JP2000082283A - Magnetic storage device and addressing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a magnetic storage device having an array of magnetic materials as a storage carrier, while solving problems associated with writing using a magnetic field, such as generation of crosstalk and reduction of coercive force due to miniaturization, while integrating. An addressing function indispensable for a circuit element is realized. SOLUTION: An arbitrary storage carrier which has a structure in which a coupling control layer is sandwiched between two magnetic layers, and exchange exchange acting between two magnetic layers via the coupling control layer is selected as a target of writing or reading. And use it as a means to achieve the desired operation. That is, when writing or reading is performed by selecting an arbitrary storage carrier, a change in the exchange interaction between the two magnetic layers caused by applying a stimulus such as an electrical stimulus or a photostimulus to the coupling control layer is used. I do.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a magnetic storage device having an array of a plurality of separated magnetic materials as a storage carrier,
And an addressing method for such a magnetic storage device.

[0002]

2. Description of the Related Art An element utilizing a magnetic material is attractive in two points as compared with a semiconductor device. First,
Since a conductive metal can be used as an element of the element, a high carrier density and a low resistance value can be realized. Therefore, an element using a magnetic material is expected to be suitable for miniaturization and high integration. Second, there is a possibility that the bistability of the magnetization direction of the magnetic material can be used for a nonvolatile memory. That is, if the bistability of the magnetization direction of the magnetic material is used, it is expected that a solid-state nonvolatile memory in which stored information is not lost even if the power of the circuit is turned off.

A solid-state non-volatile memory in which stored information is not lost even when the power of a circuit is turned off is expected to be put to practical use in various fields as an ultimate power-saving memory. Specifically, for example, a solid-state nonvolatile memory consumes no power when inactive, and is therefore expected as a key technology for reducing the capacity and weight of a battery in a portable electronic information device or the like. Also, with the rise of the satellite media business, the demand for solid-state nonvolatile memories is high as that for supporting activities of satellites in the shadow of the earth where solar cells cannot be used.

[0004] Elements using a magnetic material include:
(1) non-volatility, (2) no degradation due to repetition, (3) high-speed writing is possible,
There are advantages such as (4) suitable for miniaturization and high density, and (5) excellent radiation resistance. Less than,
These advantages will be described.

(1) As in the case of a magnetic recording medium such as a magnetic tape or a magnetic disk having non-volatility, the magnetic material itself has a bistability (bist
information), the information written as the magnetization direction is maintained even when the driving force is lost.

(2) No Deterioration Due to Repetition A ferroelectric memory (F-RAM: Ferroelectric Random Access Memory) using a ferroelectric material exhibiting bistability like a magnetic material
Have also been proposed as candidates for solid-state nonvolatile memories.
In the F-RAM, the memory state is rewritten by reversing the spontaneous dielectric polarization. However, the inversion of spontaneous dielectric polarization corresponding to the rewriting of the memory state
Since ion migration in the crystal lattice is involved, if rewriting is repeated over a million times, crystal defects develop. For this reason, in the F-RAM, there is a problem that the element life cannot be exceeded due to the fatigue of the material. On the other hand, since the magnetization reversal of the magnetic material does not involve ion movement or the like, the element using the magnetic material is not limited to the fatigue of the material,
Rewriting can be repeated almost infinitely.

(3) High-speed writing is possible The speed of magnetization reversal of the magnetic material is very fast, about 1 ns or less. Therefore, high-speed writing can be performed by utilizing the high switching speed.

(4) Magnetic alloys suitable for miniaturization and high density The magnetic properties of the magnetic alloy can be variously changed by selecting the composition and the structure. Therefore, in the element using the magnetic material, the degree of freedom in design is extremely high. In an element using a magnetic material, for example, a magnetic alloy having conductivity can be used. When a magnetic alloy having conductivity is used, the current density in the device can be increased as compared with the case where a semiconductor is used, so that it is possible to further reduce the size and density of the semiconductor device. As an element utilizing such characteristics, for example, a spin transistor has been proposed as described in the Journal of the Japan Society of Applied Magnetics, Vol. 19, 684 (1995).

(5) Excellent radiation resistance In an element, such as a D-RAM (Dynamic Random Access Memory), in which a memory state is created by charging an electric capacity, discharge occurs when ionizing radiation passes through the element. , The memory information is lost. On the other hand, since the magnetization direction of the magnetic body is not disturbed by ionizing radiation, the element using the magnetic body has excellent radiation resistance. Therefore, an element using a magnetic material is particularly useful in applications requiring high radiation resistance, such as a communication satellite. Actually, a magnetic bubble memory, which is one of the memories using a magnetic material, has already been used as a memory mounted on a communication satellite and has many achievements.

[0010] As described above, the element using the magnetic material has various advantages, and a solid-state magnetic memory has been devised as a device utilizing these advantages. A solid-state magnetic memory is a magnetic storage device using an array of magnetic materials as a storage carrier, and performs a storage operation without moving the storage carrier, unlike a magnetic tape or a magnetic disk. In a conventional solid-state magnetic memory, a simple addressing method is adopted by utilizing the characteristics of a magnetic material. Hereinafter, an addressing method in a conventional solid-state magnetic memory will be described.

In a solid-state magnetic memory, a magnetic thin film having uniaxial magnetic anisotropy is usually used as a storage carrier. The magnitude of a magnetic field required to cause the magnetization reversal of the magnetic thin film depends on the application of the magnetic field. Depends on direction. That is, when a magnetic field is applied in a direction at an angle of about 45 ° from the easy axis of magnetization, magnetization reversal can be generated with a smaller magnetic field intensity than when a magnetic field is applied in parallel to the direction of the easy axis of magnetization. it can. In a conventional solid-state magnetic memory, by utilizing such properties for addressing recording bits, it is possible to employ a very simple addressing method.

That is, in the conventional solid-state magnetic memory, as shown in FIG. 23, word lines W1, W2, W3,
, And bit lines B1, B2, B3,... Are arranged so as to be orthogonal to each other, and storage carriers A-1, A-2,. B-2, ...
., C-1, C-2,... Thus, in the conventional solid-state magnetic memory, the storage carrier is x,
The memory chips are configured in a y matrix arrangement.
Note that the axis of easy magnetization of the storage carrier is aligned with the direction of the word line.

For example, when one word line W2 and one bit line B1 are selected and an appropriate current is supplied to both at the same time, the storage carrier B-1 at the intersection of these two lines is selected.
Only the magnetization reversal occurs. At this time, both the word line W2 and the bit line B1 to which the current is supplied apply a magnetic field to a plurality of storage carriers arranged along the word line W2 and the bit line B1. The magnetic field from either one alone is not enough to cause magnetization reversal. Then, only when the magnetic field H _W from the word line W2 and the magnetic field H _B from the bit line B1 are combined and the magnetic field applied to the storage carrier becomes a magnetic field in a 45 ° direction from the axis of easy magnetization (ie, In the example of FIG. 23, only the storage carrier B-1) causes magnetization reversal. That is, in the conventional solid-state magnetic memory, the fact that the magnetization reversal occurs in the storage carrier only when the magnetic field applied to the storage carrier is a magnetic field in the direction of 45 ° from the axis of easy magnetization is used for selecting a specific storage carrier. That is.

As described above, in the conventional solid-state magnetic memory, it is possible to select a specific storage carrier and cause magnetization reversal only by a very simple tool of intersecting conductive wires. In addition, a simple address system can be adopted.

[0015]

However, the conventional solid-state magnetic memory has a problem that crosstalk occurs with miniaturization. That is, in the conventional solid-state magnetic memory, writing to the memory is performed by applying a magnetic field, but the magnetic field is a long-distance force. Has a non-negligible effect, and crosstalk occurs. To prevent this, for example, ZGWang, et al., IEE
E Trans Magn., Mag33, 4498 (1997) also reports a design example of a memory cell having a magnetic field shielding structure, but has a drawback that the structure becomes complicated.

Further, the conventional solid-state magnetic memory also has a problem that the coercive force decreases with miniaturization. In a conventional solid-state magnetic memory, the generation of a write magnetic field depends on the current, but there is a limit determined by the material on the current density that can be carried by the conductor. Therefore, as the design rule becomes finer and the diameter of the conductive wire becomes smaller, the upper limit of the available current decreases. Therefore, the maximum magnetic field that can be used generally decreases in proportion to the design rule. on the other hand,
The coercive force of the storage carrier must be designed such that a magnetization reversal is achieved with an externally applied magnetic field. For this reason, when the magnetic field that can be applied to the storage carrier decreases with miniaturization, the coercive force of the storage carrier needs to be reduced accordingly. That is, in the conventional solid-state magnetic memory, it is necessary to reduce the coercive force of the storage carrier with miniaturization.
However, if the coercive force of the storage carrier is too small, the reliability is reduced. This is a serious problem especially for a memory for a portable electronic device which is often used in an environment where a disturbance magnetic field is applied from the surroundings.

Each of these problems is caused by applying a magnetic field to a storage carrier for writing in a conventional solid-state magnetic memory. Therefore, in order to solve these problems, it is necessary to reconsider an addressing method for achieving an intended operation by designating an arbitrary storage medium selected as a target of writing or reading.

The present invention has been proposed in view of the above-described conventional circumstances. In a magnetic storage device having an array of magnetic materials as a storage carrier, generation of crosstalk and coercive force due to miniaturization are considered. It is an object of the present invention to realize an addressing function that is indispensable for an integrated circuit element, while solving a problem associated with magnetic field writing, such as deterioration.

[0019]

A magnetic storage device according to the present invention is a magnetic storage device having an array of a plurality of separated magnetic materials as a storage carrier, wherein an arbitrary one selected as an object to be written or read is provided. The present invention is characterized by utilizing exchange interaction propagating in a solid as means for designating a memory carrier and achieving a desired operation.

