JP7186622B2 - Magnetoresistive element and method for manufacturing magnetoresistive element - Google Patents

Description

The present invention relates to a magnetoresistive element and a method of manufacturing a magnetoresistive element.

Ferromagnetic tunnel junction (MTJ) elements can exhibit a large magnetoresistive ratio (MR ratio), and are therefore used in magnetic devices such as magnetic heads and nonvolatile random access magnetic memories (MRAM) in recent years. An MTJ element is composed of, for example, two ferromagnetic layers and a tunnel barrier layer provided between the two ferromagnetic layers. The ferromagnetic layer includes, for example, Co, Fe, Co alloy, Fe alloy, CoFe alloy, and Co-based Heusler alloy, and the tunnel barrier layer includes, for example, MgAl ₂ O ₄ having a spinel structure. This creates a coherent tunnel effect, resulting in a large MR ratio.

Patent Document 1 and Patent Document 2 both disclose a ferromagnetic tunnel junction, a magnetoresistive effect element using the same, and a spintronics device. Patent Document 1 discloses a ferromagnetic tunnel junction in which the tunnel barrier layer includes MgAl ₂ O ₄ with a spinel structure, and Patent Document 2 discloses a tunnel barrier layer in which the tunnel barrier layer has a disordered spinel structure (Mg _1−x Al _x )-O(0<x≦1). In any of the patent documents, the ferromagnetic tunnel junction has a structure in which a spinel-structured crystalline oxide tunnel barrier layer is directly sandwiched between two ferromagnetic layers.

Japanese Patent No. 5586028 JP 2013-175615 A

2. Description of the Related Art In MTJ elements used in magnetic devices such as magnetic heads and MRAMs, film development is being carried out to reduce area resistance (RA), which is a resistance value per unit area of the MTJ element. On the other hand, in MTJ elements used in magnetic devices such as new type magnetic field sensors and analog memories using domain wall motion, it is desired to increase the RA of the MTJ elements. A known technique for increasing the RA of the MTJ element is to increase the thickness of the tunnel barrier layer. However, according to this technique, although the RA is increased, there arises a problem that the MR ratio is reduced due to the inclusion of pinholes in the tunnel barrier layer. Therefore, it has been difficult to increase the RA of the MTJ element while suppressing the reduction of the MR ratio.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a magnetoresistive element having a high RA while suppressing a decrease in the MR ratio, and a method of manufacturing such a magnetoresistive element. and

A magnetoresistive element according to one aspect of the present invention includes a first ferromagnetic layer, a second ferromagnetic layer, and a tunnel layer stacked between the first ferromagnetic layer and the second ferromagnetic layer. a barrier layer, the tunnel barrier layer having a top surface and a bottom surface opposite the top surface, wherein the maximum height roughness of the top surface and the maximum height roughness of the bottom surface of the tunnel barrier layer are , Rz/t<1, where Rz is the greater maximum height roughness, t is the thickness of the tunnel barrier layer, and the tunnel barrier layer includes at least one amorphous region.

According to this magnetoresistive element, the tunnel barrier layer has a maximum height roughness of Rz on its top surface and/or bottom surface and includes at least one amorphous region, so that a high RA can be realized. can be done. In addition, since Rz/t<1 is satisfied in the tunnel barrier layer, it is possible to prevent the MR ratio from being reduced due to the maximum height roughness becoming excessive with respect to the thickness of the tunnel barrier layer.

In the magnetoresistive element according to one aspect of the present invention, the tunnel barrier layer further includes at least one crystallized region, and in a cross section along the stacking direction, the at least one crystallized region occupies the area of the entire tunnel barrier layer. may be greater than 5% and less than 65%.

According to this magnetoresistive element, the ratio of the crystallized region is greater than 5% in the cross section along the lamination direction, so that the reduction of the MR ratio due to the excessive ratio of the amorphous region can be suppressed. In addition, since the percentage of the crystallized region is less than 65%, the percentage of the amorphous region is sufficiently large to prevent deterioration of the ferromagnetic layer that may occur due to the process treatment for partially crystallizing the tunnel barrier layer. The RA is sufficiently high while preventing the reduction of the MR ratio.

In the magnetoresistive element according to one aspect of the present invention, at least one crystallized region may have a crystal structure in which the (001) plane is oriented in the lamination direction.

According to this magnetoresistive element, a high coherent tunneling effect can be obtained from the crystallized region in which the (001) plane is oriented. This magnetoresistive element can have a high MR ratio.

In the magnetoresistive element according to one aspect of the present invention, at least one crystallized region has a width along the in-plane direction of the tunnel barrier layer, and the maximum width may be less than 8 nm. .

According to this magnetoresistive element, even if lattice mismatch occurs between the tunnel barrier layer and at least one of the first ferromagnetic layer and the second ferromagnetic layer in the crystallized region of the tunnel barrier layer, A region where lattice mismatch occurs can be reduced. As a result, this magnetoresistive element can have a higher MR ratio.

In the magnetoresistive element according to one aspect of the present invention, at least part of the at least one crystallized region may be lattice-matched with at least one of the first ferromagnetic layer and the second ferromagnetic layer.

A high coherent tunnel effect can be obtained with this magnetoresistive element. This magnetoresistive element can have a higher MR ratio.

In the magnetoresistive element according to one aspect of the present invention, at least one of the first ferromagnetic layer, the second ferromagnetic layer, and the tunnel barrier layer may contain twin crystals.

According to this magneto-resistive element, the twin crystals can relax strain that may occur in the vicinity of the interface between the layers due to the roughness of the upper and lower surfaces of the tunnel barrier layer. As a result, lattice strain at the interface between the tunnel barrier layer and the first and/or second ferromagnetic layers is relaxed.

In the magnetoresistive element according to one aspect of the present invention, the twin crystal has a face-centered cubic lattice structure and the twin crystal plane is the (111) plane, or the twin crystal has a body-centered cubic lattice structure. and the twin plane may be the (112) plane.

According to this magnetoresistive element, the twin crystal can have an appropriate twin plane according to the type of crystal structure of the twin crystal. Lattice strain at the interface with the layer is further relaxed. This magnetoresistive element can have a higher MR ratio.

In the magnetoresistive element according to one aspect of the present invention, the tunnel barrier layer may contain an oxide material having a spinel structure represented by formula (1).
A _1-x B _x O _y (1)
In formula (1), A represents at least one element selected from the group consisting of Mg, Zn, Cu, Cd, Li, Ni, Co, Fe, Mn, Cr, Hg, and V; represents at least one element selected from the group consisting of Al, Ga, In, Sc, Ti, V, Cr, Mn, Fe, Rh, Ir, Ge, and Co; 35≤y≤1.7.

According to this magnetoresistive element, lattice mismatch with the first and second ferromagnetic layers can be further reduced in the crystallized region of the tunnel barrier layer.

In the magnetoresistive element according to one aspect of the present invention, A may represent Mg and B may represent Al in the formula (1).

According to this magnetoresistive element, this configuration allows the tunnel barrier layer to sufficiently exhibit the function of realizing a large MR ratio.

In the magnetoresistive element according to one aspect of the present invention, 0<x<0.5 may be satisfied in the formula (1).

According to this magnetoresistive element, the function of the tunnel barrier layer to realize a large MR ratio in the formula (1) can be more sufficiently exhibited.

In the magnetoresistive element according to one aspect of the present invention, at least one of the first ferromagnetic layer and the second ferromagnetic layer is Co _z Fe _100-z (0<z<80) in contact with the tunnel barrier layer. You may have a region consisting of

According to this magnetoresistive element, the spin polarization of the region can be increased and the lattice mismatch between the region and the tunnel barrier layer can be reduced.

A method for fabricating a magnetoresistive element according to one aspect of the present invention includes the steps of forming a first ferromagnetic layer, forming a MgAl alloy layer on the first ferromagnetic layer, and forming the MgAl alloy layer. Oxidized to form at least one amorphous region and at least one crystallized region having a spinel structure represented by Mg _1-x Al _x O _y (0<x≦1, 0.35≦y≦1.7) and forming a second ferromagnetic layer on the tunnel barrier layer, the tunnel barrier layer having a top surface and a bottom surface opposite to the top surface. of the maximum height roughness of the upper surface and the maximum height roughness of the lower surface of the tunnel barrier layer, and let Rz be the larger maximum height roughness, and let t be the thickness of the tunnel barrier layer, then Rz /t<1.

