Thin layer magnetic structures for binary information stores
United States Patent 3880602
In binary data information stores, the storing member may comprise a thin layer magnetic structure made of at least one ferromagnetic layer, or of two such layers interacting through a non-magnetic film coupling. In such structures, the so-called "creep" phenomenon, and other disturbing phenomenons, lessen when the coercive magnetic field increases. Said coercive magnetic field is made of an increased apparent value from providing one of the faces of such a structure with a thin lamina of a magnetic material having a rigid magnetic lattice which is randomly inorganized at the microscopic scale.
US Patent References:
Magnetic material
Meiklejohn - June 1961 - 2988466
Magnetic material
Bean - November 1963 - 3110613
Magnetic recording medium
Prentky - November 1965 - 3219353
Magnetic bodies having magnetic anisotropy comprising conjoined thin films of molybdenum and nickel coated on a non-conductive substrate
Lommel et al. - March 1966 - 3243269
Magnetic data storage devices
Tsu - July 1967 - 3330631
Magnetic material
Meiklejohn - June 1961 - 2988466
Magnetic material
Bean - November 1963 - 3110613
Magnetic recording medium
Prentky - November 1965 - 3219353
Magnetic bodies having magnetic anisotropy comprising conjoined thin films of molybdenum and nickel coated on a non-conductive substrate
Lommel et al. - March 1966 - 3243269
Magnetic data storage devices
Tsu - July 1967 - 3330631
Inventors:
Valin, Jean (Les Essarts le Roi, FR)
Bruyere, Jean Claude (Sayssinet, FR)
Bruyere, Jean Claude (Sayssinet, FR)
Application Number:
05/108587
Publication Date:
04/29/1975
Filing Date:
01/21/1971
View Patent Images:
Export Citation:
Assignee:
Centre National de la Recherche Scientifique (Paris, FR)
, Compagnie Internationale Pour L'informatique (Louveciennes, FR)
, Compagnie Internationale Pour L'informatique (Louveciennes, FR)
Primary Class:
Other Classes:
428/941, 428/656, 428/668, 148/103, 428/928, 29/DIG.028, 148/101, 148/312, 148/102, 428/630
International Classes:
H01F10/10; H01F10/32; H01F10/00; H01F10/02
Field of Search:
117/234-240 29/195 148/31.55,57,101-103 340/174TF 252/62.54
US Patent References:
Primary Examiner:
Martin, William D.
Assistant Examiner:
Pianalto, Bernard D.
Attorney, Agent or Firm:
Kemon, Palmer & Estabrook
Parent Case Data:
This application is a continuation of our prior application Ser. No. 746,090, filed July 19, 1968 and now abandonded.
Claims:
We claim
1. A thin layer magnetic structure for use as a binary information store comprising:
2. Structure according to claim 1, wherein the face of said thin ferromagnetic layer opposite to said lamina is coated by a thin film of a non-magnetic material itself coated by a further layer of a ferromagnetic material.
3. Structure according to claim 2, wherein the material of said further layer is magnetically softer than the material of the ferromagnetic layer cooperating with said lamina.
4. Structure according to claim 1, wherein the material of said one lamina is selected from the group consisting of nickel-iron-manganese alloy and nickel-cobalt-iron-manganese alloy, and wherein said ferromagnetic layer consists of a material selected from the group consisting of nickel-iron alloy and nickel-cobalt-iron alloy.
5. Structure according to claim 1 wherein said laminae have at least one common metallic component which is an alloy including at least iron and nickel and in which said one lamina includes manganese as a distinct component to render it antiferromagnetic in character.
6. Structure according to claim 5, wherein the said common metallic component consists of a ternary iron-nickel-cobalt alloy.
7. Structure according to claim 5, wherein said one lamina is comprised of a part of the said ferromagnetic alloy containing manganese diffused therein.
1. A thin layer magnetic structure for use as a binary information store comprising:
2. Structure according to claim 1, wherein the face of said thin ferromagnetic layer opposite to said lamina is coated by a thin film of a non-magnetic material itself coated by a further layer of a ferromagnetic material.
3. Structure according to claim 2, wherein the material of said further layer is magnetically softer than the material of the ferromagnetic layer cooperating with said lamina.