In the magnetic storage device according to the present invention, when utilizing the exchange interaction propagating in a solid, for example, a structure in which a coupling control layer is sandwiched between two magnetic layers is used. The exchange interaction used at this time is an exchange interaction acting between the two magnetic layers via the coupling control layer.
Then, when writing or reading is performed by selecting an arbitrary storage carrier, a change in the exchange interaction between the two magnetic layers caused by applying a stimulus such as an electrical stimulus or a light stimulus to the coupling control layer is performed. Use.

The above-mentioned bonding control layer is, for example,
A semiconductor layer is used. At this time, the exchange interaction is mediated by conduction electrons of the semiconductor layer. When writing or reading is performed by selecting an arbitrary storage carrier, a change in the exchange interaction between the two magnetic layers caused by applying an electrical stimulus to the semiconductor layer is used.

Further, as the binding control layer, for example,
A dielectric layer may be used. At this time, the exchange interaction is mediated by electrons moving between the magnetic layers by the tunnel effect via the dielectric layer. And, when performing writing or reading by selecting an arbitrary storage carrier,
The change of the exchange interaction between the two magnetic layers caused by changing the tunnel barrier height of the dielectric layer is used.

Further, as the above-mentioned bonding control layer, for example,
A conductor layer may be used. The exchange interaction used at this time is an exchange interaction acting between the two magnetic layers via the conductor layer. When writing or reading is performed by selecting an arbitrary storage carrier, a change in the exchange interaction between the two magnetic layers caused by passing a current through the conductive layer is used.

Further, as the binding control layer, for example,
A material having a film thickness of 10 nm or more and containing a magnetic material may be used. Specifically, a multilayer structure in which a magnetic layer and a non-ferromagnetic layer are laminated, a layer in which magnetic particles are dispersed in a non-magnetic material, and the like are preferable.

In the above magnetic storage device, a magnetic layer made of a hard magnetic material may be formed below the structure in which the coupling control layer is sandwiched between the two magnetic layers. Further, as the magnetic layer sandwiching the coupling control layer, a layer in which a pair of magnetic layers are stacked via an intermediate layer so that the magnetization directions are antiparallel to each other may be used. In addition, a thin film made of an electrically insulating material that mediates magnetic coupling may be disposed between the magnetic layer and the coupling control layer.

In the magnetic storage device according to the present invention, for example, a plurality of linear members are arranged so as to intersect, and individual storage carriers are arranged at positions corresponding to intersections of the linear members. You. Then, when writing or reading is performed by selecting an arbitrary storage carrier, the magnetic interaction exerted on the storage carrier from two or more linear members is combined to perform the writing or reading operation on the selected storage carrier. Do. Then, as at least one of the magnetic interactions, an exchange interaction propagating in a solid is used.

In the magnetic storage device according to the present invention, for example, a plurality of linear members are arranged so as to intersect, and individual storage carriers are arranged at positions corresponding to intersections of the linear members. You. Then, when an arbitrary storage medium is selected and a write operation is performed, the magnetization direction of one storage medium is controlled by a combination of magnetic interactions exerted on the storage medium by three or more linear members. Then, as at least one of the magnetic interactions, an exchange interaction propagating in a solid is used.

An addressing method according to the present invention is an addressing method in a magnetic storage device having an array of a plurality of separated magnetic materials as a storage carrier, wherein an arbitrary storage carrier is selected to perform writing or reading. When performing, it is characterized by utilizing an exchange interaction propagating in a solid.

In the addressing method according to the present invention, when utilizing the exchange interaction propagating in a solid, for example, a structure in which a coupling control layer is sandwiched between two magnetic layers is used. The exchange interaction used at this time is an exchange interaction acting between the two magnetic layers via the coupling control layer. Then, when writing or reading is performed by selecting an arbitrary storage carrier, a change in the exchange interaction between the two magnetic layers caused by applying a stimulus such as an electrical stimulus or a light stimulus to the coupling control layer is performed. Use.

As the above-mentioned bonding control layer, for example,
A semiconductor layer is used. At this time, the exchange interaction is mediated by conduction electrons of the semiconductor layer. When writing or reading is performed by selecting an arbitrary storage carrier, a change in the exchange interaction between the two magnetic layers caused by applying an electrical stimulus to the semiconductor layer is used.

Further, as the binding control layer, for example,
A dielectric layer may be used. At this time, the exchange interaction is mediated by electrons moving between the magnetic layers by the tunnel effect via the dielectric layer. And, when performing writing or reading by selecting an arbitrary storage carrier,
The change of the exchange interaction between the two magnetic layers caused by changing the tunnel barrier height of the dielectric layer is used.

Further, as the binding control layer, for example,
A conductor layer may be used. The exchange interaction used at this time is an exchange interaction acting between the two magnetic layers via the conductor layer. When writing or reading is performed by selecting an arbitrary storage carrier, a change in the exchange interaction between the two magnetic layers caused by passing a current through the conductive layer is used.

Further, as the binding control layer, for example,
A material having a thickness of 10 nm or more and containing a magnetic material may be used. Specifically, a multilayer structure in which a magnetic layer and a non-ferromagnetic layer are laminated, a layer in which magnetic particles are dispersed in a non-magnetic material, and the like are preferable.

The magnetic storage device to which the addressing method according to the present invention is applied is, for example, arranged so that a plurality of linear members intersect with each other, and at a position corresponding to the intersection of the linear members. Individual storage carriers are arranged. Then, when writing or reading is performed by selecting an arbitrary storage medium, a magnetic interaction exerted on the storage medium from two or more linear members is combined to perform a write or read operation on the selected storage medium. I do. In the addressing method according to the present invention, as at least one of the magnetic interactions, an exchange interaction propagating in a solid is used.

The magnetic storage device to which the addressing method according to the present invention is applied is, for example, arranged so that a plurality of linear members intersect with each other and at a position corresponding to the intersection of the linear members. Individual storage carriers are arranged. When a write operation is performed by selecting an arbitrary storage carrier, the magnetization direction of one storage carrier is set to three.
It is controlled by a combination of magnetic interactions exerted on the storage carrier from one or more linear members. In the addressing method according to the present invention, as at least one of the magnetic interactions, an exchange interaction propagating in a solid is used.

[0036]

Embodiments of the present invention will be described below in detail with reference to the drawings.

1. A solid-state magnetic memory to which the present invention is applied A magnetic storage device according to the present invention is a magnetic storage device having an array of a plurality of separated magnetic materials as a storage carrier,
This is a so-called solid magnetic memory. In the solid-state magnetic memory to which the present invention is applied, the magnetization direction of the storage carrier, which is an element element responsible for storage, is changed by a magnetic interaction (exchange interaction: exch interaction: exch interaction) in the solid without applying an external magnetic field.
Ange interaction) is controlled as driving force. In the following description, such a solid-state magnetic memory is referred to as an “exchange-coupled magnetic memory”.

The above-described problems in the conventional solid-state magnetic memory (problems such as generation of crosstalk and reduction of coercive force due to miniaturization) are caused by applying a magnetic field to the storage carrier for writing. . Therefore, in the exchange-coupling type solid-state magnetic memory in which the magnetization direction of the storage carrier is controlled using the exchange interaction, the above-described problem in the conventional solid-state magnetic memory can be solved.

It is to be noted that the exchange interaction is none other than the origin of aligning the magnetic moments of atoms in one direction inside the ferromagnetic material. Further, as shown in FIG. 1, when the magnetic body A and the magnetic body B are in contact with each other, an exchange interaction acts between the two through the interface S in contact. Further, as shown in FIG. 2, the magnetic body A and the magnetic body B are not in direct contact,
Even if the intermediate layer C exists between the magnetic material A and the magnetic material B, the exchange interaction between the magnetic material A and the magnetic material B is
It may propagate through the intermediate layer C. Here, when the intermediate layer C is a magnetic material, the exchange interaction naturally propagates, but the intermediate layer C is a magnetic material that does not exhibit magnetism by itself.
It has been confirmed that even in the case of a nonmagnetic metal such as u or Au or a semiconductor such as Si or Ge, the exchange interaction propagates through the intermediate layer C. A theory (such as the RKKY model) that explains the origin of such exchange interaction propagation has also been proposed.

[0040] 2. Example of exchange-coupled solid-state magnetic memory An example of an exchange-coupled solid-state magnetic memory is shown in FIG. In addition,
The exchange-coupled solid-state magnetic memory 1 shown in FIG. 3 is a memory that can be written only once by controlling the magnetization direction of a storage carrier 2 made of a magnetic material.

In this exchange-coupled solid-state magnetic memory 1, the storage carrier 2 is sandwiched between two fixed magnetic layers 3 and 4 for giving a reverse bias. The current supplied to the coupling control layer 6 by the input circuit 5 has the function of cutting off the bias from the fixed magnetic layer 3 to the storage carrier 2.

That is, the input to the coupling control layer 6 is OFF.
When a current does not flow through the coupling control layer 6, an exchange interaction occurs between the fixed magnetic layer 3 and the storage carrier 2, and a bias from the fixed magnetic layer 3 acts on the storage carrier 2.
When there is a bias from the fixed magnetic layer 3,
The bias from the fixed magnetic layer 3 and the bias from the fixed magnetic layer 4 cancel each other, and no net driving force acts on the magnetization of the storage carrier 2.

On the other hand, when the input to the coupling control layer 6 is turned ON and a current flows through the coupling control layer 6, the exchange interaction between the fixed magnetic layer 3 and the storage medium 2 does not occur. Bias from the fixed magnetic layer 3 does not work. Then, when the bias from the fixed magnetic layer 3 does not work, the magnetization from the storage carrier 2 is caused by the bias from the fixed magnetic layer 4.