According to this method of manufacturing a magnetoresistive element, it is possible to manufacture a magnetoresistive element having a high RA while suppressing a decrease in the MR ratio.

In the method of manufacturing a magnetoresistive element according to one aspect of the present invention, the step of oxidizing the MgAl alloy layer includes a step of oxidizing the MgAl layer by at least one of radical oxidation, plasma oxidation, ozone oxidation, and natural oxidation. may contain.

According to this method of manufacturing a magneto-resistive element, the tunnel barrier layer of the predetermined mode as described above can be efficiently formed.

ADVANTAGE OF THE INVENTION According to this invention, the magnetoresistive effect element which has high RA, and the method of manufacturing such a magnetoresistive effect element can be provided, suppressing the fall of MR ratio.

FIG. 1 is a diagram showing a cross section along the stacking direction of a magnetoresistive element according to an embodiment. FIG. 2 is an enlarged view showing an example of the area E1 shown in FIG. FIG. 3 is an enlarged view showing another example of the area E1 shown in FIG. FIG. 4 is a diagram showing a magnetoresistive device capable of evaluating the MR ratio of the magnetoresistive effect element according to Example 1. FIG.

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments for carrying out the present invention will be described in detail with reference to the accompanying drawings. In each drawing, the same reference numerals are used for the same elements when possible. Also, the dimensional ratios within and between the constituent elements in the drawings are arbitrary for ease of viewing of the drawings. In this embodiment, the X-axis, Y-axis, and Z-axis based on the orthogonal coordinate system are set with respect to FIGS. 1 to 3, but these axes are set for convenience of explanation. , the Z-axis is set in the stacking direction of the magnetoresistive elements.

FIG. 1 is a diagram showing a cross section along the stacking direction of a magnetoresistive element according to an embodiment. The magnetoresistive element 1 includes, for example, a substrate 10, an underlying layer 20, a magnetoresistive layer 30, and a cap layer 40 in this order. In FIG. 1, the magnetoresistive effect element 1 is drawn with the thickness of each layer in the stacking direction (Z-axis direction) enlarged for easy understanding.

The substrate 10 includes, for example, silicon single crystal, metal oxide single crystal, silicon single crystal with a thermal silicon oxide film, sapphire single crystal, ceramic, quartz, and glass. The material included in the substrate 10 is not particularly limited as long as it has appropriate mechanical strength and is suitable for heat treatment and fine processing. Metal oxide single crystals include, for example, MgO single crystals, and an epitaxially grown film can be easily formed on a substrate containing MgO single crystals by, for example, sputtering.

The underlayer 20 is provided to improve the crystallinity of the magnetoresistive layer 30 and can be used as an electrode for measuring the magnetoresistive characteristics of the magnetoresistive effect element 1 . The base layer 20 includes, for example, a lower electrode layer 21 for electrodes and a buffer layer 22 provided on the lower electrode layer 21 .

The lower electrode layer 21 includes, for example, at least one metal element selected from Al, Ag, Au, Cu, Cr, Pt, Ru, Ta, V, and W, an alloy of these metal elements, or an alloy of these metal elements. A laminate of two or more materials may be included. Alloys of metallic elements include, for example, cubic AgZn alloys, AgMg alloys, and NiAl alloys. Examples of the layered body of a material composed of two or more kinds of metal elements include a layered body of a material containing Ta and Ru. Specifically, the lower electrode layer 21 can include a three-layer stack of Ta, Ru, and Ta.

The lower electrode layer 21 has an upper surface 21a, and the upper surface 21a can be subjected to CMP (Chemical Mechanical Polishing). For CMP polishing, for example, a CMP apparatus is used to polish the upper surface 21a of the lower electrode layer 21 so that the upper surface 21a has a desired roughness. The roughness of the upper surface 21a can be controlled by the polishing time of CPM polishing. In one example, the roughness of the upper surface 21a can be reduced as the polishing time of the CPM polishing is increased. Alumina, for example, is used as slurry for CPM polishing.

The thickness of the lower electrode layer 21 in the Z-axis direction is, for example, in the range of 10 nm or more and 1000 nm or less. Also, the thickness is, for example, in the range of 30 nm or more and 150 nm or less.

A buffer layer 22 can be laminated on the upper surface 21 a of the lower electrode layer 21 after polishing, and the buffer layer 22 contains, for example, the same metal element as the lower electrode layer 21 . The metal element is, for example, at least one of Al, Ag, Au, Cu, Cr, Pt, Ru, Ta, V, and W, and the buffer layer 22 is an alloy of these metal elements or an alloy of these metal elements. A laminate of two or more materials may be included. Alloys of metal elements are, for example, cubic AgZn alloys, AgMg alloys and NiAl alloys. The buffer layer 22 can specifically include a two-layer stack of Ta and Ru.

The thickness of the buffer layer 22 in the Z-axis direction is, for example, in the range of 1 nm or more and 10 nm or less. Also, the thickness is, for example, in the range of 2 nm or more and 5 nm or less.

Between the underlayer 20 and the substrate 10, a crystal orientation layer may be provided for controlling the crystal orientation of the upper layer, if necessary. The crystal orientation layer includes, for example, at least one of MgO, TiN and NiTa alloys.

The magnetoresistive layer 30 includes a first ferromagnetic layer 31, a second ferromagnetic layer 32, a tunnel barrier layer 33, an antiferromagnetic layer 34, a third ferromagnetic layer 35, and a magnetic coupling layer 36. and The tunnel barrier layer 33 is made of an insulating material and laminated between the first ferromagnetic layer 31 and the second ferromagnetic layer 32 . The antiferromagnetic layer 34 is arranged on the opposite side of the second ferromagnetic layer 32 from the tunnel barrier layer 33 side, and is stacked on the third ferromagnetic layer 35 in this embodiment. The first ferromagnetic layer 31 is laminated on the underlayer 20, for example, and includes the first ferromagnetic layer 31, the tunnel barrier layer 33, the second ferromagnetic layer 32, the magnetic coupling layer 36, and the third ferromagnetic layer. The magnetic layers 35 are arranged in this order. The first ferromagnetic layer 31 functions as a magnetization free layer, and the second ferromagnetic layer 32, the magnetic coupling layer 36, the third ferromagnetic layer 35 and the antiferromagnetic layer 34 are jointly magnetized. It can function as a fixed layer.

The first ferromagnetic layer 31 functioning as a magnetization free layer is made of a ferromagnetic material having soft magnetic properties such as Fe, Co--Fe, Co--Fe--B, and a ferromagnetic Heusler alloy.

The thickness of the first ferromagnetic layer 31 in the Z-axis direction is, for example, in the range of 0.5 nm or more and 100 nm or less. Also, the thickness is, for example, in the range of 0.8 nm or more and 20 nm or less.

In this embodiment, the magnetization fixed layer has a structure called a synthesized anti-ferromagnetic structure (SAF structure). That is, the magnetoresistive layer 30 is provided between the second ferromagnetic layer 32 and the antiferromagnetic layer 34 in addition to the second ferromagnetic layer 32 and the antiferromagnetic layer 34 as a magnetization fixed layer. It further has a third ferromagnetic layer 35 and a magnetic coupling layer 36 provided between the second ferromagnetic layer 32 and the third ferromagnetic layer 35 . The second ferromagnetic layer 32 and the third ferromagnetic layer 35 are antiferromagnetically coupled to each other by exchange coupling via the magnetic coupling layer 36 . The third ferromagnetic layer 35 can have a configuration similar to that of the second ferromagnetic layer 32 . The magnetic coupling layer 36 is made of a nonmagnetic metal such as Ru, Rh, or Ir, and its thickness in the Z-axis direction is, for example, in the range of 0.3 nm or more and 2 nm or less. In this embodiment, the magnetization pinned layer may not have the SAF structure, if necessary. In this case, the magnetization fixed layer does not have the magnetic coupling layer 36 and the third ferromagnetic layer 35 and the antiferromagnetic layer 34 is laminated directly on the second ferromagnetic layer 32 .