4. Structure according to claim 1, wherein the material of said one lamina is selected from the group consisting of nickel-iron-manganese alloy and nickel-cobalt-iron-manganese alloy, and wherein said ferromagnetic layer consists of a material selected from the group consisting of nickel-iron alloy and nickel-cobalt-iron alloy.
5. Structure according to claim 1 wherein said laminae have at least one common metallic component which is an alloy including at least iron and nickel and in which said one lamina includes manganese as a distinct component to render it antiferromagnetic in character.
6. Structure according to claim 5, wherein the said common metallic component consists of a ternary iron-nickel-cobalt alloy.
7. Structure according to claim 5, wherein said one lamina is comprised of a part of the said ferromagnetic alloy containing manganese diffused therein.
Description:
SUMMARY OF THE INVENTION
The present invention concerns the thin layer magnetic structures which comprise at least one layer of ferromagnetic character made at least from iron and nickel, the hysteretic cycle of which is substantially rectangular in at least one privileged orientation. Such layers can be made either with an isotropic condition, no privileged magnetic orientation, and in such layers there is more iron than nickel; or they can be made with an anisotropic condition, then being uniaxially magnetically orientated along a privileged direction, i.e. an easy magnetization axis and in such layers there is more nickel than iron.
Binary date information stores can be made by adding to such thin layer ferromagnetic structures appropriate electrical conductor arrays enabling the operations of read-in and read-out of the information bits in and from so-called "memory points" as defined by the crossing places of the conductors of such arrays with, in some cases, additional definition from an appropriate cutting of the layers in strips and/or small areas underlying said crossing places.
In each and any of such stores the same principle of representation of the information bits applies: conventionally, the binary digital value 1 is affected to one magnetization condition of a memory point, as defined by a first local orientation of said magnetization condition, and the binary digital value 0 is affected to another condition of magnetization, usually characterized from a different orientation of the magnetization than the first one.
The magnetization conditions of the memory points must be stable with time in order to avoid deterioration of the content of the store, independantly of any further read-in and/or read-out operation at other memory points in the store. It is known that each read-in point gives rise to a demagnetizing field which is, at first approximation, substantially inversely proportional to the length of the memory point and substantially proportional to the thickness of the thin layer structure at this point. Such a demagnetizing field must firstly be lower than the coercive field of the magnetic material. The higher the value of the coercive field, the higher the demagnetizing field value at a memory point can be without appreciable drawback; hence, the higher the coercive field, the higher the density of information bits can be since each memory point can be made of a lesser area. Secondly, the demagnetizing field must not disturb the magnetization conditions of the neighbouring memory points and, in this respect too, the higher the coercive field value the lower will be the risks of deleterious interactions from memory point to memory point. Further, the higher is the value of the coercive field, the lesser the risks of disturbing the magnetizations of the memory points from stray or parasitic fields which are temporarily generated from read-in and read-out electrical currents for other memory point operations.
Referring now more particularly to magnetic structures made of thin anisotrope magnetic layers, wherein the memory points are characterized by locally saturated areas in the one and/or the contrary diraction in the orientation of the easy magnetization axis of the magnetic material, a known phenomenon is the so-called "creep" effect which, from repeated actions from parasitic fields of any kinds, slowly varies the magnetization conditions of the memory points from displacement of the walls separating the magnetization regions of these points, which obviously entails a deterioration of the content of the store. The creep facility is related to the value of the coercive field of the ferromagnetic material and the higher said value, the lower the creep.
It is the object of the invention to provide a thin layer magnetic structure comprising at least one ferromagnetic character layer consisting of at least iron and nickel, which is specially adapted for acting as a storing member in a binary data information store as herein-above defined in that it presents a value of the apparent coercive field much higher than the usual value of conventional thin layer magnetic structures.
According to a feature of the invention, such a thin layer magnetic structure is characterized in that it includes, on one face of a thin ferromagnetic character layer, a lamina of a material having a rigid magnetic lattice which is randomly inorganized at the microscopic scale.
According to a further feature of the invention, said lamina is comprised of an antiferromagnetic character material.
According to a further feature of the invention, said lamina is formed from part of a ferromagnetic layer wherein, in addition to iron and nickel, a material is alloyed which gives to said part the said antiferromagnetic character.