The reading of the magnetization direction of the storage carrier 2 is achieved by the principle of the spin valve in the example of FIG. The fixed magnetic layer 7 is separated by a non-magnetic intermediate layer 8 so as not to have a strong influence on the magnetization of the storage carrier 2. The current supplied from the output circuit 9 and flowing from the fixed magnetic layer 7 to the storage carrier 2 via the non-magnetic intermediate layer 8 is generated when the magnetization direction of the storage carrier 2 and the magnetization direction of the fixed magnetic layer 7 are parallel. It becomes larger and smaller in the case of anti-parallel,
Thereby, the magnetization direction of the storage carrier 2 can be detected.

In such an exchange-coupled solid-state magnetic memory 1, since exchange coupling is used to control the magnetization of the storage carrier 2, magnetic field-based writing such as generation of crosstalk and reduction of coercive force due to miniaturization is performed. Can be solved.

Here, the relationship between the dimension L of a storage unit portion (hereinafter, referred to as a memory cell) of a solid-state magnetic memory and a drive magnetic field H that can be used to drive a storage carrier is explained by applying a current to a conductor. In the case of the current magnetic field system using the magnetic field generated in (1), that is, in the case of the conventional solid-state magnetic memory shown in FIG. FIG. 4 shows a comparison with such an exchange-coupled solid-state magnetic memory). In the current magnetic field method, the diameter of the conductor was assumed to be 0.8 times the memory cell size L. As shown in FIG. 4, in the current magnetic field method, as the memory cell dimension L decreases, the magnetic field that can be applied from the conductor decreases. On the other hand, since the exchange interaction does not depend on the memory cell size L, the exchange coupling method becomes more advantageous as miniaturization progresses.

As described above, the magnetic field conversion value of the exchange interaction does not depend on the memory cell size L. Therefore, if the magnetization of the storage carrier is controlled using the exchange interaction, even if the miniaturization proceeds, A magnetic material having a large coercive force can be used as a storage carrier. Specifically, as can be seen from FIG. 4, even if the memory cell size L becomes very small, a magnetic material having a coercive force of several tens Oe or more can be used for the storage carrier. By using a memory carrier having a large coercive force, for example, high reliability is guaranteed even for a portable electronic device used in an environment where a disturbance magnetic field is received from the surroundings.

[0048] 3. Addressing of Solid-State Magnetic Memory In the present invention, an addressing function indispensable for an integrated circuit element is further added to the above-described exchange-coupled solid-state magnetic memory.

In general, a writing process for a solid-state magnetic memory having a plurality of memory cells comprises the following series of operations. That is, first, an arithmetic processing device using a solid-state magnetic memory selects a memory cell to be written. Next, from the arithmetic processing unit to the solid-state magnetic memory, information indicating that the magnetization reversal of the storage carrier of the memory cell should be performed toward a target memory cell among a large number of memory cells in the solid-state magnetic memory. Is sent. next,
Based on the information, a driving force for magnetization reversal is exerted on the storage carrier of the corresponding memory cell, and the magnetization of the storage carrier is reversed. Then, selecting a specific memory cell and causing the memory cell to perform a specific operation in this way is generally called addressing.

For example, in a solid-state magnetic memory in which the magnetization direction of a storage carrier is controlled by an electric input, in order to realize an addressing function, a wiring for carrying an electric signal from an arithmetic processing unit to a memory cell (so-called wiring). Address line). That is, an address line is provided for each memory cell, and an electric signal is sent to the address line corresponding to the memory cell to be operated, so that a specific memory cell can be selected and the memory cell can be operated. .

However, if address lines are individually provided for individual memory cells, the structure becomes very complicated. For example, when there are m memory cells in the vertical direction and n memory cells in the horizontal direction, if an individual address line is to be provided for each memory cell, only one address line is provided for one memory cell. In this case, m × n address lines are required. In this case, the structure becomes very complicated, and it is difficult to form an integrated circuit device.

On the other hand, in the conventional solid-state magnetic memory shown in FIG. 23, an addressing function is realized only by a very simple tool of a group of intersecting conductors. That is, in the conventional solid-state magnetic memory shown in FIG. 23, m + vertical wirings and n + horizontal wirings for a total of n + m wirings are provided for m memory cells in the vertical direction and n memory cells in the horizontal direction. Only by providing the memory cell, it is possible to select a specific memory cell and operate the memory cell.

In the following description, the addressing performed by using the group of intersecting conductors is referred to as "matrix type addressing". Such a matrix type addressing is very suitable especially for forming an integrated circuit device because the number of necessary wirings is very small and the structure is simple even if the number of memory cells is increased.

By the way, in the conventional solid-state magnetic memory shown in FIG. 23, since the magnetization reversal of the storage carrier is performed by using the superposition of the magnetic fields, matrix type addressing can be easily realized. However, in the exchange-coupled solid-state magnetic memory, the superposition of magnetic fields is not used, so that application of matrix addressing is not easy.

That is, in the conventional exchange-coupled solid-state magnetic memory, in order to select a specific memory cell and operate the memory cell, only the specific memory cell is selected and a current or a voltage is supplied. A mechanism is required, and therefore, matrix-type addressing cannot be simply applied to the exchange-coupled solid-state magnetic memory that has been conventionally devised. In other words, in the conventional exchange-coupled solid-state magnetic memory, if matrix-type addressing is forcibly applied, some contrivance is required beyond simply connecting the address line and the memory cell. . Specifically, for example, it is necessary to connect a non-linear element such as a diode between an address line and a memory cell, or to attach a select transistor used in a semiconductor memory to the memory cell. . However, these are not preferred because they lead to a complicated and complicated structure.

[0056] 4. Exchange coupled solid state magnet to which the present invention is applied
Basic Structure of Memory As described above, in realizing matrix-type addressing in an exchange-coupled solid-state magnetic memory, using a non-linear element, a selection transistor, or the like is not preferable because the structure becomes complicated and complicated. . Therefore, in the present invention, matrix addressing can be realized in an exchange-coupled solid-state magnetic memory without using a nonlinear element, a selection transistor, or the like. Less than,
A basic configuration of an example of such an exchange-coupled solid-state magnetic memory to which the present invention is applied will be described.

4-1 Overall Configuration First, the exchange-coupled solid-state magnetic memory is provided with a plurality of linear members formed in an elongated linear or band shape. The linear members are provided with a function as a signal transmission line for specifying a memory cell and a function of controlling the magnetization direction of a storage carrier in the memory cell. In the following description, such a linear member is referred to as a “drive line”.

More specifically, for example, assuming that the directions orthogonal to each other are the x direction and the y direction, respectively, a plurality of drive lines (hereinafter, referred to as "drive lines") arranged substantially in parallel to the x direction.
These drive lines are called x-direction drive lines. ), And a plurality of drive lines arranged substantially parallel to the y direction (hereinafter, these drive lines are referred to as y direction drive lines). Then, the storage carrier is arranged at a position of a lattice point formed by the intersection of the x-direction drive line and the y-direction drive line.

These drive lines are used for all of a plurality of storage carriers arranged along the drive lines, as in the case of the word lines and bit lines in the conventional solid-state magnetic memory shown in FIG. It acts to change the magnetization direction of the carrier. However, here, since it is an exchange-coupling type solid-state magnetic memory, an exchange interaction is used as an operation for changing the magnetization direction of the storage carrier. In the following description, an operation of causing the magnetization direction of the storage carrier to be directed in a certain direction is referred to as a driving operation.

4-2 Source of Matrix Type Addressing
This is achieved by combining drive lines as described above.
Matrix addressing in an exchange-coupled solid-state magnetic memory will be described.

4-2-1 Configuration of Memory Cell First, a memory cell of an exchange-coupled solid-state magnetic memory in which matrix addressing is performed is described with reference to FIGS.
This will be described with reference to FIG. FIG. 5 is a diagram showing a portion corresponding to one memory cell, and FIG. 6 is a diagram for explaining the driving principle.

As shown in FIG. 5, the memory cell 10 includes a first y-direction drive line 11 and a second y-direction drive line 12 arranged in parallel with each other, and a first and second y-direction drive line. 1
An x-direction drive line 14 disposed orthogonal to the first and second x-direction drive lines 14 and a storage carrier 13 disposed between the first and second y-direction drive lines 11 and 12 and the x-direction drive line 14. Consists of Here, the storage carrier 13 is placed under the influence of the first y-direction drive line 11, the second y-direction drive line 12 and the x-direction drive line 14. That is, the storage carrier 13 is placed under the influence of three driving sources.

In the memory cell 10, the first y-direction drive line 11 is formed by forming a laminate of a first fixed magnetic layer 11 a magnetized in a predetermined direction and a first conductor layer 11 b into an elongated strip shape. It is formed in. The second y-direction drive line 12 includes a second pinned magnetic layer 12a magnetized in a direction different from that of the first pinned magnetic layer 11a, and a second conductor layer 12a.
The laminate with b is formed in an elongated strip shape. Then, the storage carrier 13 is provided with the first y so as to face the fixed magnetic layers 11a and 12a via the conductor layers 11b and 12b.
It is formed from a part on the direction drive line 11 to a part on the second y-direction drive line 12.

Although not shown in FIG. 5, especially when the electric resistance of the storage carrier 13 is low, the first conductive layer 1
1b and the storage carrier 13, between the second conductive layer 12b and the storage carrier 13, between the storage carrier 13 and the x-direction drive line 1
It is better to form an insulating layer between the first and fourth layers.

In this memory cell 10, the first y-direction drive line 11 is a drive source that acts on the storage carrier 13 to drive a drive action A 1 to direct the magnetization direction of the storage carrier 13 in a predetermined direction. It has become. Similarly, the second
The y-direction drive line 12 is also a drive source that acts on the storage carrier 13 to actuate a drive action A2 to direct the magnetization direction of the storage carrier 13 to a predetermined direction. In addition,
5 and 6, the direction of the arrow A1 indicates the direction of the driving action acting on the storage carrier 13 from the first y-direction drive line 11, and the direction of the arrow A2 indicates the direction of the second y-direction drive. The direction of the driving action acting on the storage carrier 13 from the line 12 is shown.