The magnetization directions of the second ferromagnetic layer 32 and the third ferromagnetic layer 35 of the magnetization fixed layer change with respect to the external magnetic field compared to the magnetization direction of the first ferromagnetic layer 31 functioning as the magnetization free layer. hard to do The magnetization directions of the second ferromagnetic layer 32 and the third ferromagnetic layer 35 are substantially fixed with respect to an external magnetic field that may be applied during use, such as an external magnetic field to be measured, and preferably not substantially changed. In the present embodiment, the magnetization directions of the second ferromagnetic layer 32 and the third ferromagnetic layer 35 are exchange-coupled with the antiferromagnetic layer 34 and the third ferromagnetic layer 35, This is realized by imparting unidirectional magnetic anisotropy directly to the third ferromagnetic layer 35 and indirectly to the second ferromagnetic layer 32 . Since this mode of magnetization direction can be realized without using the antiferromagnetic layer 34 , the magnetoresistive layer 30 does not have to have the antiferromagnetic layer 34 . Even if the magnetization fixed layer does not have the SAF structure, the magnetoresistive layer 30 may not have the antiferromagnetic layer 34 for the same reason. On the other hand, the first ferromagnetic layer 31 functioning as a magnetization free layer is made of, for example, a soft magnetic material, and its magnetization direction is not substantially fixed. Therefore, when an external magnetic field is applied to the magnetoresistive layer 30, the magnetization direction easily changes in that direction. When an external magnetic field is applied to the magnetoresistive layer 30 and the relative magnetization directions of the first ferromagnetic layer 31 and the second ferromagnetic layer 32 change, the resistance value of the magnetoresistive layer 30 changes, The magnetoresistive element 1 exerts a magnetoresistive effect.

The second ferromagnetic layer 32 and the third ferromagnetic layer 35 can be composed of ferromagnetic material. Ferromagnetic materials for the magnetization fixed layer include, for example, metals or alloys having at least one element selected from Ni, Fe, Ru, Mn, Ir, and Co. Specific examples of alloys include Co--Fe alloys, Ni--Fe alloys, Co--B alloys, Fe--B alloys and Co--Fe--B alloys. Heusler alloys such as Co--Fe--Al alloys, Co--Fe--Si alloys, Co--Mn--Si alloys, Co--Mn--Ge alloys, Co--Fe--Al--Si alloys, and Co--Fe--Ga--Ge alloys. mentioned. The second ferromagnetic layer 32 can have a structure in which multilayer films such as Co--Fe--B alloy and Co--Fe alloy are laminated repeatedly.

The thickness of the second ferromagnetic layer 32 in the Z-axis direction is, for example, in the range of 2 nm or more and 10 nm or less. Also, the thickness is, for example, in the range of 3 nm or more and 7 nm or less. The thickness of the third ferromagnetic layer 35 in the Z-axis direction is, for example, in the range of 1 nm or more and 8 nm or less. Also, the thickness is, for example, in the range of 2 nm or more and 6 nm or less.

The antiferromagnetic layer 34 is composed of an antiferromagnetic material such as FeMn alloy, PtMn alloy, PtCrMn alloy, NiMn alloy, IrMn alloy, NiO, _Fe2O3 , or the _like .

The thickness of the antiferromagnetic layer 34 in the Z-axis direction is, for example, in the range of 5 nm or more and 20 nm or less.

In the magnetoresistive layer 30, one of the first ferromagnetic layer 31 and the second ferromagnetic layer 32 may function as a magnetization free layer, and the other may function as part of the magnetization fixed layer. Therefore, as a modification of this embodiment, the first ferromagnetic layer 31 may function as part of the magnetization fixed layer, and the second ferromagnetic layer 32 may function alone as the magnetization free layer. In this case, the antiferromagnetic layer 34 is formed directly on the opposite side of the first ferromagnetic layer 31 from the tunnel barrier layer 33 side, or by connecting the magnetic coupling layer 36 and the third ferromagnetic layer 35 so as to form an SAF structure. , and directly or indirectly imparts unidirectional magnetic anisotropy to the first ferromagnetic layer 31 .

The magnetoresistive element 1 can have a cap layer 40 on the magnetoresistive layer 30 . A cap layer 40 is provided to protect the magnetoresistive layer 30 . The cap layer 40 is, for example, one or more metal elements of Ru, Ag, Al, Cu, Au, Cr, Mo, Pt, W, Ta, Pd, and Ir, alloys of these metal elements, or Laminates of two or more materials may be included. If necessary, an upper electrode may be provided on the cap layer 40 to allow current to flow through the magnetoresistive element 1 along the Z-axis direction.

FIG. 2 is an enlarged view showing an example of the region E1 shown in FIG. 1, and is a view showing one example of a cross section along the stacking direction (Z-axis direction) of the magnetoresistive layer 30. As shown in FIG. As shown in FIG. 2, the tunnel barrier layer 33 has an upper surface 33a and a lower surface 33b opposite the upper surface 33a. The upper surface 33a can be positioned on the second ferromagnetic layer 32 side, and the lower surface 33b can be positioned on the first ferromagnetic layer 31 side. The upper surface 33a of the tunnel barrier layer 33 has surface roughness. Due to the surface roughness, the upper surface 33a has, for example, one or more protrusions 33c and one or more depressions 33d.

The upper surface 33a of the tunnel barrier layer 33 has a maximum height roughness Rz1 as the surface roughness. This maximum height roughness Rz1 is defined by the JISB 0601:2013 standard. According to this standard, the maximum height roughness Rz1 of the upper surface 33a is defined by the average line Av1 within the range of the reference length Lr1 of the roughness curve of the upper surface 33a (the roughness of the upper surface 33a within the range of the reference length Lr1 defined as the sum of the height Rp1 of the projection 33c, which is the highest from the arithmetic mean height of the curve), and the depth Rv1 of the recess 33d, which is the lowest from the average line Av1. The roughness curve of the upper surface 33a is measured by, for example, TEM (transmission electron microscope), STEM (scanning transmission electron microscope), SEM (scanning electron microscope) or AFM (atomic force microscope). The reference length Lr1 is a predetermined length along the in-plane direction (XY in-plane direction) of the magnetoresistive layer 30, and is, for example, 4 nm, 8 nm, 10 nm, 20 nm, or 30 nm.

Like the upper surface 33a, the lower surface 33b of the tunnel barrier layer 33 has surface roughness, and due to the surface roughness, for example, one or more protrusions 33f and one or more depressions 33g are formed. have. The maximum height roughness Rz2 of the lower surface 33b is defined by the JISB 0601:2013 standard, like the maximum height roughness Rz1 of the upper surface 33a. The maximum height roughness Rz2 is the height Rp2 of the highest protrusion 33f from the average line Av2 and the depth of the lowest depression 33g from the average line Av2 within the range of the reference length Lr2 of the roughness curve of the lower surface 33b. It is defined as the value of the sum with Rv2. The roughness curve of the lower surface 33b is also measured by TEM or the like, and the reference length Lr2 is, for example, 4 nm, 8 nm, 10 nm, 20 nm, or 30 nm along the in-plane direction of the magnetoresistive layer 30. Although FIG. 2 illustrates an example in which the reference length Lr1 is the same length as the reference length Lr2, the reference length Lr1 may be a length different from the reference length Lr2.

In the magnetoresistive element 1, the larger maximum height roughness Rz can be selected from the maximum height roughness Rz1 of the upper surface 33a of the tunnel barrier layer 33 and the maximum height roughness Rz2 of the lower surface 33b. The upper surface 33a of the tunnel barrier layer 33 may have the maximum height roughness Rz, and the lower surface 33b may have the maximum height roughness Rz. When the maximum height roughness Rz1 of the upper surface 33a is the same as the maximum height roughness Rz2 of the lower surface 33b, both the upper surface 33a and the lower surface 33b may have the maximum height roughness Rz. An upper surface 33a and/or a lower surface 33b of the tunnel barrier layer 33 has a maximum height roughness Rz. FIG. 2 shows the case where the upper surface 33a has a larger maximum height roughness Rz than the lower surface 33b.

