffusion annealing should take place at a higher temperature than a conventional stress relief annealing at about 1425ºF (774ºC) which transformer manufacturers may use after fabrication. Temperatures on the order of up to 1650ºF (899ºC) are sufficient for the heat-resistant domain temper without the need for an additional release coating to prevent thermal welding of adjacent coil windings during annealing. Lower annealing temperatures can also be used successfully.

As is well known, virtually complete homogeneity is a highly desirable condition for soft magnetic materials. It has been found that a suitable time and temperature will develop and stabilize the permanent pores and also diffuse the impurities into the steel, thereby obtaining a virtually homogeneous steel sheet throughout the steel thickness. In general, tempering at higher temperatures promotes homogeneity. In all tempering or annealing operations according to the present invention, the strip can be annealed either in coil form or as an annealing strand of the continuously moving strip after application of the metallic impurity.

For a better understanding of the present invention, the following examples are presented. Unless otherwise stated, the metallic impurities or contaminants used in the examples were obtained from the The plating solutions given in Table I were selected and electrolytically deposited. Table 1 Plating metal solutions and conditions Tin Nickel Copper Zinc Antimony Tin(II) sulfate Sulfuric acid Room temperature Stainless steel anodes or Tin anodes Nickel sulfate Nickel chloride Boric acid Temperature: Nickel anodes Copper cyanide Sodium cyanide Sodium hydroxide Room temperature Copper anodes Zinc sulfate Ammonium chloride Stainless steel anodes or Tin(II) chloride Nickel chloride Ammonium bifluoride Ammonium hydroxide (adjustment of a pH value of 2.5) Antimony oxide Hydrofluoric acid Beta-naphthol 1 eye drops/1 Ambient temperature Stainless steel Anodes or stainless steel * The current density applies to the entire strip area

Example I

Two Epstein packets of nominally 8 mil (0.2 mm) thick, high permeability, grain oriented silicon steel sheet of composition Steel 1 were mechanically scribed after washing. The scribe effectively removed portions of the primer in a pattern of substantially parallel lines substantially transverse to the rolling direction of the steel sheet. Each Epstein packet contained twelve (12) strips. Each strip was 3 cm wide and had the scribe lines spaced approximately 5 mm apart. Immediately after mechanical scribe, a stress relief annealing was performed at 1500ºF (816ºC) for 2 hours. The specimens were then electroplated using the copper solution shown in Table I and then annealed in a protective atmosphere at 1650ºF (899ºC) for 5 hours to diffuse the metallic contaminant into the silicon steel sheet body. The percentages in parentheses indicate a change compared to the initial properties. The magnetic properties were determined in a manner usual for Epstein packages. Table II Initial after washing Mechanically scratched 2h stress relief annealing at 1500ºF (816ºC) Chemically stripped (copper) + 1650ºF (899ºC)/5h tempering Epstein Package Permeability at 10H Core Loss at 1.5T at 1.7T

The results of Table II show that mechanical scribing improved core loss somewhat, as expected, due to some plastic deformation damage to the underlying steel sheet. However, this improvement was not thermally stable. After stress relief annealing at 1500ºF (816ºC), the plastic deformation disappeared and the properties essentially returned to the initial (pre-scribing) values of the primed and final-structure tempered silicon steel. The steel showed only the pattern of the exposed underlying metal. The chemical copper line treatment of the present invention followed by diffusion annealing at 1650ºF (899ºC) shows a significant improvement in core loss, averaging about 6%. This clearly demonstrates that the chemical lining treatment of the present invention can be effective regardless of any plastic or thermal stresses or deformation of the steel. After subsequent stress relief annealing at 1450ºF (788ºC) for 2 hours, the specimens show permanent core loss improvement indicative of heat resistant domain tempering on each specimen. These specimens confirm that thermal or plastic deformation of the exposed silicon steel plays no role in heat resistant domain tempering.

Fig. 1 is a scanning electron microscope photomicrograph of a notch, i.e. the silicon steel exposed through the primer, which is filled with copper after copper plating of the specimen.

Fig. 2 is a 150x magnification photograph of an X-ray chart showing copper in the line pattern of the silicon steel specimen.