That is, the first fixed magnetic layer 11a constituting the first y-direction drive line 11 is magnetized in the −x direction, and the driving action A1 from the first y-direction drive line 11 to the storage carrier 13 is performed. Works to orient the magnetization direction M1 of the storage carrier 13 in the −x direction. Further, the second pinned magnetic layer 12a constituting the second y-direction drive line 12 is magnetized in the + x direction, and the driving action A2 from the second y-direction drive line 12 to the storage carrier 13 is 13 acts to orient the magnetization direction M1 in the + x direction.

The storage carrier 13 formed from a part on the first y-direction drive line 11 to a part on the second y-direction drive line 12 is made of a magnetic material having uniaxial magnetic anisotropy. , And x directions are the easy axes of magnetization. In the memory cell 10, binary recording can be performed depending on the direction of magnetization of the storage carrier 13.

On the other hand, the x-direction drive line 14 is made of a conductive material, and
And it is formed in an elongated strip shape so that the longitudinal direction is the x direction. In the memory cell 10, a magnetic field is generated by flowing a current through the x-direction drive line 14, and the magnetic field is applied to the storage carrier 13. 5 and 6, A3 indicates a magnetic field generated when a current flows through the x-direction drive line 14.

By the way, as described in the description of the conventional solid-state magnetic memory shown in FIG. 23, the magnitude of the magnetic field required to cause the magnetization reversal in the magnetic material depends on the direction of application of the magnetic field. In general, when a magnetic field is applied in a direction at an angle of about 45 ° from the easy axis, magnetization reversal occurs with a smaller magnetic field strength than when a magnetic field is applied in parallel to the direction of the easy axis. be able to.

Therefore, in this memory cell 10, when only the driving action A 1 from the first y-direction driving line 11 or only the driving action A 2 from the second y-direction driving line 12, the storage carrier 13 To prevent magnetization reversal,
When there is both the driving action A1 from the y-direction drive line 11 and the action by the magnetic field A3 generated by flowing a current through the x-direction drive line 14, or from the second y-direction drive line layer 12. When both the driving action A2 and the action by the magnetic field A3 generated by flowing the current through the x-direction driving line 14, the magnetization reversal occurs in the storage carrier 13. That is, in this memory cell, a current flows through the first conductive layer 11b forming the first y-direction drive line 11 and a current flows through the second conductive layer 12b forms the second y-direction drive line 12. By controlling the current and the current flowing through the x-direction drive line 14, the magnetization direction M <b> 1 of the storage carrier 13 is controlled, and binary recording is performed according to the magnetization direction of the storage carrier 13.

4-2-2 Principle of Driving Memory Cell The principle of driving the memory cell 10 will be described in detail with reference to FIG.

First, FIG. 6A shows the current supply to the x-direction drive line 14, the current supply to the first conductive layer 11b constituting the first y-direction drive line 11, and the second y-direction drive line 11. This shows a state in which the magnetization direction of the storage carrier 13 is maintained in the + x direction (right direction in the figure) without supplying current to the second conductor layer 12b constituting the direction drive line 12. In this case, the first y
Since the current supply to the first conductor layer 11b constituting the direction drive line 11 and the current supply to the second conductor layer 12b constituting the second y-direction drive line 12 are not performed, the storage carrier In 13, both the driving action A1 from the first y-direction driving line 11 and the driving action A2 from the second y-direction driving line 12 act. However, the direction of the driving action A1 from the first y-direction driving line 11 and the direction of the driving action A2 from the second y-direction driving line 12 are opposite to each other.
The driving action A1 from the y-direction driving line 11 and the driving action A2 from the second y-direction driving line 12 cancel each other. Therefore, the magnetization direction M1 of the storage carrier 13 is stabilized by the uniaxial magnetic anisotropy of the storage carrier itself, and the previous state (here, the magnetization direction of the storage carrier 13 is in the + x direction) is changed. It is kept as it is.

Next, as shown in FIG. 6B, when the magnetization direction M1 of the storage carrier 13 is in the + x direction (right direction in the figure), the current supply to the x-direction drive line 14 and the first The current is not supplied to the first conductive layer 11b constituting the y-direction drive line 11, and the second conductor constituting the second y-direction drive line 12 is not supplied.
Shows a state in which a current is supplied only to the conductive layer 12b. At this time, the driving action A2 from the second y-direction driving line 12 to the storage carrier 13 does not work. On the other hand, the first y
Since the current is not supplied to the conductor layer 11b constituting the direction drive line 11, the drive action A1 from the first y-direction drive line 11 acts on the storage carrier 13. At this time, the first
The drive action A1 acting on the storage carrier 13 from the y-direction drive line 11 does not exceed the coercive force of the storage carrier 13. If the driving action A1 from the first y-direction drive line 11 does not exceed the coercive force of the storage carrier 13, then as shown in FIG.
Even if the drive action A2 from
Is maintained in the + x direction (right direction in the figure).

Next, as shown in FIG. 6C, when the magnetization direction M1 of the storage carrier 13 is in the + x direction (right direction in the figure), the first y-direction drive line 11 constituting the first y-direction drive line 11 is formed. When supplying current to the x-direction drive line 14 without supplying current to the conductor layer 11b and supplying current to the second conductor layer 12b constituting the second y-direction drive line 12, Indicates the status. At this time, the driving action A2 from the second y-direction driving line 12 to the storage carrier 13 does not work. Meanwhile, the first
No current is supplied to the conductor layer 11b constituting the y-direction drive line 11 of FIG.
The driving action A1 from the direction driving line 11 operates. Further, since the current is supplied to the x-direction drive line 14, the y-direction magnetic field A3 generated by flowing the current to the x-direction drive line 14 is generated.
Acts on the storage carrier 13.

At this time, the total vector of the drive action A1 from the first y-direction drive line 11 and the action by the magnetic field A3 generated by flowing a current through the x-direction drive line 14 is the magnetization vector of the storage carrier 13. Since the direction is shifted from the easy axis, the magnetization reversal can be caused in the storage carrier 13 with a smaller magnetic field intensity as compared with the action acting in parallel to the easy axis. Then, the storage carrier 13 thus has a driving action A1 from the first y-direction driving line 11,
When both the action by the magnetic field A3 generated by flowing the current through the x-direction drive line 14 is present, the magnetization reversal is caused to occur. As a result, as shown in FIG. 6C, the magnetization direction M1 of the storage carrier 13 is reversed from the + x direction (right direction in the drawing) to the −x direction (left direction in the drawing),
The magnetization direction M1 of No. 3 is aligned with the direction of the driving action A1 from the first y-direction driving line 11.

Thereafter, a current is supplied to the x-direction drive line 14,
And when the current supply to the second conductor layer 12b is stopped,
The magnetization direction M1 of the storage carrier 13 inverted in the x direction is kept as it is. When the current supply to the x-direction drive line 14 and the current supply to the second conductor layer 12b are stopped, the first y
Both the drive action A1 from the directional drive line 11 and the drive action A2 from the second y-direction drive line 12 are in a state where they are in operation. , Since the directions of the driving action A2 from the second y-direction driving line 12 are opposite to each other, these driving actions A1,
A2 is offset. Therefore, the magnetization direction M1 of the storage carrier 13 is stabilized by the uniaxial magnetic anisotropy of the storage carrier itself, and the state up to that point (here, the magnetization direction of the storage carrier 13 is in the −x direction). Is kept as it is.

Next, as shown in FIG. 6D, when the magnetization direction M1 of the storage carrier 13 is in the −x direction (left direction in the figure), the current supply to the x-direction drive line 14 and the second No current is supplied to the second conductive layer 12b constituting the y-direction drive line 12 of FIG.
Shows a state in which a current is supplied only to the conductive layer 11b. At this time, the driving action A1 from the first y-direction driving line 11 to the storage carrier 13 does not work. On the other hand, the second y
Since the current is not supplied to the conductor layer 12b constituting the direction drive line 12, the drive action A2 from the second y-direction drive line 12 acts on the storage carrier 13. At this time, the second
The drive action A2 acting on the storage carrier 13 from the y-direction drive line 12 does not exceed the coercive force of the storage carrier 13. If the driving action A2 from the second y-direction drive line 12 is set so as not to exceed the coercive force of the storage carrier 13, the first y-direction drive line 11 as shown in FIG.
Even if the driving action A1 from
Is maintained in the −x direction (left direction in the figure).

Next, as shown in FIG. 6E, when the magnetization direction M1 of the storage carrier 13 is in the −x direction (left direction in the figure), the second y-direction drive line 12 When the current is supplied to the x-direction drive line 14 without supplying the current to the conductor layer 12b, and the current is supplied to the first conductor layer 11b constituting the first y-direction drive line 11. The state of is shown. At this time, the driving action A1 from the first y-direction driving line 11 to the storage carrier 13 does not work. On the other hand, the second
Current is not supplied to the conductive layer 12b constituting the y-direction drive line 12 of FIG.
The driving action A2 from the direction driving line 12 operates. Further, since the current is supplied to the x-direction drive line 14, the y-direction magnetic field A3 generated by flowing the current to the x-direction drive line 14 is generated.
Acts on the storage carrier 13.

At this time, the total vector of the driving action A2 from the second y-direction drive line 12 and the action by the magnetic field A3 generated by flowing the current through the x-direction drive line 14 is the magnetization vector of the storage carrier 13. Since the direction is shifted from the easy axis, the magnetization reversal can be caused in the storage carrier 13 with a smaller magnetic field intensity as compared with the action acting in parallel to the easy axis. Then, the storage carrier 13 thus has a driving action A2 from the second y-direction driving line 12,
When both the action by the magnetic field A3 generated by flowing the current through the x-direction drive line 14 is present, the magnetization reversal is caused to occur. As a result, as shown in FIG. 6E, the magnetization direction M1 of the storage carrier 13 is reversed from the −x direction (left direction in the drawing) to the + x direction (right direction in the drawing).
The magnetization direction M1 of No. 3 is aligned with the direction of the driving action A2 from the second y-direction driving line 12.