When the thickness of the tunnel barrier layer 33 in the Z-axis direction is t, Rz/t<1 in at least one of the cross-sections of the magnetoresistive layer 30 along the Z-axis direction. When the maximum height roughness Rz1 of the upper surface 33a is the maximum height roughness Rz, the thickness t of the tunnel barrier layer 33 in the Z-axis direction is, for example, within the range of the reference length Lr1. It is obtained by arithmetically averaging the thickness in the Z-axis direction. The thickness t of the tunnel barrier layer 33 in the Z-axis direction is measured by, for example, TEM (transmission electron microscope), STEM (scanning transmission electron microscope), or SEM (scanning electron microscope). When the maximum height roughness Rz2 of the lower surface 33b is the maximum height roughness Rz, the thickness t of the tunnel barrier layer 33 in the Z-axis direction is, for example, within the range of the reference length Lr2. It is obtained by arithmetically averaging the thickness in the Z-axis direction.

In the present embodiment, it is preferable that Rz/t<1 in at least two cross sections perpendicular to each other among cross sections along the Z-axis direction of the magnetoresistive layer 30, and along the Z-axis direction of the magnetoresistive layer 30 More preferably, Rz/t<1 in substantially all cross sections.

The surface roughness of the upper surface 33a and the lower surface 33b of the tunnel barrier layer 33 is the same as that of the upper surface of the element below the tunnel barrier layer 33 of the magnetoresistive element 1 (for example, the upper surface 21a of the lower electrode layer 21 (see FIG. 1)). It is also caused by surface roughness. Therefore, the maximum height roughness Rz of the tunnel barrier layer 33 can be adjusted to a predetermined size by controlling the surface roughness (for example, maximum height roughness) of the upper surface (for example, the upper surface 21a) of the element. can be controlled. The magnitude of the maximum height roughness Rz of the tunnel barrier layer 33 may be about the same as, greater than, or less than the magnitude of the maximum height roughness of the upper surface of the element. good too.

The maximum height roughness Rz of the tunnel barrier layer 33 is, for example, preferably 0.2 nm or more and 3.2 nm or less, and more preferably 0.6 nm or more and 1.8 nm or less. The thickness of the tunnel barrier layer 33 in the Z-axis direction is, for example, preferably 1 nm or more and 3.2 nm or less, and more preferably 1.6 nm or more and 2.2 nm or less.

As shown in FIG. 2, tunnel barrier layer 33 includes at least one amorphous region 61 and at least one crystallized region 62 . Amorphous region 61 and crystallized region 62 can be separated from each other, for example, by boundary 33e. The tunnel barrier layer 33 may substantially consist of at least one amorphous region 61 and at least one crystallized region 62 only. The amorphous region 61 and the crystallized region 62 have substantially the same composition, but may have substantially different compositions.

At least one amorphous region 61 is included in the tunnel barrier layer 33 . The number of amorphous regions 61 may be, for example, two or five in the entire tunnel barrier layer 33 . At least one crystallized region 62 is included in the tunnel barrier layer 33 . The number of crystallized regions 62 may be, for example, two or five in the entire tunnel barrier layer 33 . In the entire tunnel barrier layer 33, the number of amorphous regions 61 and the number of crystallized regions 62 may be the same or different.

Local electron diffraction using TEM and STEM can be used to identify the crystal structure of the tunnel barrier layer 33 . The local electron diffraction method is sometimes called a nanobeam electron diffraction method when the electron beam is focused to the nanometer scale, and an angstrom electron diffraction method when the electron beam is focused to the sub-nanometer scale. When the localized region of the tunnel barrier layer 33 is crystallized, the corresponding electron diffraction image is a spot pattern, and when the localized region of the tunnel barrier layer 33 is amorphous, the corresponding electron diffraction image is a halo pattern. Become. Therefore, according to the local electron diffraction method, it can be determined whether the local region of the tunnel barrier layer 33 is crystallized or amorphous.

In the magnetoresistive element 1 as described above, the upper surface 33a and/or the lower surface 33b of the tunnel barrier layer 33 has the maximum height roughness Rz. Therefore, compared with the case where the upper surface 33a and/or the lower surface 33b are substantially flat, the area of the upper surface 33a and/or the lower surface 33b is increased, so that the RA (per unit area in the XY plane) of the magnetoresistance effect element 1 resistance) increases. Furthermore, since the tunnel barrier layer 33 includes at least one amorphous region 61, the RA of the magnetoresistive element 1 is higher than when the region is crystallized. In addition, since the tunnel barrier layer 33 satisfies Rz/t<1, the reduction in the MR ratio caused by the maximum height roughness Rz of the upper surface 33a being excessive with respect to the thickness t of the tunnel barrier layer 33 is suppressed. can be prevented. As a result, according to the magnetoresistive element 1 of the present embodiment, it is possible to realize a high RA while suppressing a decrease in the MR ratio.

In the tunnel barrier layer 33 of the magnetoresistive element 1 of this embodiment, 0.1<Rz/t is preferable, and 0.32<Rz/t is more preferable. As a result, the area of the upper surface 33a is sufficiently increased, so that the RA of the magnetoresistive element 1 can be sufficiently increased. This condition regarding the lower limit value of Rz/t may be satisfied in at least one of the cross sections of the magnetoresistive layer 30 along the Z-axis direction, similarly to the above-described condition regarding the upper limit value of Rz/t. Of the cross sections along the Z-axis direction of the magnetoresistive layer 30, preferably at least two cross sections that are orthogonal to each other are filled, and substantially all cross sections along the Z-axis direction of the magnetoresistive layer 30 are filled. is more preferred.

FIG. 2 shows an example in which the amorphous region 61 and the crystallized region 62 are arranged in the tunnel barrier layer 33 generally in the in-plane direction (XY in-plane direction). 61 and crystallized region 62 may be arranged in the stacking direction (Z-axis direction). For example, in the tunnel barrier layer 33, the crystallized region 62 may be arranged on the amorphous region 61. For example, the amorphous region 61, the crystallized region 62, and the amorphous region 61 are arranged in this order in the stacking direction. good too.

At least one of the cross sections of the tunnel barrier layer 33 along the Z-axis direction (preferably at least two cross sections of the tunnel barrier layer 33 along the Z-axis direction, more preferably at least two cross sections perpendicular to each other, more preferably the tunnel barrier layer 33) In substantially the entire cross section of the layer 33 along the Z-axis direction), the ratio of the area of the at least one crystallized region 62 to the area of the entire tunnel barrier layer 33 is greater than 5% and greater than 65%. Small is preferred, more preferably greater than 20% and less than 50%.

If the area ratio is greater than 5% (preferably 20%), it is possible to suppress the reduction in the MR ratio due to the excessive proportion of the amorphous region 61 . In addition, when the ratio of the area is less than 65% (preferably 50%), the ratio of the amorphous region 61 is sufficiently large, so that there is a possibility that it may be caused by the process treatment for partially crystallizing the tunnel barrier layer 33. A decrease in the MR ratio due to deterioration of at least one of the first ferromagnetic layer 31 and the second ferromagnetic layer 32 is prevented, and RA is sufficiently high. The proportion of the area of the at least one crystallized region 62 can be controlled, for example, by the annealing temperature of the annealing process performed when forming the tunnel barrier layer 33, and the proportion of the area of the at least one amorphous region 61 can be controlled. may be greater than, approximately equal to, or less than .

Also in at least one cross section along the in-plane direction, the ratio of the area of at least one amorphous region 61 to the area of the entire tunnel barrier layer 33 is equal to that of the at least one crystallized region 62 to the area of the entire tunnel barrier layer 33. may be larger, substantially equal, or smaller than the ratio of the area of . In this embodiment, the area ratio of at least one crystallized region 62 is preferably greater than 5% and less than 65%, more preferably greater than 20% and less than 50%.