Example II

Single-strip Epstein specimens 8 mils (0.2 mm) thick and 1 1/2 inch wide of the steel composition of Example I were chemically pickled in HCl-1% HF acid to remove all insulating primer from the annealed strips. A plastic scriber with slots was attached to the steel surface so that the slot pattern formed substantially parallel lines substantially transverse to the rolling direction of the steel strip as in Example I. Each specimen with the scriber thereon was electrolytically copper plated as in Example I and then annealed at 1650ºF (899ºC) for 2 hours (after removal of the scriber) to diffuse the metallic contaminant into the silicon steel body. Percentages in parentheses represent changes from the original properties. Magnetic properties were determined in a manner conventional for single-strip tests. Table III Chemically striped (copper) + 1650ºF (899ºC)/5h tempering Sample No. Initial permeability at 10H Properties Kerner loss at 1.5T at 1.7T

The results in Table III show that, although the test specimens were never subjected to plastic deformation or stress, they had significantly improved core loss properties after diffusion annealing. The improved properties clearly demonstrate that plastic deformation plays no role in domain tempering by the chemical lining or striping according to the invention.

Metallographic examination of the various specimens in the diffusion annealed zone showed no excessive attack of the steel substrate by the plated deposit. This is consistent with the small amount of contaminant deposited and the relatively low diffusion temperatures used. Structures at high magnification tended to change and complicate, as is not uncommon in diffusion coupling metallurgy. The results suggest that the domain annealing of the invention is independent of the development and/or maintenance of subtle composition gradients within the chemically lined or treated region. Instead, this effect appears to be the Kirkendall porosity phenomenon well known in diffusion coupling metallurgy. Without being bound to any particular theory, the mechanism related to the Kirkendall phenomenon seems to suggest that contaminants of other chemical natures can be successfully used as chemical line contaminants and that the precise chemical character is not as important as the diffusion rate with respect to the ferrous base metal. Furthermore, adhering to the Kirkendall diffusion theory, it appears that lower diffusion temperatures tend to exaggerate the Kirkendall porosity and that preferably the post-annealing should be done within the lower range of 1400 to 1700ºF (760 to 927ºC).

Example III

Individual strips of high permeability grain oriented silicon steel according to Examples 1 and 2 were mechanically or in some cases electrically discharge scribed after washing. The scribe effectively removed portions of the primer in a pattern of substantially parallel lines substantially transverse to the rolling direction of the steel strip. The lines were approximately 3 mm wide and spaced approximately 5 mm apart on individual 8-mil (0.2 mm) thick Epstein strip specimens. The specimens were then electrolytically plated with a variety of metallic contaminants from the plating baths shown in Table I. The plating resulted in at least half of the recesses in the primer being filled with the respective metallic contaminant as shown by microscopic examination. After plating, the specimens were diffusion annealed in an atmosphere consisting essentially of hydrogen as indicated. Table IV Original Properties Chemically striped Annealing Sample No. Permeability at 10H Kerner loss at 1.5T 1.7T Metal Duration Temperature Table IV (continued) Original Properties Chemically striped Annealing Sample No. Permeability at 10H Kerner loss at 1.5T 1.7T Metal Duration Temperature

The results contained in Table IV show that for the tin, nickel and antimony impurities provided by the plating baths and plating conditions given in Table I, the steels showed an improvement in core loss properties at both 1.5 and 1.7 Tesla with only a slight loss of permeability. Since the specimens were heated to temperatures above the typical stress relief annealing temperature of 1425ºF (774ºC), the core loss improvements were permanent with respect to heating to that temperature. In other words, the improvements were "heat resistant."

Example IV

Additional samples were tested to determine the thermal stability of the magnetic properties of the samples treated in accordance with the invention. If the improvement in magnetic properties is the result of Kirkendall porosity, then these improvements should be extremely stable. All samples were 8 or 12 Epstein ribbon packages. Each ribbon of nominally 8-mil (0.2 mm) thick steel of the composition of Example 1 was given lines approximately 0.25 mm wide and spaced approximately 5 mm apart by lightly mechanically scratching through the thin glass primer to expose the bare steel and electroplating with the metals shown in Table V. The ribbons were then stacked and diffusion annealed at 1650°F (899°C) for 5 hours. A significant improvement in core loss properties was evident. The Epstein packages were then further annealed at 2100ºF (1150ºC) for 2 hours and 10 hours as indicated. The percentages in parentheses indicate a change from the initial properties. Table V Initial Properties Chemically stripped + 1650ºF (899ºC)/5h anneal After further annealing at 2100ºF (1149ºC) Sample No. Permeability at 10H Kerner Loss at 1.5T at 1.7T Zinc Strip Nickel-Tin Strip Copper Strip ** Sample was annealed for 10 hours at 2100ºF (1149º) Table V (continued) Initial Properties Chemically Stripped + 1650ºF (899ºC)/5h Anneal After Further Annealing at 2100ºF (1149ºC) Sample No. Permeability at 10H Kerner Loss at 1.5T at 1.7T * Sample was annealed for 2 hours at 2100ºF (1149º)