LIST OF FIGURES AND SHORT DESCRIPTION THEREOF
In order to illustrate these and further features of the invention, reference is made to the accompanying drawings, wherein:
FIG. 1 shows a thin layer magnetic structure in a first example of embodiment of the invention, comprising a single ferromagnetic layer 2 formed over and linked to a lamina 1 of a material having a rigid magnetic lattice of inorganized condition at the microscopic scale, said lamina being carried on any appropriate non magnetic substrate 5. As such a structure may be used in a store with or without magnetic orientation along an easy magnetization axis, the direction of said axis A is only shown in dot line at 6 in said FIG. 1.
FIG. 2 shows a second example of embodiment comprising two ferromagnetic layers 2 and 3, which may be or not of the same magnetic material and may be or not of identical thickness, said layers being coupled by means of an intercalated film 4 of a non magnetic material; said film 4 is preferably conducting. The remaining part of FIG. 2 is identical to the structure of FIG. 1.
A ferromagnetic two coupled layer structure is described in French Pat. No. 1,383,012 filed Oct. 18, 1963 by CENTRE NATIONAL DE LA RECHERCHE SCIENTIFIQUE. As known, the coercive field of such coupled layers is lower than the coercive field of each of the layers and consequently, the present invention will find an advantageous use in such structures since, from another point of view, the use of coupled layer structures is of advantage in magnetic stores as the coupling effect reduces to a very small value their response to external parasitic fields.
As coupled magnetic layer structures are specially adapted to uniaxial magnetization conditions, the direction of the easy magnetization axis 6 is shown in full line in FIG. 2.
The term "thin film" as used in the present specification must be understood as referring to thicknesses of layers from about 100 to some thousands of Angstroms, a very thin layer or film then being of a relatively appreciably smaller thickness; Illustratively, and for instance, when in FIG. 2 (or FIG. 1), the magnetic layer 2 is of about 800 A, the layer 3, of about 400 A, the layer 4 may be of about 80 A thick.
DETAILED DESCRIPTION:
In both FIGS. 1 and 2, the material of the lamina 1 must be such that its configuration or pattern of spins, which are randomly orientated at the microscopic scale of the crystallites, in the superficial lattice contacting the ferromagnetic layer 2 at least, be stable or rigid and unaffected by the action of any kind of external magnetic fields. Further, the energy of interaction between the spins of the lamina 1 and the spins of the ferromagnetic layer 2 at the contact region must be sufficiently low not to essentially disturb the intrinsic magnetic properties of of the layer 2, i.e. its capabilities of response to read-in binary digits or bits from local modifications of its magnetization. For uniaxial layers, further, the said energy of interaction must not unduly increase the dispersion (angular dispersion) of the components of the easy magnetization axis at the memory points, and must not unduly disturb the quasi-coherent rotation of the magnetization towards the difficult direction, perpendicular to the easy axis, as said rotation is conventionally used, as well known per se, for read-in and read-out operations in anisotrope thin layers.
As known an anti-ferromagnetic material, when used at a lower temperature than its order/disorder transition temperature, T N as conventionally denoted, preserves a rigid magnetic lattice which is not modified by application of any kind of magnetic field. It is consequently of advantage to constitute the lamina 1 in such an anti-ferromagnetic material, as the magnetic stores will always be operated at temperatures lower than T N , and consequently the random organization of the magnetic spins will not vary in such a material for such stores.
After the thin layer structure is made according to the invention, which will be later hereinbelow described, an important increase of the coercive field is obtained: for the single ferromagnetic layer structure shown in FIG. 1, the coercive field is at least twice the value of the coercive field of the ferromagnetic layer alone and, for the coupled layer structure of FIG. 2, the value of the coercive field is as high as 8 times the value such a coupled layer structure had without application of the invention.
Such important increases of the apparent value of the coercive field may be easily understood:- as currently known, the expression of the coercive field includes, inter alia, a term relating to the "exchange." The application or combination of the lamina 1 of antiferromagnetic material and the layer 2 of ferromagnetic material, the said lamina 1 having inorganized though stable condition at the microscopic scale of its crystallites, produces an increase of said "exchange" term by reason of the spin to spin coupling between the ferromagnetic and antiferromagnetic zones in contact, which coupling acts for opposing magnetization displacements in the ferromagnetic material.