At least one amorphous region 61 has a width W61 along the in-plane direction of the tunnel barrier layer 33 and at least one crystallized region 62 has a width W62 along the in-plane direction of the tunnel barrier layer 33 . The width W61 of the amorphous region 61 is, for example, 1 nm or more and 30 nm or less, and the width W62 of the crystallized region 62 is, for example, 0.15 nm or more and less than 8 nm. In the present embodiment, both the width W61 and the width W62 indicate the maximum width defined along the in-plane direction (XY in-plane direction). FIG. 2 shows an example in which both the width W61 and the width W62 are the maximum width values along the X-axis direction.

In the magnetoresistive element 1, the width W62 of the crystallized region 62 may be less than 8 nm. Even if lattice mismatch occurs with at least one of the two ferromagnetic layers 32, the area where lattice mismatch occurs can be reduced. As a result, the magnetoresistive element 1 can have a higher MR ratio. This effect is exhibited more remarkably when the width W62 of the crystallized region 62 is less than 5 nm.

At least one amorphous region 61 has a height H61 along the Z-axis direction, and at least one crystallized region 62 has a height H62 along the Z-axis direction. A height H61 of the amorphous region 61 is, for example, 0.2 nm or more and 3 nm or less, and a height H62 of the crystallized region 62 is, for example, 0.2 nm or more and 3 nm or less.

In this embodiment, at least one crystallized region 62 may have a crystal structure in which the (001) plane is oriented in the Z-axis direction. The magnetoresistive element 1 can obtain a high coherent tunnel effect due to the crystallized region 62 oriented in the (001) plane, and as a result, can have a high MR ratio. This effect is particularly pronounced when the crystallized region 62 is composed of MgO, MgAl ₂ O ₄ and γ-alumina.

At least part of the at least one crystallized region 62 may be lattice-matched with at least one of the first ferromagnetic layer 31 and the second ferromagnetic layer 32 . Thereby, a high coherent tunnel effect can be obtained. The magnetoresistive element 1 can have a higher MR ratio. In the magnetoresistive element 1, the crystallized region 62 is lattice-matched with the first ferromagnetic layer 31 and the second ferromagnetic layer 32, for example, by inverse Fourier transform of a lattice image obtained by structural observation. can be confirmed by the image.

The tunnel barrier layer 33 can contain an oxide material having a spinel structure represented by the following formula (1).
A _1-x B _x O _y (1)
In formula (1), A represents at least one element selected from the group consisting of Mg, Zn, Cu, Cd, Li, Ni, Co, Fe, Mn, Cr, Hg, and V; represents at least one element selected from the group consisting of Al, Ga, In, Sc, Ti, V, Cr, Mn, Fe, Rh, Ir, Ge, and Co; Further, 0<x≦1 and 0.35≦y≦1.7, and O represents oxygen. The above spinel structure also includes a disordered spinel structure in which the atomic arrangement of A and B in formula (1) is disordered.

In this case, the lattice mismatch between the first ferromagnetic layer 31 and the second ferromagnetic layer 32 can be further reduced in the crystallized region 62 of the tunnel barrier layer 33 . This is because the mismatch between the lattice constant of the crystallized region 62 of the tunnel barrier layer 33 and the lattice constants of the first ferromagnetic layer 31 and the second ferromagnetic layer 32 can be reduced.

In formula (1), A may represent Mg and B may represent Al. With this configuration, the tunnel barrier layer 33 can sufficiently exhibit the function of realizing a large MR ratio.

Moreover, in the formula (1), 0<x<0.5 may be satisfied. This allows the tunnel barrier layer 33 to fully exhibit its function of achieving a high MR ratio.

Moreover, in the formula (1), 0<x<0.5 and 1≦y≦1.33 may be satisfied. In this case, the function of the tunnel barrier layer 33 to achieve a high MR ratio can be more fully exhibited.

The tunnel barrier layer 33 can be made of MgO or γ-alumina in addition to the oxide material having the spinel structure represented by the above formula (1).

In the magnetoresistive layer 30, at least one of the first ferromagnetic layer 31, the second ferromagnetic layer 32, and the tunnel barrier layer 33 may contain twins. As a result, the twin crystals can relax strain that may occur in the vicinity of the interface between the upper and lower surfaces 33a and 33b of the tunnel barrier layer 33 due to the roughness of these layers. As a result, the lattice strain at the interface between the tunnel barrier layer 33 and the first ferromagnetic layer 31 and/or the second ferromagnetic layer 32 is relaxed.

The twin has a face-centered cubic lattice structure and the twin plane is the (111) plane, or the twin has a body-centered cubic lattice structure and the twin plane is the (112) plane. It can be a face. Since the twins can have suitable twin planes depending on the type of crystal structure of the twins, the tunnel barrier layer 33 and the first ferromagnetic layer 31 and/or the second ferromagnetic layer 32 The lattice strain at the interface with is further relaxed. The magnetoresistive element 1 can have a higher MR ratio.

At least one of the first ferromagnetic layer 31 and the second ferromagnetic layer 32 may have a region of Co _z Fe _100-z (0<z<80) in contact with the tunnel barrier layer 33 . Thereby, the spin polarization of the region can be increased, and the lattice mismatch between the region and the tunnel barrier layer 33 can be reduced.

FIG. 3 is an enlarged view showing another example of the region E1 shown in FIG. 1, showing another example of one of the cross sections along the stacking direction (Z-axis direction) of the magnetoresistive layer 30. It is a diagram.

The tunnel barrier layer 33 shown in FIG. 3 differs from the tunnel barrier layer 33 shown in FIG. 2 in that it further has an intervening region 63 between the amorphous region 61 and the crystallized region 62 . The intervening region 63 is composed of a material that is neither crystallized nor amorphous among the materials that constitute the tunnel barrier layer 33 . The amorphous regions 61 and the intervening regions 63 can be separated from each other, for example by a boundary 61a, and the crystallized regions 62 and the intervening regions 63 can be separated from each other, for example by a boundary 62a.

Tunnel barrier layer 33 of FIG. 3 can have a structure similar to tunnel barrier layer 33 of FIG. The tunnel barrier layer 33 of FIG. 3 has a top surface 33a and a bottom surface 33b opposite the top surface 33a. The upper surface 33a of the tunnel barrier layer 33 has a maximum height roughness Rz1 as surface roughness. Due to the surface roughness, the upper surface 33a has, for example, one or more protrusions 33c and one or more depressions 33d. Like the upper surface 33a, the lower surface 33b of the tunnel barrier layer 33 has surface roughness, and due to the surface roughness, for example, one or more protrusions 33f and one or more depressions 33g are formed. have.

In this embodiment, the magnetoresistive element 1 is manufactured by forming layers from the underlayer 20 to the cap layer 40 on the substrate 10 by a manufacturing method such as sputtering and electron beam evaporation. When forming each layer, a heat treatment may be performed as necessary, and a magnetic field application treatment for imparting unidirectional magnetic anisotropy may be performed as necessary. A heat treatment may be performed simultaneously with the magnetic field application treatment. Further, the magnetoresistive element 1 may be microfabricated into a shape that enables evaluation of magnetoresistive characteristics by lithography using an electron beam or the like and dry etching using Ar ions or the like. The magnetoresistive element 1 is a magnetoresistive element having a CPP (Current Perpendicular to Plane) structure in which a detection current is passed along the lamination direction (the direction perpendicular to the film surface of each layer, the Z-axis direction).

The method of manufacturing the magnetoresistive element according to the present embodiment includes a step of forming the first ferromagnetic layer 31 on the underlayer 20 and a step of forming the tunnel barrier layer 33 on the first ferromagnetic layer 31. and forming the second ferromagnetic layer 32 on the tunnel barrier layer 33 .

The step of forming the first ferromagnetic layer 31 can be performed by a sputtering method, and in this step, a two-layer laminate consisting of, for example, a CoFeB layer and a CoFe layer is formed on the underlayer 20. .