The data and results in Table V demonstrate that the core loss improvement was sustained up to 2100ºF (1150ºC). This demonstrates the unique and exceptional stability of the domain temper of the invention. Furthermore, a scanning electron microscope (SEM) analysis and a chemical analysis using X-rays showed that the previously chemically ruled areas now had the same composition as the matrix steel. In other words, the high temperature had homogenized the strip. The contaminants were no longer localized, but formed part of the total residual contamination in the bulk of the specimen. The homogeneity and low stress state of the soft magnetic material is a desired result.

In view of the findings of the invention, it is expected that the metal impurities chosen will perform as expected if their diffusion rates are lower than the inherent diffusion rate of the iron in the iron-based alloy. Furthermore, the lower the diffusion rate through the iron, the more suitable the metal is as a contaminant to provide permanent porosity. For example, the diffusion of copper through iron is on the order of 4 times slower than that of iron. Nickel diffuses on the order of 500 times slower. Such metallic elements with slower diffusion rates in iron should result in the Kirkendall porosity phenomenon and the advantages of the invention.

Although the metallic impurities can provide a heat-resistant domain temper as a result of domain tempering of stress-relieved, grain-oriented silicon-iron (alloys) with a good primer thereon, the reproducibility was poor in cases where the primer was unsuitable. Apparently, the final structure annealing result, Naturally occurring forsterite primer sometimes allows unwanted plating of the metallic contaminant through the pores and cracks in the forsterite. This was evident when antimony nodules oozed and grew directly from pores in the forsterite primer in areas remote from the exposed silicon steel below and between the virtually parallel lines (Fig. 4).

It has been found that, in accordance with one embodiment of the invention, the use of an additional sealing coating or barrier layer on the forsterite prior to application of the metallic contaminant to the exposed silicon steel strips results in a significant improvement in the consistency and reproducibility of a heat resistant domain coating. The primary purpose of introducing the barrier layer into the process is to seal the pores and cracks in the forsterite coating. Table VI identifies several coatings believed to be suitable as the barrier layer of the invention. The similarity between all of these coatings is that they are all water soluble and cure at relatively low temperatures. Furthermore, these barrier layers contain phosphorus or silicates or combinations or compounds thereof as the major component of the coating. Preferably, the major component of the coating is a metal phosphate or metal silicate. In particular, the coating should be one which, after curing, deposits substantially as a magnesium phosphate layer. Table VI Name Barrier and Conditions Concentration Phosphoric Acid (85%) Magnesium Oxide Nalcoaq (1050) Registered Trademark Chromium Trioxide Water Hardened: 1000ºF (538ºC) - 1 min (air) Sodium Silicate (40-42 Be) Water Hardened: 800ºF (427ºC) - 1 min (air) Phosphoric Acid (85%) Magnesium Oxide Kasil #1 Registered Trademark Ammonium Hydroxide (58%) Chromium Trioxide Dupanol Registered Trademark (2%) Water Hardened: 800ºF (427ºC) - 1 min (air) Phosphoric Acid (85%) Magnesium Oxide Ammonium Hydroxide (58%) Chromium Trioxide Dupanol Registered Trademark (2%) Water Hardened: 800ºF (427ºC) - 1 min (air) Balance

The following examples serve to better understand this embodiment (of the invention):

Example V

Tests were conducted to demonstrate the effect of the barrier coating on improving the process for producing a heat-resistant domain temper. All specimens were made from various melts of a silicon steel of typical composition corresponding to Steel 1 and were nominally 8-mil (0.2 mm). Individual strips of Steel 1 were mechanically scribed to the scribed condition. Scribing effectively removed portions of the primer in a pattern of substantially parallel lines substantially transverse to the rolling direction of the steel strip. The lines were approximately 0.25 mm wide and spaced approximately 5 mm apart. Each specimen was coated with a barrier coating "P" as shown in Table VI after removal of the primer. All specimens were then either zinc-plated or copper-plated from a plating bath as shown in Table I. Magnetic properties were determined on a single 30 x 3 cm Epstein strip. After electroplating, all specimens were diffusion annealed for the time and temperature specified in Table VII. Percentages represent a change in core loss characteristics compared to the original characteristics. Magnetic properties were determined in a manner customary for individual strip tests. Table VII Original Properties After 5h/1650ºF (899ºC) Diffusion Annealing Permeability at 10H Core Loss at 1.5T at 1.7T See Change Contaminant Zinc Copper Average Improvement (all bands)