If the lamina 1 did not present such a random condition of its spins, such a coupling would be of unidirectional character and the magnetization in the ferromagnetic layer would depend upon that of the lamina. As, on the other hand, the magnetic lattice of the material of the lamina 1 is randomly inorganized, everything is such that the exchange coupling finally gives or results in an anisotrope interaction at the macroscopic scale since the global or overall effect is the same for any portion one may consider in the ferromagnetic layer 2, and consequently for any volume comprising a memory point in said layer. It remains possible to control the orientation of the material at each memory point in said ferromagnetic layer from application of magnetic fields which, at any temperature lower than T N of the antiferromagnetic material, do not affect the random inorganization of the magnetic lattice of spins in the lamina 1.
The increase of the value of the coercive field makes it possible, as already said, to provide a higher density of the memory points whilst practically eliminating deteriorations of the content of the store from the action of the demagnetizing fields from said points and of the parasitic fields, either purely external or operational.
There are several ways of making the above-described thin layer magnetic structures:
Ferromagnetic materials are well known in themselves, and for instance and mainly, such materials comprise the nickel-iron and nickel-iron-cobalt alloys. Similarly, antiferromagnetic materials are well known in themselves, for instance cobalt oxide, chromium oxide, and the ternary alloy comprised of iron, nickel and manganese. It is sensible, by reason of the similarities of component elements, to consider as interesting couples ferromagnetic/antiferromagnetic materials such as nickel-iron and nickel-iron-cobalt for the ferromagnetic material and such as nickel-iron-manganese for the antiferromagnetic material, though such a choice must not be considered as excluding others.
Generally known, a thin magnetic layer is made from simultaneous evaporation of its component elements on a non-magnetic substrate such, for instance, a glass or ceramics of high melting point, which is heated at an appropriate temperature during deposition of the material, the elements of which are arranged in crucibles for a controlled evaporation thereof thus controlling a required proportion of such elements in the material formed on the substrate.
It is further known that, for obtaining a ternary alloy comprised of nickel, iron and manganese, a method may be used which consists of firstly depositing on the substrate a layer of manganese, secondly depositing a layer of a nickel-iron layer over the manganese layer, and thereafter annealing the product for producing thermal diffusion of the manganese within the adjacent part of the nickel-iron layer. Controlling the operation for only a partial diffusion of the manganese element within the ferromagnetic layer of nickel and iron (it would be the same with a layer of nickel-iron-cobalt) enables the attainment of a structure according to the present invention.
Illustratively considering such a method, examples may be detailed as follows:
EXAMPLE 1:
The substrate is heated at a temperature of the order from 300° to 350° C. A D.C. orientating magnetic field is applied to the location of said substrate. Evaporation of a layer of manganese up to, for instance, about 100 Angstroms is conventionally made. Thereafter, evaporation is made in conventional fashion of a layer comprised of nickel, iron and cobalt with, illustratively, about 20% cobalt, and, in the remaining 80%, a weight proportion of 81 to 19 of nickel with respect to iron, up to a thickness of the order of about 800 A.
Without changing such conditions of substrate temperature and D.C. field application (the value of said field being, for instance, about 20 Oe), the product is annealed from a period of 10 to 15 minutes. The manganese diffuses from thermal diffusion within the ferromagnetic layer so that in the lower part of said last named layer, adjacent the substrate, a lamina of antiferromagnetic material is obtained. Obviously, the concentration in manganese in said lamina varies with the thickness, decreasing from the substrate surface but this has no deteriorating action for the sought result since, in any transition, coupling is obtained between the spins of an antiferromagnetic material and the spins of a ferromagnetic material at a certain level of the structural complex which is obtained.
When the coercive field of the final structure is measured, it is seen that it is of the order of 7 to 7.3 Oersteds with the above defined conditions. In the same conditions, but without manganese, a ferromagnetic layer was obtained, which layer presents a measured value by 3.6 Oersteds only of coercive field. Consequently, it can be stated that a structure according to the invention presents a coercive field of about twice the value of that of a ferromagnetic layer.
Further, however, the antiferromagnetic lamina thus obtained in the structure reveals a satisfactorily inorganized random distribution of its spins so that there is no unidirectional coupling between it and the ferromagnetic material in the structure, which latter constitutes the storing member proper.