Thereafter, a current is supplied to the x-direction drive line 14,
And when the current supply to the first conductive layer 11b is stopped,
The magnetization direction M1 of the storage carrier 13 inverted in the x direction is kept as it is. When the current supply to the x-direction drive line 14 and the current supply to the first conductor layer 11b are stopped, the drive action A1 from the first y-direction drive line 11 and the drive action A1 from the second y-direction drive line 12 And the direction of the driving action A1 from the first y-direction driving line 11 and the direction of the driving action A2 from the second y-direction driving line 12 are opposite to each other. The driving action A
1, A2 are offset. Therefore, the magnetization direction M1 of the storage carrier 13 is stabilized by the uniaxial magnetic anisotropy of the storage carrier itself, and the state up to that point (here, the storage carrier 13
(A state in which the magnetization direction is in the + x direction).

As described above, in this memory cell 10,
First conductor layer 11 constituting first y-direction drive line 11
b and ON / OFF of the current supply to the second conductor layer 12b constituting the second y-direction drive line 12.
N / OFF, ON / OFF of current supply to the x-direction drive line 14
By switching OFF, the magnetization direction M1 of the storage carrier 13 can be reversed, and binary recording can be performed depending on the magnetization direction of the storage carrier 13. Moreover, in the memory cell 10, the magnetization direction M1 of the storage carrier 13 can be repeatedly inverted, and the recorded information can be repeatedly rewritten. Further, the memory cell 10
In order to maintain the magnetization direction M1 of the storage carrier 13, there is no need to supply a current to the first conductor layer 11b, the second conductor layer 12b, and the x-direction drive line 14, and the memory becomes a nonvolatile memory. I have.

4-2-3 Matrix Addressing In the memory cell 10 described above, three of the first y-direction drive line 11, the second y-direction drive line 12, and the x-direction drive line 14 correspond to the storage carrier 13 Is a driving source for reversing the magnetization direction M1 of the storage medium 13 with only one driving source.
Cannot be reversed, and when the two driving sources are simultaneously turned on, the magnetization reversal occurs in the storage carrier 13. Therefore, a plurality of elongated y-direction drive lines 11, 12 and x-direction drive lines 14 are arranged vertically and horizontally on the substrate, and the memory cells as shown in FIG. 5 and FIG. By arranging 10, a magnetic storage device in which a large number of memory cells 10 are arranged in an integrated manner can be obtained.

That is, the first and second y-direction drive lines 1
1 and 12 are arranged so that their longitudinal directions are parallel to each other, and the x-direction drive line 14 is orthogonal to the longitudinal direction of the first and second y-direction drive lines 11 and 12. 23, by arranging a plurality of storage carriers 13 at positions corresponding to the intersections of the first and second y-direction drive lines 11 and 12 and the x-direction drive line 14, respectively.
As in the conventional solid-state magnetic memory described above, an exchange-coupled solid-state magnetic memory having an address function capable of selecting and writing an arbitrary memory cell in a simple matrix arrangement (that is, an exchange-coupled solid-state magnetic memory having a matrix-type addressing function). Type solid-state magnetic memory).

Specifically, for example, as shown in FIG. 7, a plurality of first y-direction drive lines 11A, 11B,... And a plurality of second y-direction drive lines 12A, 12B,. And
A plurality of first y-direction drive lines and a plurality of second y-direction drive lines are arranged in parallel in the y-direction. That is,
First y-direction drive line 11A and second y-direction drive line 12A
21A, a combination 21B of the first y-direction drive line 11B and the second y-direction drive line 12B,.
Are arranged parallel to the y direction. Further, a plurality of x-direction drive lines 14A, 14B, 14C,... Are arranged in parallel in the x-direction. And the storage carrier 13A-
1, 13A-2, ..., 13B-1, 13B-2, ...
..., 13C-1, 13C-2, ... are arranged respectively.

Then, for example, one y-direction drive line 11
When A and one x-direction drive line 14B are selected and the appropriate currents I ₁ and I ₂ are supplied to both at the same time, magnetization reversal occurs only in the storage carrier 13B-1 at the intersection of these two lines. At this time, the y-direction drive line 11A to which the current is supplied
The x-direction drive line 14B and the x-direction drive line 14B exert a drive action on a plurality of storage carriers arranged along the y-direction drive line 11A and the x-direction drive line 14B.
Is not enough to cause magnetization reversal. Then, a driving action caused by passing a current through the y-direction driving line 11A and a driving action caused by passing a current through the x-direction driving line 14B are combined, so that the driving action on the storage carrier 13 is 45 degrees from the easy axis. Only when the direction is ゜ (that is, FIG.
In the example of (1), magnetization reversal occurs only in the storage carrier 13B-1).

By implementing matrix type addressing in the exchange-coupled solid-state magnetic memory as described above, the conventional solid-state magnetic memory shown in FIG. With a simple configuration like a memory, writing to any memory cell is possible.

Such an exchange-coupled solid-state magnetic memory has the following features.
Even if the matrix type addressing is adopted, since it is not necessary to use a non-linear element or a selection transistor, it is possible to use only a metal material or an insulating material, and it is not necessary to use a semiconductor sensitive to contamination. In the case of using only a metal material or an insulating material, there is also an advantage that a manufacturing process can be greatly simplified because a semiconductor sensitive to contamination is not used.

[0088] 5. Exchange coupled solid state magnet to which the present invention is applied
Specific Embodiment 5-1 of Memory An array of a plurality of magnetic thin films formed on a substrate is used as an overall configuration storage carrier. In addition, a plurality of drive lines are used as addressing transmission lines for selecting an individual storage carrier and transmitting an operation for achieving a write operation to only one of the exchange-coupled solid-state magnetic memories to the storage carrier. Place the lines on the substrate. In order to achieve effective addressing with as few drive lines as possible, a plurality of drive line sets (for example, a set of drive lines running in the x direction and a set of drive lines running in the y direction) are provided. The storage carrier is arranged at a position of a lattice point formed by the intersection of the set of drive lines.

5-2 Available Exchange Coupling Mechanisms There are various origins of the mechanism in which the exchange interaction for transmitting the driving force from the drive line to the storage carrier occurs as follows. There are also input methods suitable for externally controlling the magnitude of the exchange interaction.

5-2-1 Magnetic Coupling via a Semiconductor Layer The carrier in the semiconductor that comes into contact with the magnetic material has a spin density distribution that attenuates vibrationally with distance from the magnetic material,
A magnetic interaction (RKKY interaction) occurs with another magnetic ion or magnetic substance within a distance that the polarization (shift of the average spin of the carrier from zero) extends. This interaction results in exchange coupling between the two magnetic layers separated by the semiconductor layer.

[0091] The period of the vibration according to the magnitude and distance of the magnetic interaction depends on the carrier density. In addition, the carrier density of the semiconductor can be changed by an external stimulus such as electrical stimulation (voltage application, current injection, or the like) or light irradiation. Therefore, by applying an external stimulus to the semiconductor layer, the magnetic coupling between the upper and lower magnetic layers can be changed. Therefore, for example, as shown in FIG.
, A magnetic metal thin film 32 having a fixed magnetization direction and a magnetic metal thin film 33 having a movable magnetization direction are arranged to face each other, and by switching ON / OFF of a voltage, the magnetization vector of the magnetic metal thin film 33 is inverted. Driving force can be generated.

In particular, in magnetic coupling via a semiconductor layer,
Due to the oscillatory nature of the spin density distribution, not only the strength of the magnetic coupling but also the sign of the magnetic coupling may change. In other words, in the case of magnetic coupling via the semiconductor layer, it is determined whether the magnetization of the upper and lower magnetic layers is likely to be parallel (ferromagnetic) or antiparallel (antiferromagnetic). It may be possible to control with external stimulus given. By using a drive line capable of reversing the drive direction in this manner, for example, the functions of the two drive lines (first and second y-direction drive lines 11 and 12) in the memory cell shown in FIG. Can be realized by one drive line. Therefore, for example, when there are m memory cells in the vertical direction and n memory cells in the horizontal direction,
As in the conventional solid-state magnetic memory shown in FIG.
With only one drive line, matrix addressing is achieved.

5-2-2 Magnetic Coupling via Dielectric Layer An exchange coupling may be provided between the magnetic layers via the dielectric layer. At this time, exchange coupling between the magnetic layers is mediated by tunnel electrons connecting both magnetic layers. Therefore, for example, as shown in FIG. 9, a magnetic metal thin film 42 having a fixed magnetization direction and a magnetic metal thin film 43 having a movable magnetization direction are arranged to face each other via a dielectric layer 41, ,
When a voltage is applied from the electrode 43 or an electrode provided separately therefrom to change the potential distribution of the laminated structure, the tunnel probability of electrons passing through the dielectric layer 41 changes, and the metal magnetic thin film 42 and the metal magnetic thin film 43 The exchange coupling between them changes. This can be used as a driving force for reversing the magnetization direction.

When a plurality of dielectric layers 41a and 41b are formed as shown in FIG. 10, the structure has a plurality of potential barriers. The probability that electrons pass through a structure having a plurality of potential barriers shows a remarkable maximum when the electrons have energy that resonates through a potential well formed between the barriers. If the energy distribution of the electrons or the potential distribution of the structure is changed between the resonance and the non-resonance, a relatively small external electrical stimulus can cause a large change in the tunnel probability, and as a result, the exchange by the tunnel electrons can occur. Significant changes in binding can occur.