In comparison to control samples which were magnetically tested prior to any treatment, Table VII contains data demonstrating that most samples were on the order of 10 to 12% better in core loss on average. For comparison purposes, samples from the same batches of material after standard mechanical scribing techniques were capable of providing a 15 to 20% improvement in core loss. Another advantage of the present invention is that the improvement in core loss can be the result of a heat resistant domain temper.

Example 6

In further examples, additional tests were conducted to demonstrate the domain annealing process using various silicon steels of compositions 1, 2, and 3 for Epstein test packages. Each sample was prepared as in Example 5 above, but with modifications to produce different grain oriented silicon steels of nominal 7-mil (0.18 mm) and 8-mil (0.2 mm) size. The samples were then machined as in the previous example under the test conditions shown in Table VIII with parallel bands of treated areas spaced 5 mm apart. Magnetic properties were determined in a manner conventional for Epstein packages. Table VIII Initial Properties 1650ºF (899ºC) Diffusion Annealed Steel Dimension (mil) (mm) Impurity Permeability at 10H Core Loss at 1.5T at 1.7T Copper Zinc

The data in Table VIII demonstrate that the domain tempering process of the invention can reduce core loss in an 8-mil (0.2 mm) gauge material corresponding to Steel 1 by as much as 11% compared to the original properties. The greatest improvement is achieved when copper is used as an impurity. The core loss in the 7-mil (0.18 mm) gauge material corresponding to Steel 2 was reduced by about 5% at 1.5T and by 5% at 1.7T. The core loss in the 7-mil (0.18 mm) gauge material corresponding to Steel 3 was reduced by about 7% at 1.5T and by about 4% at 1.7T.

Example 7

Further tests were conducted to compare the results of a domain temper with and without a barrier coating on a silicon steel of the composition of Steel 1 for individual Epstein ribbons. Each specimen was prepared as in Example V. The ribbons were mechanically scribed to remove the thin primer in a pattern of virtually parallel lines transverse to the rolling direction. After the primer lines were removed, the "P" barrier coating was applied to the specimens thus marked. The specimens were then nickel plated or nickel-tin plated as shown in Table IX using suitable plating baths as shown in Table I. Finally, the specimens were diffusion annealed in hydrogen at 1600 to 1650ºF (871 to 899ºC) for 1 to 5 hours. The magnetic properties are results of individual Epstein ribbons of a nominal 8-mil (0.2 mm) thick 30 x 3 cm ribbon. The percentages in parentheses represent a change from the original properties. Table IX Without barrier layer Chemically nickel-plated lines or stripes Sample No. Original properties Chemical lining + diffusion annealing Permeability Kerner loss Average Table IX (continued) Without barrier layer Chemically nickel-tinned strips or lines Sample No. Original properties Chemical lining + diffusion annealing Permeability Kerner loss Average Table IX (continued) Barrier Electroless Nickel Plated Strips Sample No. Original Properties Chemical Lining + Diffusion Annealing Permeability Kerner Loss Average Table IX (continued) With barrier layer Chemically nickel-tinned strips or lines Sample No. Original properties Chemical lining + diffusion annealing Permeability Kerner loss Average

The results of Table IX clearly demonstrate the advantages achievable by the present invention. For comparable samples, the consistency and reproducibility of results were improved for the samples with the barrier coating. All samples with the barrier coating had better core loss values than comparable samples without such a barrier coating. Furthermore, none of the samples with the barrier coating had any nodules of the metallic contaminant on the sample surface in areas remote from the exposed line pattern. This indicates that the barrier coating blocked any pores or cracks in the glass primer that might have exposed bare metal (which would then have been plated with the contaminant).

In the course of these tests, it was unexpectedly found that the barrier layer not only seals the pores and cracks in the primer of the grain oriented silicon steel, but also acts synergistically with the main contaminant in the ruled area during and after diffusion annealing. In particular, it was found that in the case of zinc or nickel-tin, phosphorus was the main contaminant present in the permanent defect created in the steel.