EXAMPLE 2:
The steps are the same as in Example 1, however the annealing step is maintained during about 1 hour. It is then seen that the antiferromagnetic lamina presents a unidirectional coupling with the ferromagnetic layer so that the structure cannot be used as such for a magnetic store of high coercive field storing member.
In order to destroy such unidirectional coupling, an alternating magnetic field (or even better, a rotating magnetic field resulting from application of two alternating fields in phase quadrature) is substituted to the D.C. field and annealing is carried out for a short length of time. The coupling proved to be destroyed and the results finally similar to those in Example 1.
It must be noted that, as a variation, the D.C. field is applied only during depositions of manganese and ferromagnetic alloy and annealing is solely made in presence of an A.C. or rotating magnetic field, during a length of time of about one quarter of an hour. Results are then satisfactory and substantially as herein before indicated.
EXAMPLE 3:
Coupled layer structure is sought. The first operations are made according to Example 1 or Example 2. Thereafter, evaporation of a very thin film of such a metal as molybdenum, of the order of 80 A in thickness for instance, is ensured over the ferromagnetic layer at a lower temperature of the substrate (preferably lower than 300° C.) Thereafter a further layer of ferromagnetic material is evaporated over the molybdenum film: for instance a layer of nickel-iron alloy with 81 to 19% ratio of components, without any cobalt, is evaporated up to a thickness of the order of 400 A.
It is seen that the antiferromagnetic lamina and the first ferromagnetic layer remain with no unidirectional coupling though the D.C. orientating field had been re-applied for the evaporation of the second ferromagnetic layer, by reason of the lower operative temperature. The measured coercive field is of the order of 4 Oersteds. However, with a similarly made coupled layer structure, omitting the manganese layer at the first step, hence omitting the antiferromagnetic lamina, the coercive field is only of a value substantially equal to 0.5 oersted. Consequently the structure according to the invention presents a value of coercive field increased by a coefficient 8 with respect to the corresponding conventional structure.
Obviously the above examples are not limitative in themselves. For instance, cobalt may be omitted for single ferromagnetic layer structures, as cobalt is thought of interest mainly with coupled layer structures; the nickel-iron ratios may be varied when the structure is not to have a negligible magnetostrictive coefficient; relative thickness can be varied at will, as being only indicatively given and such thicknesses chosen in accordance with considerations of the stores wherein the structures will be used.
Obviously, further, the superposition of two independant layers, antiferromagnetic and ferromagnetic made from successive evaporations on the substrate and without interpenetration thereof remains within the field of the invention.
The present invention concerns the thin layer magnetic structures which comprise at least one layer of ferromagnetic character made at least from iron and nickel, the hysteretic cycle of which is substantially rectangular in at least one privileged orientation. Such layers can be made either with an isotropic condition, no privileged magnetic orientation, and in such layers there is more iron than nickel; or they can be made with an anisotropic condition, then being uniaxially magnetically orientated along a privileged direction, i.e. an easy magnetization axis and in such layers there is more nickel than iron.
Binary date information stores can be made by adding to such thin layer ferromagnetic structures appropriate electrical conductor arrays enabling the operations of read-in and read-out of the information bits in and from so-called "memory points" as defined by the crossing places of the conductors of such arrays with, in some cases, additional definition from an appropriate cutting of the layers in strips and/or small areas underlying said crossing places.
In each and any of such stores the same principle of representation of the information bits applies: conventionally, the binary digital value 1 is affected to one magnetization condition of a memory point, as defined by a first local orientation of said magnetization condition, and the binary digital value 0 is affected to another condition of magnetization, usually characterized from a different orientation of the magnetization than the first one.
The magnetization conditions of the memory points must be stable with time in order to avoid deterioration of the content of the store, independantly of any further read-in and/or read-out operation at other memory points in the store. It is known that each read-in point gives rise to a demagnetizing field which is, at first approximation, substantially inversely proportional to the length of the memory point and substantially proportional to the thickness of the thin layer structure at this point. Such a demagnetizing field must firstly be lower than the coercive field of the magnetic material. The higher the value of the coercive field, the higher the demagnetizing field value at a memory point can be without appreciable drawback; hence, the higher the coercive field, the higher the density of information bits can be since each memory point can be made of a lesser area. Secondly, the demagnetizing field must not disturb the magnetization conditions of the neighbouring memory points and, in this respect too, the higher the coercive field value the lower will be the risks of deleterious interactions from memory point to memory point. Further, the higher is the value of the coercive field, the lesser the risks of disturbing the magnetizations of the memory points from stray or parasitic fields which are temporarily generated from read-in and read-out electrical currents for other memory point operations.