The step of forming the tunnel barrier layer 33 includes forming a MgAl alloy layer on the first ferromagnetic layer 31 by, for example, a sputtering method, oxidizing the MgAl alloy layer to form at least one amorphous region 61, At least one crystallized region 62 having a spinel structure represented by, for example, Mg _1-x Al _x O _y (0<x≦1, 0.35≦y≦1.7) is formed. and the step of In the tunnel barrier layer 33 formed by this process, Rz is the larger maximum height roughness of the maximum height roughness Rz1 of the upper surface 33a and the maximum height roughness Rz2 of the lower surface 33b. When the thickness of the layer is t, Rz/t<1 in the cross section along the stacking direction.

Also, the step of oxidizing the MgAl alloy layer may include a step of oxidizing the MgAl layer by at least one of radical oxidation, plasma oxidation, ozone oxidation, and natural oxidation. In this case, it is possible to efficiently form the tunnel barrier layer 33 having a predetermined form as described above.

The step of forming the second ferromagnetic layer is performed, for example, by a sputtering method, and in this step, a three-layer laminate including, for example, a CoFe layer, a CoFeB layer and a CoFe layer is formed on the tunnel barrier layer 33. be.

According to this method of manufacturing a magnetoresistive element, it is possible to manufacture the magnetoresistive element 1 having a high RA while suppressing a decrease in the MR ratio.

The magnetoresistive effect element will be further described below with reference to examples and comparative examples of the present invention, but the present invention is not limited to the following examples.

(Example 1)
As Example 1, a magnetoresistive element having the same configuration as the magnetoresistive element 1 of the embodiment shown in FIG. 1 was manufactured. In the description of the first embodiment, the reference numerals used for the description of the magnetoresistive effect element 1 are used. Note that the reference numerals used for explaining the magnetoresistive effect element 1 are also used in the examples after the second example and the comparative examples.

In Example 1, the magnetoresistive element 1 was produced by the following procedure. First, a substrate 10 made of Si was prepared. A thermally oxidized silicon film was provided on the substrate 10 . The substrate 10 provided with the thermally oxidized silicon film was placed in a magnetron sputtering apparatus to form the lower electrode layer 21 of the underlying layer 20 . The lower electrode layer 21 is composed of a three-layer laminate in which a Ta layer (thickness 5 nm), a Ru layer (thickness 100 nm), and a Ta layer (thickness 30 nm) are laminated in this order. The temperature during the formation of the lower electrode layer 21 was room temperature.

The substrate 10 with the lower electrode layer 21 formed thereon was taken out from the magnetron sputtering apparatus and placed in a CMP apparatus. Using a CMP apparatus, the upper surface of the lower electrode layer 21 was polished so as to have a desired roughness. Alumina was used as slurry for polishing. Polishing time was 60 seconds.

After the CMP polishing, the substrate 10 was placed in the magnetron sputtering apparatus again, and the buffer layer 22 was formed on the polished lower electrode layer 21 by sputtering. The buffer layer 22 was a two-layer laminate consisting of a Ta layer (2 nm thick) and a Ru layer (2 nm thick). The temperature during formation of the buffer layer 22 was room temperature.

Next, the magnetoresistive layer 30 was formed on the buffer layer 22 of the underlayer 20 by sputtering. In forming the magnetoresistive layer 30, first, the first ferromagnetic layer 31 was formed as a magnetization free layer. The first ferromagnetic layer 31 was a two-layer stack consisting of a CoFeB layer (thickness 3 nm) and a CoFe layer (thickness 1 nm). The temperature at which the first ferromagnetic layer 31 was formed was room temperature. No heat treatment was performed after forming the first ferromagnetic layer 31 .

Subsequently, a tunnel barrier layer 33 was formed on the first ferromagnetic layer 31 . In forming the tunnel barrier layer 33, first, an MgAl alloy layer was formed on the first ferromagnetic layer 31, and then the MgAl alloy layer was oxidized. The MgAl alloy layer was formed by sputtering, and the MgAl alloy layer was oxidized by natural oxidation. After performing an annealing treatment (at a temperature of 300° C.) for 15 minutes, a tunnel barrier layer 33 made of MgAl ₂ O ₄ was produced. The thickness t of the tunnel barrier layer 33 was set to 1.80 nm.

Next, a magnetization fixed layer having an SAF structure was formed on the tunnel barrier layer 33 . The magnetization pinned layer is a three-layer laminate consisting of a CoFe layer (thickness 1 nm) as the second ferromagnetic layer 32, a CoFeB layer (thickness 3 nm) and a CoFe layer (thickness 1 nm), and a magnetic coupling layer 36 of A multi-layer laminate comprising a Ru layer (thickness 0.8 nm), a CoFe layer (thickness 3 nm) as the third ferromagnetic layer 35, and an IrMn layer (thickness 8 nm) as the antiferromagnetic layer 34 was formed. The temperature during the formation of the magnetization fixed layer was room temperature. No heat treatment was performed after forming the magnetization fixed layer. A magnetoresistive layer 30 was fabricated by forming the magnetization fixed layer.

After manufacturing the magnetoresistive layer 30 , a cap layer 40 was formed on the magnetization fixed layer of the magnetoresistive layer 30 . The cap layer 40 was a two-layer laminate including a Ru layer (thickness 3 nm) and a Ta layer (thickness 5 nm). The temperature during formation of the cap layer 40 was room temperature. No heat treatment was performed after the cap layer 40 was formed.

After forming the cap layer 40, heat treatment in a magnetic field is performed to exchange-couple the IrMn layer as the antiferromagnetic layer 34 and the third ferromagnetic layer 35, thereby forming a CoFe layer as the third ferromagnetic layer 35 and One-way magnetic anisotropy was imparted to a three-layered laminate consisting of a CoFe layer, a CoFeB layer and a CoFe layer as the second ferromagnetic layer 32 . In this magnetic field heat treatment, the heat treatment temperature was set at 300° C., the heat treatment time was set at 3 hours, and the intensity of the applied magnetic field was set at 10 kOe (798 kA/m). By this heat treatment in a magnetic field, the fabrication of the magnetoresistive element 1 according to Example 1 was completed.

(Measurement of MR ratio)
FIG. 4 is a diagram showing a magnetoresistive device capable of measuring the MR ratio of the magnetoresistive effect element according to Example 1. FIG. The magnetoresistive device 50 includes a first electrode layer 51 and a second electrode layer 52 sandwiching the magnetoresistive element 1 together with the first electrode layer 51 . The magnetoresistive element 1 was microfabricated into a shape suitable for measurement of magnetoresistive characteristics. The first electrode layer 51 is connected to the underlying layer 20 on the substrate 10 of the magnetoresistive element 1 , and the second electrode layer 52 is connected to the cap layer 40 of the magnetoresistive element 1 . The magnetoresistive device 50 further comprises a power source 53 and a voltmeter 54 , both connected to the first electrode layer 51 and the second electrode layer 52 . A current is applied to the magnetoresistive element 1 in the stacking direction by the power supply 53 , and the voltage applied to the magnetoresistive element 1 at this time can be monitored by the voltmeter 54 . In this embodiment, a voltage applied to the magnetoresistive element 1 is monitored by a voltmeter 54 while a magnetic field is swept to the magnetoresistive element 1 from the outside while a constant current is passed through the magnetoresistive element 1 in the stacking direction. By doing so, the resistance change of the magnetoresistive effect element 1 was measured.

In this example, the MR ratio of the magnetoresistive element 1 was calculated from the measurement result of the resistance change. The MR ratio is expressed as a percentage and can be calculated by the following formula (i).
MR ratio (%)=((R _AP −R _P )/R _P )×100(%) (i)
In this formula (i), _RAP is the resistance of the magnetoresistive element 1 when the magnetization direction of the first ferromagnetic layer 31 and the magnetization direction of the second ferromagnetic layer 32 are antiparallel. It's size. _RP is the resistance of the magnetoresistive element 1 when the magnetization direction of the first ferromagnetic layer 31 and the magnetization direction of the second ferromagnetic layer 32 are parallel. The MR ratio was similarly measured in other examples and comparative examples described later.