A metallographic examination of the various specimens in the diffusion annealed zone showed no excessive attack of the steel substrate by the plated deposit. This is logical given the small amount of deposited contamination and the relatively low diffusion temperatures maintained. The structures at high magnification tended to change and complicate things, which is not common in diffusion coupling metallurgy. Confirmation of mutual diffusion between steel and contamination as well as between the steel, the contamination and the phosphorus is shown in Fig. 5 to 10. In Fig. 5, a cross section (at 375x magnification) through the Ni-Sn line after diffusion annealing at 1650ºF (899ºC) for 2 hours, the appearance of nodules on the steel surface is evident. Fig. 6 shows the same nodule at 3000X magnification. Figs. 7, 8, 9 and 10 clearly show the inclusion of nickel, tin, phosphorus and iron in the diffusion zone in a scanning electron microscope/X-ray mapping.

It was an object of the present invention to develop a process for providing a heat resistant domain temper for grain oriented silicon steels to improve core loss values. Another advantage of the inventive process is the ability to remove portions of the primer to expose a pattern of the underlying silicon steel, e.g. lines substantially transverse to the rolling direction, by conventional or previously unused means, provided that the steel exposed through the primer is free of thermal and plastic stresses. The barrier layer further enhances the core loss improvements and their reproducibility. An advantage of the present invention is that a semi-finished sheet product with a barrier layer and a metallic contamination can be produced for downstream tempering by the customer before or after processing into transformer cores.

Claims (14)

1. A method for tempering the magnetic domain wall spacing of grain-oriented silicon steel sheet or strip with an insulating primer thereon, characterized in that the method the removal of parts of the primer to expose a line pattern of the underlying silicon steel; the application of a metallic impurity to the exposed silicon steel, the metallic impurity having a lower diffusion rate than the iron in the silicon steel and the exposed steel being free of thermal and plastic stresses, and subsequently tempering the steel and the contaminant thereon at a temperature of 760ºC (1400ºF) or more in a protective atmosphere to diffuse sufficient and controlled amounts of the contaminant into the exposed steel to produce lines of permanent pores while ensuring heat-resistant domain tempering and reduced core loss of the stress-relieved steel includes. 2. The method of claim 1, wherein the step of tempering the steel and the impurity thereon includes temperatures up to 982ºC (1800ºF) to produce lines of up to 1149ºC (2100ºF) stable permanent pores. 3. A method according to claim 1 or 2, wherein the step of Tempering the steel and the impurity thereon to develop and stabilize the remaining pores at a temperature of 760ºC to 927ºC (1400 to 1700ºF). 4. The method of claim 1, wherein the step of tempering the steel and the impurity thereon includes temperatures up to 1149ºC (2100ºF) to further diffuse the impurities into the steel with the goal of providing a substantially homogeneous steel sheet through the steel thickness. 5. A method according to any one of the preceding claims, wherein the metallic impurity consists of copper, tin, nickel, zinc or antimony or combinations or compounds thereof. 6. A method according to any preceding claim, additionally comprising (prior to application of the metallic contaminant) the application of a barrier layer having a major component in the form of phosphorus or silicate or combinations or compounds thereof for sealing the primers to the primed steel. 7. The method of claim 6, wherein after application of the barrier layer, parts of both the barrier layer and the primers are removed and then the metallic contaminants are applied. 8. The method of claim 6, wherein portions of the primer are removed before the barrier layer is applied to the primer and the metallic contaminant is applied to the exposed steel after the barrier layer is applied. 9. The method of claim 6, 7 or 8, wherein the barrier layer is a metal phosphate based layer containing at least 25% by weight of phosphorus in the dried layer. 10. The method of claim 9, wherein the barrier layer consists of a magnesium oxide based layer. 11. A method according to any one of claims 6 to 10, wherein in the step of tempering the steel a reducing atmosphere of hydrogen or nitrogen or mixtures thereof is used. 12. A method according to any preceding claim, wherein the pattern comprises generally parallel lines of exposed steel extending substantially transversely to the rolling direction of the steel. 13. A method according to any one of the preceding claims, wherein the primer consists of forsterite. 14. A method according to any one of the preceding claims, wherein before the tempering step and after application of the metallic impurity, the silicon steel is processed into an article.