Referring now more particularly to magnetic structures made of thin anisotrope magnetic layers, wherein the memory points are characterized by locally saturated areas in the one and/or the contrary diraction in the orientation of the easy magnetization axis of the magnetic material, a known phenomenon is the so-called "creep" effect which, from repeated actions from parasitic fields of any kinds, slowly varies the magnetization conditions of the memory points from displacement of the walls separating the magnetization regions of these points, which obviously entails a deterioration of the content of the store. The creep facility is related to the value of the coercive field of the ferromagnetic material and the higher said value, the lower the creep.
It is the object of the invention to provide a thin layer magnetic structure comprising at least one ferromagnetic character layer consisting of at least iron and nickel, which is specially adapted for acting as a storing member in a binary data information store as herein-above defined in that it presents a value of the apparent coercive field much higher than the usual value of conventional thin layer magnetic structures.
According to a feature of the invention, such a thin layer magnetic structure is characterized in that it includes, on one face of a thin ferromagnetic character layer, a lamina of a material having a rigid magnetic lattice which is randomly inorganized at the microscopic scale.
According to a further feature of the invention, said lamina is comprised of an antiferromagnetic character material.
According to a further feature of the invention, said lamina is formed from part of a ferromagnetic layer wherein, in addition to iron and nickel, a material is alloyed which gives to said part the said antiferromagnetic character.
LIST OF FIGURES AND SHORT DESCRIPTION THEREOF
In order to illustrate these and further features of the invention, reference is made to the accompanying drawings, wherein:
FIG. 1 shows a thin layer magnetic structure in a first example of embodiment of the invention, comprising a single ferromagnetic layer 2 formed over and linked to a lamina 1 of a material having a rigid magnetic lattice of inorganized condition at the microscopic scale, said lamina being carried on any appropriate non magnetic substrate 5. As such a structure may be used in a store with or without magnetic orientation along an easy magnetization axis, the direction of said axis A is only shown in dot line at 6 in said FIG. 1.
FIG. 2 shows a second example of embodiment comprising two ferromagnetic layers 2 and 3, which may be or not of the same magnetic material and may be or not of identical thickness, said layers being coupled by means of an intercalated film 4 of a non magnetic material; said film 4 is preferably conducting. The remaining part of FIG. 2 is identical to the structure of FIG. 1.
A ferromagnetic two coupled layer structure is described in French Pat. No. 1,383,012 filed Oct. 18, 1963 by CENTRE NATIONAL DE LA RECHERCHE SCIENTIFIQUE. As known, the coercive field of such coupled layers is lower than the coercive field of each of the layers and consequently, the present invention will find an advantageous use in such structures since, from another point of view, the use of coupled layer structures is of advantage in magnetic stores as the coupling effect reduces to a very small value their response to external parasitic fields.
As coupled magnetic layer structures are specially adapted to uniaxial magnetization conditions, the direction of the easy magnetization axis 6 is shown in full line in FIG. 2.
The term "thin film" as used in the present specification must be understood as referring to thicknesses of layers from about 100 to some thousands of Angstroms, a very thin layer or film then being of a relatively appreciably smaller thickness; Illustratively, and for instance, when in FIG. 2 (or FIG. 1), the magnetic layer 2 is of about 800 A, the layer 3, of about 400 A, the layer 4 may be of about 80 A thick.
DETAILED DESCRIPTION:
In both FIGS. 1 and 2, the material of the lamina 1 must be such that its configuration or pattern of spins, which are randomly orientated at the microscopic scale of the crystallites, in the superficial lattice contacting the ferromagnetic layer 2 at least, be stable or rigid and unaffected by the action of any kind of external magnetic fields. Further, the energy of interaction between the spins of the lamina 1 and the spins of the ferromagnetic layer 2 at the contact region must be sufficiently low not to essentially disturb the intrinsic magnetic properties of of the layer 2, i.e. its capabilities of response to read-in binary digits or bits from local modifications of its magnetization. For uniaxial layers, further, the said energy of interaction must not unduly increase the dispersion (angular dispersion) of the components of the easy magnetization axis at the memory points, and must not unduly disturb the quasi-coherent rotation of the magnetization towards the difficult direction, perpendicular to the easy axis, as said rotation is conventionally used, as well known per se, for read-in and read-out operations in anisotrope thin layers.