(Measurement of RA)
In Example 1, the magnetoresistive device 50 was used to measure the resistance change of the magnetoresistive element 1, and from the measurement results, the RA (area resistance) of the magnetoresistive element 1 was measured. In this embodiment, when the area of the magnetoresistive effect element 1 in a plan view is A, R _p A obtained by multiplying R _p in the formula (i) by A is defined as RA (area resistance). RA is the resistance value per unit area obtained by dividing the resistance value obtained by dividing the applied bias voltage by the current flowing in the lamination direction of the magnetoresistive effect element 1 by the reciprocal of the area of the surface where each layer is joined. It is a normalized numerical value. The value of RA (resistance area) includes measurement error within 3%. RA was also measured in the same manner in other examples and comparative examples described later.

(Measurement of maximum height roughness Rz)
In this example, the maximum height roughness Rz1 of the upper surface 33a and the maximum height roughness Rz2 of the lower surface 33b of the tunnel barrier layer 33 were measured by TEM observation of the cross section along the stacking direction. These maximum height roughness Rz1 and maximum height roughness Rz2 were measured according to the JISB 0601:2013 standard, and the reference length Lr1 and the reference length Lr2 were both set to 20 nm. In the present embodiment, of the maximum height roughness Rz1 of the upper surface 33a and the maximum height roughness Rz2 of the lower surface 33b of the tunnel barrier layer 33, the maximum height roughness Rz1 of the upper surface 33a has a larger maximum height roughness. Therefore, the maximum height roughness Rz1 of the upper surface 33a is defined as the maximum height roughness Rz. In other examples and comparative examples described later, the maximum height roughness Rz was similarly measured. Also in these other examples and comparative examples, the maximum height roughness Rz1 of the upper surface 33a showed a larger maximum height roughness, so the maximum height roughness Rz1 of the upper surface 33a was regarded as the maximum height roughness Rz. did.

(Estimation of percentage of crystallized area)
In this example, the ratio of the crystallized region 62 in the tunnel barrier layer 33 was estimated by TEM observation of the cross section along the stacking direction. In other examples and comparative examples to be described later, the ratio of the crystallized region 62 was similarly measured.

(Example 2)
In Example 2, the same procedure as in Example 1 was performed except that the polishing time for CMP polishing the upper surface of the lower electrode layer 21 was set to 30 seconds and the thickness t of the tunnel barrier layer 33 was set to 1.81 nm. A magneto-resistive element 1 having a similar configuration was produced by the method.

(Example 3)
Example 3 was the same as Example 1 except that the polishing time for CMP polishing the upper surface of the lower electrode layer 21 was 15 seconds and the thickness t of the tunnel barrier layer 33 was 1.79 nm. A magneto-resistive element 1 having a similar configuration was produced by the method.

(Example 4)
In Example 4, a magnetoresistive element 1 having the same structure as in Example 1 was fabricated by the same method as in Example 1, except that the thickness t of the tunnel barrier layer 33 was set to 2.00 nm.

(Example 5)
Example 5 was the same as Example 1 except that the polishing time for CMP polishing the upper surface of the lower electrode layer 21 was set to 30 seconds and the thickness t of the tunnel barrier layer 33 was set to 2.01 nm. A magneto-resistive element 1 having a similar configuration was produced by the method.

(Example 6)
Example 6 was the same as Example 1 except that the polishing time for CMP polishing the upper surface of the lower electrode layer 21 was set to 15 seconds and the thickness t of the tunnel barrier layer 33 was set to 1.96 nm. A magneto-resistive element 1 having a similar configuration was produced by the method.

(Example 7)
In Example 7, a magnetoresistive element 1 having the same configuration as in Example 1 was fabricated by the same method as in Example 1, except that the material and thickness of the tunnel barrier layer 33 were different.

The tunnel barrier layer 33 is formed by first forming an Mg layer on the first ferromagnetic layer 31 and then oxidizing the Mg layer. The Mg layer was formed by sputtering, and the Mg layer was oxidized by natural oxidation. After performing an annealing treatment (at a temperature of 250° C.) for 15 minutes, a tunnel barrier layer 33 made of MgO was produced. The thickness t of the tunnel barrier layer 33 was set to 1.79 nm.

(Example 8)
Example 8 was the same as Example 7, except that the polishing time for CMP polishing the upper surface of the lower electrode layer 21 was 30 seconds and the thickness t of the tunnel barrier layer 33 was 1.82 nm. A magneto-resistive element 1 having a similar configuration was produced by the method.

(Example 9)
Example 9 was the same as Example 7, except that the polishing time for CMP polishing the upper surface of the lower electrode layer 21 was 15 seconds, and the thickness t of the tunnel barrier layer 33 was 1.81 nm. A magneto-resistive element 1 having a similar configuration was produced by the method.

(Example 10)
In Example 10, the same structure as in Example 3 was used except that the thickness t of the tunnel barrier layer 33 was set to 1.83 nm and the annealing conditions (temperature and treatment time) of the tunnel barrier layer 33 were used. A magneto-resistive element 1 was produced. The annealing temperature was 380° C. and the annealing time was 15 minutes. In this example, the ratio of the crystallized region 62 was changed by changing the annealing temperature as compared with the third example.

(Example 11)
In Example 11, the same structure as in Example 3 was used except that the thickness t of the tunnel barrier layer 33 was set to 1.82 nm and the annealing conditions (temperature and treatment time) of the tunnel barrier layer 33 were used. A magneto-resistive element 1 was produced. The annealing temperature was 350° C. and the annealing time was 15 minutes. In this example, the ratio of the crystallized region 62 was changed by changing the annealing temperature as compared with the third example.

(Example 12)
In Example 12, the same structure as in Example 3 was used except that the thickness t of the tunnel barrier layer 33 was set to 1.81 nm and the annealing conditions (temperature and treatment time) of the tunnel barrier layer 33 were used. A magneto-resistive effect element was fabricated. The annealing temperature was 330° C. and the annealing time was 15 minutes. In this example, the ratio of the crystallized region 62 was changed by changing the annealing temperature as compared with the third example.

(Example 13)
In Example 13, the same structure as in Example 3 was used except that the thickness t of the tunnel barrier layer 33 was set to 1.83 nm and the annealing conditions (temperature and treatment time) of the tunnel barrier layer 33 were used. A magneto-resistive element 1 was produced. The annealing temperature was 260° C. and the annealing time was 15 minutes. In this example, the ratio of the crystallized region 62 was changed by changing the annealing temperature as compared with the third example.

(Example 14)
In Example 14, the same structure as in Example 3 was used except that the thickness t of the tunnel barrier layer 33 was set to 1.84 nm and the annealing conditions (temperature and treatment time) of the tunnel barrier layer 33. A magneto-resistive element 1 was produced. The annealing temperature was 220° C., and the annealing time was 15 minutes. In this example, the ratio of the crystallized region 62 was changed by changing the annealing temperature as compared with the third example.

(Comparative example 1)
In Comparative Example 1, the same structure as in Example 1 was used except that the upper surface of the lower electrode layer 21 was not CMP-polished and the thickness t of the tunnel barrier layer 33 was set to 1.82 nm. A magneto-resistive effect element was fabricated.

(Comparative example 2)
In Comparative Example 2, the same structure as in Example 1 was used except that the upper surface of the lower electrode layer 21 was not CMP-polished and the thickness t of the tunnel barrier layer 33 was set to 1.98 nm. A magneto-resistive effect element was fabricated.

(Comparative Example 3)
In Comparative Example 3, the same structure as in Example 7 was used except that the upper surface of the lower electrode layer 21 was not CMP-polished and the thickness t of the tunnel barrier layer 33 was set to 1.84 nm. A magneto-resistive effect element was fabricated.

(Comparative Example 4)
In Comparative Example 4, a magnetoresistive element was produced by an epitaxial method. First, the substrate 10 made of MgO single crystal was placed in an epitaxial apparatus, and the underlying layer 20 was formed. In this comparative example, the underlying layer 20 is composed of one Cr layer (40 nm thick). After forming the underlying layer 20, a heat treatment (at a temperature of 800° C.) was performed for 60 minutes.