5-2-3 Coupling via Conductor Layer Even in a conductor layer made of a nonmagnetic metal or the like, RKKY interaction is common, and magnetic coupling can be obtained between the magnetic layers via this. However, since the conductor has a large number of carriers and a short relaxation time, it is not easy to change the number of carriers by an external stimulus like a semiconductor, and it is therefore difficult to modulate the magnetic coupling. However, modulation of magnetic coupling can be realized by devising the structure of the material.

For example, a magnetic coupling between magnetic layers can be cut off by disposing a coupling control layer composed of a laminated film of Cr / Fe-Ag between magnetic layers and supplying a current to the coupling control layer. The structure shown in FIG. 3 corresponds to an example of using such a principle. Note that this is an example using a conductor and also corresponds to an example using a composite material described in section 5-2-4. The method of controlling by current in this way has advantages that the operation speed is not limited by the electric capacity and a high-breakdown-voltage insulating material is not required.

5-2-4 Coupling via Composite Material Instead of a single-phase material, a composite material as shown in FIGS. 11 and 12 may be used for a coupling control layer for controlling magnetic coupling between magnetic layers. In addition, it is possible to control to propagate the magnetic coupling and change the strength of the coupling by an external stimulus.

FIG. 11 shows a coupling control layer having a multilayer structure in which a magnetic layer 51 and a non-ferromagnetic layer 52 are laminated. Here, the magnetic layer 5 is used as a component of the multilayer structure.
For example, a ferromagnetic metal such as Fe, Co, or Ni, or an alloy obtained by diluting them with a non-magnetic metal can be used as 1. The non-ferromagnetic layer 52 includes Ti, V, M
n, Cu, Zr, Nb, Mo, Ru, Rh, Pd, A
Most metals such as g, Hf, Ta, W, Re, Ir, Pt, and Au can be used. Note that Cr or the like, which itself is antiferromagnetic at room temperature, can also be used for the non-ferromagnetic layer 52. Whether the obtained coupling is ferromagnetic or antiferromagnetic, or its strength, etc., gives various design possibilities depending on the kind of the magnetic material to be laminated and the thickness of the non-ferromagnetic layer 52. .

In addition to the lamination structure, a fine particle dispersion structure as shown in FIG. 12 can also be used as the binding control layer. This coupling control layer has a structure in which ferromagnetic fine particles 53 made of Fe or the like are dispersed inside a nonmagnetic material 54 made of Ag or the like. At this time, the magnetic coupling propagates through the ferromagnetic fine particles 53 like a stepping stone, and as a result, the magnetic layers disposed above and below the coupling control layer are magnetically coupled.

At this time, the magnetic coupling between the ferromagnetic fine particles is very weak, and the magnetic coupling is likely to be broken by excessive electron scattering or temperature rise when a current flows. That is, in the case of the coupling control layer having a fine particle dispersion structure, the magnetic coupling between the magnetic layers disposed above and below the magnetic layer depends on the weak magnetic coupling between the ferromagnetic fine particles. Easy magnetic coupling is easily broken.

Incidentally, such a fine particle dispersion structure can also be used as an element of a laminated structure. For example, in an exchange-coupling-type solid-state magnetic memory described later, the coupling control layer
Although a Fe / Ag film is used, since the Fe-Ag film is made of a non-solid solution two-phase mixed material, it can be said that the Fe-Ag film has a fine particle dispersed structure.

When a composite material containing a magnetic substance is used for the coupling control layer, the magnetic coupling is indirectly mediated by the magnetic substance in the composite material. It can be relatively thick. Therefore, when using a composite material containing a magnetic material for the coupling control layer,
The thickness is preferably 10 nm or more. When the film thickness is 10 nm or more, it is possible to avoid the problem that the production is difficult because the thickness of the coupling control layer is too small.

Although the upper limit of the thickness of the coupling control layer made of a composite material containing a magnetic material is not particularly limited, considering the production process and the like in actual production,
It is desirable that the thickness of the coupling control layer be about 1 μm or less.

5-2-5 A material having a relatively low Curie temperature at which the coupled magnetic order via another kind of magnetic substance disappears,
The ferrimagnetic material in the state near the compensation point also significantly changes the macro magnetic characteristics by the external stimulus. This can be used for modulation of magnetic coupling between magnetic layers.

5-2-6 Method of Modulating from External Layer Exchange coupling described in Sections 5-2-1 to 5-2-5, and a Fe / Cr multilayer film, a Co / Cu multilayer film, and a F
In exchange coupling via a non-magnetic metal generally found in e / Au multilayer films and the like, the interface outside the magnetic layer (the side not in contact with the intermediate layer disposed between the magnetic layers), that is, the protective layer and the underlayer Alternatively, there is a method that utilizes the modulation of a potential barrier formed at the interface between the substrate and the magnetic layer. FIG. 13 shows an example of a structure when this method is used.

In the structure shown in FIG. 13, a magnetic layer 58a, an intermediate layer 59 and a magnetic layer 58b laminated on a substrate 57 are formed.
A two-layer film including a semiconductor layer 55a and a metal layer 55b is used as the protective layer 55 for protecting the semiconductor device. Semiconductor layer 55a
The metal layer 55b is arbitrarily selected from a combination that forms a Schottky barrier at the interface. As long as the condition for forming a Schottky barrier is satisfied, the semiconductor layer 55a may be any of an element semiconductor, a compound semiconductor, an oxide semiconductor, and a mixed crystal semiconductor, and the metal layer 55b may be a ferromagnetic material or a nonmagnetic material. But it doesn't matter. This metal / semiconductor two-layer structure is called a “Schottky barrier layer”.

In such a structure, an electrode 56 is formed on the Schottky barrier layer, and as shown in FIG. 14, when a voltage is applied to the electrode 56, the semiconductor layer 55a and the metal layer 55
An electric field is generated at the interface with b, and the height of the potential barrier changes due to the resulting Schottky effect. Accordingly, the reflectance of electrons at the interface between the semiconductor layer 55a and the metal layer 55b is modulated, and as a result, the magnetic layer 58a, the intermediate layer 59, the magnetic layer 58b, and the metal layer 55b are formed on the substrate 57.
The interference effect of the electrons of the intermediate layer in the entire structure including the semiconductor layer 55a and the semiconductor layer 55a is modulated. Thereby, the magnetic coupling between the magnetic layers 58a and 58b is modulated.

That is, in the state of FIG. 13 in which no voltage is applied to the electrode 56, the Schottky barrier between the semiconductor layer 55a and the metal layer 55b is high, and the magnetization directions of the magnetic layer 58a and the magnetic layer 58b are changed. Are in an anti-parallel state as shown by an arrow in the figure, whereas in the state of FIG. 14 in which a voltage is applied to the electrode 56, a Schottky barrier between the semiconductor layer 55a and the metal layer 55b is formed. As a result, the sign of the exchange coupling between the magnetic layer 58a and the magnetic layer 58b changes, and the magnetization direction of the magnetic layer 58a and the magnetization direction of the magnetic layer 58b become parallel as indicated by the arrow in the figure. State.

When a good Schottky barrier is formed between the magnetic layer 58b and the semiconductor layer 55a, the metal layer 55b may be omitted. If a good barrier is formed, an insulating layer may be used instead of the semiconductor layer 55a.

In the above example, the protective layer 55 is used as a Schottky barrier layer, but another layer may be used as a Schottky barrier layer. That is, for example, a method in which an underlayer is formed between a substrate and a magnetic layer and the underlayer is used as a Schottky barrier layer, or a semiconductor substrate is used as a substrate, and the semiconductor substrate is used as a semiconductor in the Schottky barrier layer. A method of using as a layer is also possible. Further, a structure having two or more Schottky barrier layers by using both the protective layer and the base layer or both the protective layer and the substrate is also possible.

The method of modulating the potential barrier is not limited to the method using the Schottky effect. For example, in a structure similar to the above structure, the amount of doping into the semiconductor is increased and the tunnel effect is used. A method of making the effective potential barrier height zero when an external bias electric field is applied, a method of utilizing avalanche breakdown by a high electric field, and the like can also be used. Further, the semiconductor junction (p-
n junction) can be used instead of the above-mentioned Schottky junction or the like.

5-3 Experiment of Addressing Operation
Verification Next, a result of actually manufacturing an exchange-coupled solid-state magnetic memory to which the present invention is applied and verifying the addressing operation thereof will be described.

5-3-1 Exchange-Coupled Solid-State Magnetic Memory
Manufacturing Procedure An exchange-coupled solid-state magnetic memory to which the present invention was applied was manufactured using a magnetron sputtering apparatus. Hereinafter, the manufacturing procedure will be described with reference to FIGS. FIG. 15 shows a process of manufacturing an exchange-coupled solid-state magnetic memory.
19 to 19 are enlarged sectional views of a portion corresponding to one memory cell.

(1) Deposition of Fixed Magnetic Layer and Cu Layer for Controlling Magnetization Direction (FIG. 15) A high coercivity Co-Pt magnetic layer (permanent magnet layer) 61 and a Co layer 62 are formed on a glass substrate 60. After the deposition, a resist mask pattern 63 was formed by electron beam engraving, and a Cu layer 64 was deposited in a strip shape in a region corresponding to one y-direction drive line.

Here, the high coercive force Co-Pt magnetic layer 61 has a thickness of 100 nm, the Co layer 62 has a thickness of 100 nm,
The thickness of the u layer 64 was 0.8 nm.

(2) Deposition of Secondary Co Layer and Bond Control Layer (FIG. 16) The resist mask pattern 63 is removed, and the secondary Co layer 65 is removed.
Was deposited to a thickness of 20 nm. Cu layer 6 of secondary Co layer 65
3 is magnetized in an anti-parallel direction to the underlying Co layer 62 by antiferromagnetic exchange interaction via the Cu layer 63.