As known an anti-ferromagnetic material, when used at a lower temperature than its order/disorder transition temperature, T N as conventionally denoted, preserves a rigid magnetic lattice which is not modified by application of any kind of magnetic field. It is consequently of advantage to constitute the lamina 1 in such an anti-ferromagnetic material, as the magnetic stores will always be operated at temperatures lower than T N , and consequently the random organization of the magnetic spins will not vary in such a material for such stores.
After the thin layer structure is made according to the invention, which will be later hereinbelow described, an important increase of the coercive field is obtained: for the single ferromagnetic layer structure shown in FIG. 1, the coercive field is at least twice the value of the coercive field of the ferromagnetic layer alone and, for the coupled layer structure of FIG. 2, the value of the coercive field is as high as 8 times the value such a coupled layer structure had without application of the invention.
Such important increases of the apparent value of the coercive field may be easily understood:- as currently known, the expression of the coercive field includes, inter alia, a term relating to the "exchange." The application or combination of the lamina 1 of antiferromagnetic material and the layer 2 of ferromagnetic material, the said lamina 1 having inorganized though stable condition at the microscopic scale of its crystallites, produces an increase of said "exchange" term by reason of the spin to spin coupling between the ferromagnetic and antiferromagnetic zones in contact, which coupling acts for opposing magnetization displacements in the ferromagnetic material.
If the lamina 1 did not present such a random condition of its spins, such a coupling would be of unidirectional character and the magnetization in the ferromagnetic layer would depend upon that of the lamina. As, on the other hand, the magnetic lattice of the material of the lamina 1 is randomly inorganized, everything is such that the exchange coupling finally gives or results in an anisotrope interaction at the macroscopic scale since the global or overall effect is the same for any portion one may consider in the ferromagnetic layer 2, and consequently for any volume comprising a memory point in said layer. It remains possible to control the orientation of the material at each memory point in said ferromagnetic layer from application of magnetic fields which, at any temperature lower than T N of the antiferromagnetic material, do not affect the random inorganization of the magnetic lattice of spins in the lamina 1.
The increase of the value of the coercive field makes it possible, as already said, to provide a higher density of the memory points whilst practically eliminating deteriorations of the content of the store from the action of the demagnetizing fields from said points and of the parasitic fields, either purely external or operational.
There are several ways of making the above-described thin layer magnetic structures:
Ferromagnetic materials are well known in themselves, and for instance and mainly, such materials comprise the nickel-iron and nickel-iron-cobalt alloys. Similarly, antiferromagnetic materials are well known in themselves, for instance cobalt oxide, chromium oxide, and the ternary alloy comprised of iron, nickel and manganese. It is sensible, by reason of the similarities of component elements, to consider as interesting couples ferromagnetic/antiferromagnetic materials such as nickel-iron and nickel-iron-cobalt for the ferromagnetic material and such as nickel-iron-manganese for the antiferromagnetic material, though such a choice must not be considered as excluding others.
Generally known, a thin magnetic layer is made from simultaneous evaporation of its component elements on a non-magnetic substrate such, for instance, a glass or ceramics of high melting point, which is heated at an appropriate temperature during deposition of the material, the elements of which are arranged in crucibles for a controlled evaporation thereof thus controlling a required proportion of such elements in the material formed on the substrate.
It is further known that, for obtaining a ternary alloy comprised of nickel, iron and manganese, a method may be used which consists of firstly depositing on the substrate a layer of manganese, secondly depositing a layer of a nickel-iron layer over the manganese layer, and thereafter annealing the product for producing thermal diffusion of the manganese within the adjacent part of the nickel-iron layer. Controlling the operation for only a partial diffusion of the manganese element within the ferromagnetic layer of nickel and iron (it would be the same with a layer of nickel-iron-cobalt) enables the attainment of a structure according to the present invention.