Next, a magnetoresistive layer 30 was formed on the underlying layer 20 . The magnetoresistive layer 30 has a first ferromagnetic layer 31 as a magnetization free layer, a tunnel barrier layer 33, and a magnetization fixed layer in this order. In this comparative example, the magnetization fixed layer does not have the SAF structure.

First, a first ferromagnetic layer 31 was formed as a magnetization free layer. The first ferromagnetic layer 31 consists of an Fe layer (30 nm thick). After forming the first ferromagnetic layer 31, an annealing treatment (at a temperature of 300° C.) was performed for 60 minutes. A tunnel barrier layer 33 made of MgAl ₂ O ₄ was formed on the first ferromagnetic layer 31 . The thickness t of the tunnel barrier layer 33 was set to 1.80 nm. After forming the MgAl ₂ O ₄ layer, an annealing treatment (at a temperature of 450° C.) was performed for 15 minutes.

Subsequently, a magnetization fixed layer was formed on the tunnel barrier layer 33 . The magnetization pinned layer was a multi-layer laminate consisting of an Fe layer (thickness 6 nm) as the second ferromagnetic layer 32 and an IrMn layer (thickness 12 nm) as the antiferromagnetic layer 34 . After forming the Fe layer, an annealing treatment (at a temperature of 300° C.) was performed for 15 minutes. No heat treatment was performed after forming the magnetization fixed layer. A magnetoresistive layer 30 was fabricated by forming the magnetization fixed layer.

After manufacturing the magnetoresistive layer 30 , a cap layer 40 made of a Ru layer (10 nm thick) was formed on the magnetization fixed layer of the magnetoresistive layer 30 . Further, after the cap layer 40 is formed, heat treatment in a magnetic field is performed to exchange-couple the IrMn layer as the antiferromagnetic layer 34 and the Fe layer as the second ferromagnetic layer 32, so that the Fe layer is One-way magnetic anisotropy was given. The heat treatment temperature in this magnetic field heat treatment was set to 300° C., and the heat treatment time was set to 3 hours. The strength of the applied magnetic field was 10 kOe (798 kA/m). The magnetic field heat treatment completed the formation of the magnetoresistive element 1 according to this comparative example.

Table 1 is a table showing the configuration of each layer included in the magnetoresistive effect elements according to Examples 1 to 3, manufacturing conditions, and measurement results such as RA. Table 2 is a table showing the configuration of each layer included in the magnetoresistive effect elements according to Examples 4 to 6, manufacturing conditions, and measurement results such as RA. Table 3 is a table showing the configuration of each layer included in the magnetoresistive effect elements according to Examples 7 to 9, manufacturing conditions, and measurement results such as RA. Table 4 is a table showing the configuration of each layer included in the magnetoresistive effect elements according to Examples 10 to 12, manufacturing conditions, and measurement results such as RA. Table 5 is a table showing the structure of each layer included in the magnetoresistive effect elements according to Examples 13 and 14, manufacturing conditions, and evaluation results of RA and MR ratios. Table 6 is a table showing the configuration of each layer included in the magnetoresistive effect elements according to Comparative Examples 1 to 3, manufacturing conditions, and measurement results such as RA. Table 7 is a table showing the configuration of each layer included in the magnetoresistive effect element according to Comparative Example 4, manufacturing conditions, and measurement results such as RA.

As shown in Tables 1 to 7, Rz/t of Examples 1 to 14 is Rz/t<1, and Rz/t of Comparative Examples 1 to 4 is Rz/t >1. Examples 1 to 14 were shown to have a high RA while suppressing a decrease in the MR ratio as compared to Comparative Examples 1 to 4. In addition, Examples 1 to 9 and 11 to 13 (the ratio of the crystallized region is more than 5% and less than 65%) are compared with Examples 10 and 14 (the ratio of the crystallized region is 5% or less or 65% or more). As a result, it was shown to have a particularly high RA while suppressing a decrease in the MR ratio.

According to this embodiment, a magnetoresistive element having a high RA while suppressing a decrease in the MR ratio and a method of manufacturing such a magnetoresistive element are provided.

DESCRIPTION OF SYMBOLS 1... Magnetoresistive effect element 31... First ferromagnetic layer 32... Second ferromagnetic layer 33... Tunnel barrier layer 33a... Upper surface 33b... Lower surface 61... Amorphous region 62... Crystallized region.

Claims (12)

a first ferromagnetic layer;
a second ferromagnetic layer;
a tunnel barrier layer laminated between the first ferromagnetic layer and the second ferromagnetic layer;
with
the tunnel barrier layer has a top surface and a bottom surface opposite the top surface;
Among the maximum height roughness of the upper surface and the maximum height roughness of the lower surface of the tunnel barrier layer, Rz is the larger maximum height roughness, and t is the thickness of the tunnel barrier layer. / t < 1,
The magnetoresistive element, wherein the tunnel barrier layer includes at least one amorphous region and at least one crystallized region having a spinel structure represented by formula (1) .
A _1-x B _x O _y (1)
[Wherein, A represents at least one element selected from the group consisting of Mg, Zn, Cu, Cd, Li, Ni, Co, Fe, Mn, Cr, Hg, and V; represents at least one element selected from the group consisting of Ga, In, Sc, Ti, Cr, Rh, Ir, Ge, and Co, where 0<x≦1 and 0.35≦y≦1.7 be. ] 2. The magnetoresistor according to claim 1, wherein the ratio of the area of the at least one crystallized region to the area of the entire tunnel barrier layer in a cross section along the stacking direction is more than 5% and less than 65%. effect element. 3. The magnetoresistive element according to claim 2, wherein said at least one crystallized region has a crystal structure in which (001) planes are oriented in said stacking direction. the at least one crystallized region has a width along the in-plane direction of the tunnel barrier layer;
4. The magnetoresistive element according to claim 1, wherein said maximum width is less than 8 nm. 5. The method according to any one of claims 1 to 4, wherein at least part of said at least one crystallized region is lattice-matched with at least one of said first ferromagnetic layer and said second ferromagnetic layer. magnetoresistive effect element. 6. The magnetoresistive element according to claim 1, wherein at least one of said first ferromagnetic layer, said second ferromagnetic layer, and said tunnel barrier layer contains twins. The twin has a face-centered cubic lattice structure and the twin plane is the (111) plane, or the twin has a body-centered cubic lattice structure and the twin plane is the (112) plane. 7. The magnetoresistive element according to claim 6, wherein 8. The magnetoresistive element according to claim 1, wherein A represents Mg and B represents Al in the formula (1). 9. The magnetoresistive element according to claim 1, wherein 0<x<0.5 in said formula ( 1 ). At least one of the first ferromagnetic layer and the second ferromagnetic layer has a region made of Co _z Fe _100-z (0<z<80) in contact with the tunnel barrier layer. 10. The magnetoresistive element according to any one of 9 . forming a first ferromagnetic layer;
forming a MgAl alloy layer on the first ferromagnetic layer;
The MgAl alloy layer is oxidized to form at least one amorphous region and at least a spinel structure represented by Mg _1-x Al _x O _y (0<x≦1, 0.35≦y≦1.7). forming a tunnel barrier layer comprising one crystallized region;
forming a second ferromagnetic layer on the tunnel barrier layer;
with
the tunnel barrier layer has a top surface and a bottom surface opposite the top surface;
Among the maximum height roughness of the upper surface and the maximum height roughness of the lower surface of the tunnel barrier layer, Rz is the larger maximum height roughness, and t is the thickness of the tunnel barrier layer. A method of manufacturing a magnetoresistive element wherein /t<1. 12. The magnetoresistive element according to claim 11 , wherein said step of oxidizing said MgAl alloy layer includes a step of oxidizing said MgAl alloy layer by at least one of radical oxidation, plasma oxidation, ozone oxidation, and natural oxidation. How to manufacture.

Images