Subsequently, an insulating bonding layer 66 was deposited. The insulating coupling layer 66 is formed by sputtering and depositing a Fe—Si target in an oxygen-containing argon atmosphere, and is made of a material that has high electric resistance but exhibits ferromagnetism and propagates magnetic coupling.

Next, a bonding control layer 67 was deposited. Fe
-Ag mosaic target (15-degree sector angle Ag
While the two plates of the six plates are arranged on the Fe target) and the two Cr targets are simultaneously sputtered, the substrate 60 alternately stays on each of the targets.
An r / Fe-Ag multilayer was deposited at room temperature. The thickness of each layer is 0.9 nm of Cr and 1.5 nm of Fe-Ag. Deposition is started from the Fe-Ag layer on the ferrite thin film, and is half-deposited in 16 cycles so that the Fe-Ag layer is at the top. finished. As described in section 5-2-4, this layer has a function of breaking magnetic coupling when a current flows. Further, an electrode pad was formed on an outer peripheral portion of a region where a memory cell is to be formed so that an electrode for supplying a current to the coupling control layer 67 can be formed.

Further, an insulating coupling layer 68 was deposited on the coupling control layer 67 so as not to cover the electrode pads derived from the coupling control layer 67.

(3) Formation of Drive Line Pattern (FIG. 17) A resist mask pattern 69 is formed so as to connect the electrode pads in a region corresponding to two y-direction drive lines (per cell width). The portion was cut down to the middle of the underlying Co layer 62. Thus, the y-direction drive lines 70 and 71 were formed.

(4) Deposition of Ni—Fe Layer and Application of Magnetic Anisotropy (FIG. 18) After the steps are filled with the insulating resin 72 and flattened, the insulating coupling layer 6 is formed.
8, a Ni—Fe layer 73 was deposited. During deposition, an external magnetic field is applied in the −x direction to remove the Ni—Fe layer 73 except for a magnetic bias that propagates from the base by heating the substrate.
A uniaxial magnetic anisotropy having an easy axis in the x-axis direction was induced.
Note that the Ni—Fe layer 73 serves as a storage carrier.

(5) Formation of Storage Carrier and Deposition / Formation of X-Direction Drive Line (FIG. 19) The Ni-Fe layer 73 is left in the dimensions of the storage carrier by a masking process, and is filled with an insulating resin 74 and then placed in the x-direction. And an x-direction drive line 75 was formed. Thereafter, a magnetic field of 2 kOe was applied in the x direction at room temperature using an electromagnet, and the magnetization directions of the high coercive force Co-Pt magnetic layers 61 and Co layers 62 were aligned in the -x direction.

As described above, an exchange-coupled solid-state magnetic memory including a drive line was manufactured. In FIGS. 15 to 19, one memory cell is shown in an enlarged manner, but actually 4 × 4 memory cells are formed. Here, FIG. 20 shows a plan structure of a solid-state magnetic memory in which 4 × 4 memory cells are formed. FIG. 21 shows an enlarged plan view of one memory cell.

In the exchange-coupled solid-state magnetic memory, the x-direction drive line 75 is a simple conductor, and the effect of the conductor on the magnetization of the storage carrier formed of the Ni—Fe layer 73 is as follows. It depends on the magnetic field created by the flowing current. This time, for simplification of the process, only the x-direction drive line 75 does not use exchange coupling. However, it is needless to say that all the driving can be realized by the use of the exchange coupling by using the drive line which gives the effect of tilting the carrier magnetization in the y direction by using the exchange coupling. In that case, 5-2-1
From among the sections shown in Sections 5-2 to 5-5, an appropriate mechanism that generates driving when the electric input is ON may be selected and used. In addition, a driving line of a type that is disconnected when the electric input used in this embodiment is turned on is used, and the driving force is balanced by superimposing a bias from another magnetic material as in the fixed magnetic layer 4 in FIG. It is needless to say that the driving may be performed when the electric input is ON as a result.

The above-described device structure includes the following ideas which are important in device fabrication.

(1) Fixed magnetic layer deposited on the entire surface of the substrate In the exchange-coupled solid-state magnetic memory, a high coercive force Co-P
A fixed magnetic layer in which a t-magnetic layer 61 and a Co layer 62 are stacked is deposited on the entire surface of the substrate. By magnetizing this pinned magnetic layer in one direction and constructing the structure up to the cell array on it, the uniformity of the magnetization direction of all the drive lines and storage carriers over the entire substrate with respect to the pinned magnetic layer Enhanced. This uniformity contributes to the uniformity of the signal particularly in the process of reading data from the memory, and enhances the reliability.

(2) Cu Layer Used for Controlling Magnetization Direction The drive directions of the two types of drive lines 70 and 71 for driving the magnetization of the storage carrier in opposite directions (such as the + x direction and the −x direction) are correct. It is expected that they are parallel and opposite.
As a means for forming such a regular magnetic domain structure in the drive lines 70 and 71, the above-described example utilizes the property that the magnetizations of the Co layers 62 and 65 on both sides via the Cu layer 63 are antiparallel to each other. ing. It is known that the same anti-parallel coupling is caused by a combination of various materials such as a coupling between Fe layers via a Cr layer, and an appropriate combination can be selected and used for memory fabrication. Note that such antiparallel coupling is described in, for example, SSP Parkin, Physica
l Review Letters, vol. 61, p. 3598-3601, (1991).

(3) The current is confined in the coupling control layer. The magnetic coupling propagates as an insulating coupling layer. As a material having a high electric resistance and mediating the magnetic coupling, in the above-described example, a Fe—Si target is sputter-deposited in an oxygen-containing atmosphere. The thin film was used. It is considered that the mixture is a mixture of a magnetic metal alloy and an oxide. Not limited to this composition,
A material having a similar function can be obtained by sputtering an alloy target containing Fe, Co, and Ni as main components in an oxygen-containing atmosphere.

5-3-2 Confirmation of Addressing Operation Using the exchange-coupled solid-state magnetic memory manufactured as described above, it was confirmed that a memory cell could be actually selected and written. Here, the magnetization direction of the storage carrier was detected using a Kerr microscope. The Kerr microscope uses a polarization microscope image based on the fact that the rotation of the polarization plane (Magneto-optical Kerr Effect) that occurs when light is reflected from the surface of a magnetic sample reflects the magnetization direction of the sample. This is a device that gives light and dark contrast depending on magnetization. In the experiment, the optical arrangement was selected so that the contrast depending on the magnetization component in the x direction, which is the easy axis direction of the storage carrier, could be detected. Prior to the observation, the insulating resin overlying the 4 × 4 memory carriers was removed by ion milling so that the Ni—Fe thin film serving as the memory carriers was exposed on the surface. This is a countermeasure for avoiding the addition of unnecessary contrast other than the magneto-optical Kerr effect due to the superposition of the birefringence and the surface reflection of the resin.

(1) In the initial magnetization state, the permanent magnet underlayer of the sample was magnetized in the −x direction and the memory carrier layer was also aligned in the −x direction, as described in the process described in section 5-3-1. . When this was observed with a Kerr microscope, all 16 storage carriers appeared to have the same brightness.

(2) Next, the sample is transferred to a micro prober to set up four electrodes, one of the y-direction drive lines and one of the x-direction drive lines are selected, and a pulse current is supplied to both at the same time. did. In order to invert the magnetization of the storage carrier in the + x direction, a drive line in which the bias in the −x direction to the storage carrier was weakened by current supply was selected.

(3) The sample was returned to the Kerr microscope again and placed in the same direction as the first observation, and the image was observed. FIG. 22 shows a schematic diagram of the observed image. As shown in FIG. 22, only the storage carrier of the selected memory cell (the storage carrier 73A in the second row from the top and the third column from the left in the figure) is observed brighter than the other storage carriers, and It was confirmed that the magnetization direction changed.

(4) Next, a pulse current was supplied to another y-direction drive line and another x-direction drive line for the purpose of returning the sample to the prober, selecting the same memory cell, and "erasing" the memory contents. After that, when a Kerr microscope was observed, 1
All six storage carriers appeared the same brightness. Thus, it was confirmed that memory writing by reversal of magnetization was performed reversibly.

(5) When the above experiment was repeatedly performed on a plurality of different memory cells, it was confirmed that writing and erasing could be performed independently for each memory cell. That is, in the exchange-coupling type solid-state magnetic memory, the addressing operation could be performed by the drive lines arranged in a matrix.

[0135]

As described above in detail, according to the present invention, in a magnetic storage device having an array of magnetic materials as a storage carrier, the use of a magnetic field such as the occurrence of crosstalk and a decrease in coercive force due to miniaturization. The addressing function indispensable for the integrated circuit element can be realized by a simple matrix-type wiring while solving the problem associated with writing.

[Brief description of the drawings]

FIG. 1 is a diagram showing a structure in which a magnetic body A and a magnetic body B are in contact with each other.

FIG. 2 is a diagram showing a structure in which an intermediate layer C exists between a magnetic body A and a magnetic body B.

FIG. 3 is a diagram illustrating an example of an exchange-coupled solid-state magnetic memory;

FIG. 4 is a diagram showing a relationship between a dimension L of a memory cell of a solid-state magnetic memory and a driving magnetic field H that can be used for driving a storage carrier.

FIG. 5 is an enlarged view of one memory cell in an example of an exchange-coupled solid-state magnetic memory to which the present invention is applied.

6A and 6B are diagrams for explaining the driving principle of the memory cell shown in FIG. 5; FIG. 6A shows a state in which the magnetization direction of the storage carrier is held to the right; ) Is the second
6C shows a state in which a current is applied only to the second conductive layer constituting the y-direction drive line, FIG. 6C shows a state in which the magnetization direction of the storage carrier is rewritten leftward, and FIG. d)
FIG. 6E is a diagram showing a state in which a current is applied only to the first conductive layer constituting the first y-direction drive line, and FIG. 6E is a diagram showing a state in which the magnetization direction of the storage carrier is rewritten rightward. is there.