Illustratively considering such a method, examples may be detailed as follows:
EXAMPLE 1:
The substrate is heated at a temperature of the order from 300° to 350° C. A D.C. orientating magnetic field is applied to the location of said substrate. Evaporation of a layer of manganese up to, for instance, about 100 Angstroms is conventionally made. Thereafter, evaporation is made in conventional fashion of a layer comprised of nickel, iron and cobalt with, illustratively, about 20% cobalt, and, in the remaining 80%, a weight proportion of 81 to 19 of nickel with respect to iron, up to a thickness of the order of about 800 A.
Without changing such conditions of substrate temperature and D.C. field application (the value of said field being, for instance, about 20 Oe), the product is annealed from a period of 10 to 15 minutes. The manganese diffuses from thermal diffusion within the ferromagnetic layer so that in the lower part of said last named layer, adjacent the substrate, a lamina of antiferromagnetic material is obtained. Obviously, the concentration in manganese in said lamina varies with the thickness, decreasing from the substrate surface but this has no deteriorating action for the sought result since, in any transition, coupling is obtained between the spins of an antiferromagnetic material and the spins of a ferromagnetic material at a certain level of the structural complex which is obtained.
When the coercive field of the final structure is measured, it is seen that it is of the order of 7 to 7.3 Oersteds with the above defined conditions. In the same conditions, but without manganese, a ferromagnetic layer was obtained, which layer presents a measured value by 3.6 Oersteds only of coercive field. Consequently, it can be stated that a structure according to the invention presents a coercive field of about twice the value of that of a ferromagnetic layer.
Further, however, the antiferromagnetic lamina thus obtained in the structure reveals a satisfactorily inorganized random distribution of its spins so that there is no unidirectional coupling between it and the ferromagnetic material in the structure, which latter constitutes the storing member proper.
EXAMPLE 2:
The steps are the same as in Example 1, however the annealing step is maintained during about 1 hour. It is then seen that the antiferromagnetic lamina presents a unidirectional coupling with the ferromagnetic layer so that the structure cannot be used as such for a magnetic store of high coercive field storing member.
In order to destroy such unidirectional coupling, an alternating magnetic field (or even better, a rotating magnetic field resulting from application of two alternating fields in phase quadrature) is substituted to the D.C. field and annealing is carried out for a short length of time. The coupling proved to be destroyed and the results finally similar to those in Example 1.
It must be noted that, as a variation, the D.C. field is applied only during depositions of manganese and ferromagnetic alloy and annealing is solely made in presence of an A.C. or rotating magnetic field, during a length of time of about one quarter of an hour. Results are then satisfactory and substantially as herein before indicated.
EXAMPLE 3:
Coupled layer structure is sought. The first operations are made according to Example 1 or Example 2. Thereafter, evaporation of a very thin film of such a metal as molybdenum, of the order of 80 A in thickness for instance, is ensured over the ferromagnetic layer at a lower temperature of the substrate (preferably lower than 300° C.) Thereafter a further layer of ferromagnetic material is evaporated over the molybdenum film: for instance a layer of nickel-iron alloy with 81 to 19% ratio of components, without any cobalt, is evaporated up to a thickness of the order of 400 A.
It is seen that the antiferromagnetic lamina and the first ferromagnetic layer remain with no unidirectional coupling though the D.C. orientating field had been re-applied for the evaporation of the second ferromagnetic layer, by reason of the lower operative temperature. The measured coercive field is of the order of 4 Oersteds. However, with a similarly made coupled layer structure, omitting the manganese layer at the first step, hence omitting the antiferromagnetic lamina, the coercive field is only of a value substantially equal to 0.5 oersted. Consequently the structure according to the invention presents a value of coercive field increased by a coefficient 8 with respect to the corresponding conventional structure.
Obviously the above examples are not limitative in themselves. For instance, cobalt may be omitted for single ferromagnetic layer structures, as cobalt is thought of interest mainly with coupled layer structures; the nickel-iron ratios may be varied when the structure is not to have a negligible magnetostrictive coefficient; relative thickness can be varied at will, as being only indicatively given and such thicknesses chosen in accordance with considerations of the stores wherein the structures will be used.
Obviously, further, the superposition of two independant layers, antiferromagnetic and ferromagnetic made from successive evaporations on the substrate and without interpenetration thereof remains within the field of the invention.