FIG. 7 is a diagram for explaining an addressing method in an exchange-coupled solid-state magnetic memory to which the present invention is applied.

FIG. 8 is a diagram illustrating a state of magnetic driving via the semiconductor layer when a semiconductor layer is used as a coupling control layer.

FIG. 9 is a diagram illustrating a state of magnetic driving via the dielectric layer when a dielectric layer is used as the coupling control layer.

FIG. 10 is a diagram showing a state of magnetic driving via the plurality of dielectric layers when a plurality of dielectric layers are used as the coupling control layer.

FIG. 11 is a diagram showing a coupling control layer having a multilayer structure in which a magnetic layer and a non-ferromagnetic layer are stacked.

FIG. 12 is a diagram showing a coupling control layer in which magnetic particles are dispersed in a non-magnetic material.

FIG. 13 is a diagram showing an example of a structure in which the magnetic coupling of the magnetic layer can be modulated by using the modulation of a potential barrier formed at the outer interface of the magnetic layer.

FIG. 14 is a diagram showing a state where a voltage is applied to the structure shown in FIG. 13;

FIG. 15 is a first diagram showing a manufacturing process of an example of the exchange-coupled solid-state magnetic memory to which the present invention is applied;

FIG. 16 is a second diagram showing the manufacturing process of one example of the exchange-coupled solid-state magnetic memory to which the present invention is applied;

FIG. 17 is a third diagram showing the manufacturing process of an example of the exchange-coupled solid-state magnetic memory to which the present invention is applied;

FIG. 18 is a fourth diagram showing the manufacturing process of one example of the exchange-coupled solid-state magnetic memory to which the present invention is applied;

FIG. 19 is a fifth diagram showing the manufacturing process of one example of the exchange-coupled solid-state magnetic memory to which the present invention is applied;

FIG. 20 is a diagram showing a plan structure of an exchange-coupled solid-state magnetic memory in which 4 × 4 memory cells are formed.

21 is an enlarged view of a portion of a circle S in FIG. 20, and is an enlarged view of a planar structure of one memory cell.

FIG. 22 is a diagram showing the results of observation with a Kerr microscope after performing a write operation on a memory cell in order to confirm an addressing operation in the exchange-coupled solid-state magnetic memory shown in FIG. 20;

FIG. 23 is a diagram illustrating an addressing method in a conventional solid-state magnetic memory.

[Explanation of symbols]

10 memory cells, 11 and 12 y-direction drive lines,
13 memory carrier, x direction drive line

Claims (25)

[Claims] 1. A magnetic storage device having, as a storage carrier, an array of a plurality of separated magnetic materials, wherein an arbitrary storage carrier selected as an object to be written or read is specified to achieve a desired operation. A magnetic storage device utilizing exchange interaction propagating in a solid as means. 2. A structure having a structure in which a coupling control layer is sandwiched between two magnetic layers, wherein the exchange interaction is an exchange interaction acting between the two magnetic layers via the coupling control layer. 2. The magnetic memory according to claim 1, wherein when writing or reading is performed by selecting a carrier, a change in exchange interaction between two magnetic layers caused by applying a stimulus to the coupling control layer is used. apparatus. 3. The coupling control layer is composed of a semiconductor layer. The exchange interaction is mediated by conduction electrons of the semiconductor layer, and when writing or reading is performed by selecting an arbitrary storage carrier, 3. The magnetic storage device according to claim 2, wherein a change in exchange interaction between two magnetic layers caused by applying an electrical stimulus to the layers is used. 4. The coupling control layer comprises a dielectric layer, wherein the exchange interaction is mediated by electrons moving between magnetic layers by a tunnel effect through the dielectric layer, and selects an arbitrary storage carrier. 3. The magnetic storage device according to claim 2, wherein a change in exchange interaction between two magnetic layers caused by changing a height of a tunnel barrier of the dielectric layer is performed when writing or reading is performed. . 5. The coupling control layer is made of a conductor layer, the exchange interaction is an exchange interaction acting between two magnetic layers via the conductor layer, and an arbitrary storage carrier is selected. 3. The magnetic memory device according to claim 2, wherein when writing or reading is performed, a change in exchange interaction between two magnetic layers caused by passing a current through the conductive layer is used. 6. The magnetic memory device according to claim 2, wherein the coupling control layer has a thickness of 10 nm or more and contains a magnetic material. 7. The magnetic memory device according to claim 6, wherein the coupling control layer is a multilayer structure in which a magnetic layer and a non-ferromagnetic layer are stacked. 8. The magnetic memory device according to claim 6, wherein the coupling control layer is formed by dispersing magnetic particles in a non-magnetic material. 9. The magnetic memory device according to claim 2, wherein a magnetic layer made of a hard magnetic material is formed below the structure in which the coupling control layer is sandwiched between the two magnetic layers. 10. At least one of the magnetic layers sandwiching the coupling control layer is formed by stacking a pair of magnetic layers via an intermediate layer such that the magnetization directions are antiparallel to each other. The magnetic storage device according to claim 2. 11. The magnetic memory device according to claim 2, wherein a thin film made of an electrically insulating material that mediates magnetic coupling is disposed between the magnetic layer and the coupling control layer. 12. A structure in which an intermediate layer is sandwiched between two magnetic layers and a coupling control layer is provided on one or both magnetic layers on a side opposite to the intermediate layer. A change in exchange interaction between two magnetic layers caused by stimulating the coupling control layer when writing or reading is performed by selecting an arbitrary storage carrier. 2. The magnetic storage device according to claim 1, wherein the magnetic storage device is used. 13. A plurality of linear members are arranged so as to intersect, and individual storage carriers are arranged at positions corresponding to intersections of the linear members, and an arbitrary storage carrier is selected and written. Or, when performing a read operation, by performing a write or read operation on the selected storage medium by combining the magnetic interaction exerted on the storage medium from the two or more linear members, 2. The magnetic storage device according to claim 1, wherein at least one is due to an exchange interaction propagating in a solid. 14. A plurality of linear members are arranged so as to intersect, and individual storage carriers are arranged at positions corresponding to intersections of the linear members, and an arbitrary storage carrier is selected and written. When performing the operation, 1
The direction of magnetization of one storage carrier is controlled by a combination of magnetic interactions exerted on the storage carrier by three or more linear members, and at least one of the magnetic interactions is exchange interaction propagating in a solid. 2. The magnetic storage device according to claim 1, wherein the magnetic storage device is operated. 15. An addressing method in a magnetic storage device having an array of a plurality of separated magnetic materials as a storage carrier, wherein when an arbitrary storage carrier is selected to perform writing or reading, it is propagated through a solid. An addressing method characterized by utilizing an exchange interaction. 16. The exchange interaction is an exchange interaction acting between two magnetic layers via a coupling control layer in a structure in which a coupling control layer is sandwiched between two magnetic layers, and selects an arbitrary storage carrier. 16. The addressing method according to claim 15, wherein when writing or reading is performed, a change in exchange interaction between the two magnetic layers caused by applying a stimulus to the coupling control layer is used. 17. The semiconductor device according to claim 1, wherein the coupling control layer is formed of a semiconductor layer, and the exchange interaction is mediated by conduction electrons of the semiconductor layer. 17. The addressing method according to claim 16, wherein a change in exchange interaction between the two magnetic layers caused by applying an electrical stimulus to the layers is used. 18. The coupling control layer comprises a dielectric layer, wherein the exchange interaction is mediated by electrons moving between magnetic layers by a tunnel effect through the dielectric layer, and selects an arbitrary storage carrier. 17. The addressing method according to claim 16, wherein when writing or reading is performed, a change in exchange interaction between two magnetic layers caused by changing a height of a tunnel barrier of the dielectric layer is used. 19. The coupling control layer is made of a conductor layer, the exchange interaction is an exchange interaction acting between two magnetic layers via the conductor layer, and an arbitrary storage carrier is selected. 17. The addressing method according to claim 16, wherein when writing or reading is performed, a change in exchange interaction between two magnetic layers caused by passing a current through the conductive layer is used. 20. The addressing method according to claim 16, wherein the coupling control layer has a thickness of 10 nm or more and contains a magnetic material. 21. The addressing method according to claim 16, wherein the coupling control layer has a multilayer structure in which a magnetic layer and a non-ferromagnetic layer are stacked. 22. The addressing method according to claim 16, wherein the coupling control layer is formed of magnetic particles dispersed in a non-magnetic material. 23. The exchange interaction described above, wherein an intermediate layer is sandwiched between two magnetic layers, and a coupling control layer is provided on one or both magnetic layers on a side opposite to the intermediate layer. An exchange interaction that acts between the magnetic layers, and utilizes the change in the exchange interaction between the two magnetic layers caused by stimulating the coupling control layer when writing or reading data by selecting an arbitrary storage carrier. 16. The addressing method according to claim 15, wherein: 24. The magnetic storage device, wherein a plurality of linear members are arranged so as to intersect, and individual storage carriers are arranged at positions corresponding to the intersections of the linear members, and an arbitrary storage device is provided. When writing or reading is performed by selecting a carrier, by combining magnetic interactions exerted on the storage carrier from two or more linear members, a write or read operation is performed on the selected storage carrier, and 16. The addressing method according to claim 15, wherein at least one of the static interactions is due to an exchange interaction propagating in a solid. 25. The magnetic storage device, wherein a plurality of linear members are arranged so as to intersect, and individual storage carriers are arranged at positions corresponding to intersections of the linear members. When selecting a carrier and performing a write operation, 1
The direction of magnetization of one storage carrier is controlled by a combination of magnetic interactions exerted on the storage carrier by three or more linear members, and at least one of the magnetic interactions is exchange interaction propagating in a solid. 16. The addressing method according to claim 15, wherein the addressing is performed by an action.