JP2008058323A - Method of evaluating metal structure of r-t-b magnet alloy - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an effective evaluation method of evaluating the metal structure of a rare earth containing alloy. <P>SOLUTION: A microscopic structure of an R-T-B magnet alloy cross section is read in as a digital image, a straight line of length L substantially parallel to a freezing face of the metal is drawn on the microscopic structure image, an R<SB>2</SB>T<SB>14</SB>B main phase and an R-rich phase are determined based on a difference between brightness of the main phase and brightness of an R-rich phase crossed with the straight line, and a value (L/N) provided by dividing the straight line of length L by a frequency N with the brightness on the straight line changed from the brightness judged as the R<SB>2</SB>T<SB>14</SB>B main phase up to the brightness judged as the R-rich phase, is evaluated as an R-rich phase interval r, in the evaluation method of evaluating the metal structure of the R-T-B magnet alloy in the present invention. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a rare earth-containing alloy flake and a method for producing the same, and more particularly to a rare earth-containing alloy produced by a strip casting method and a method for evaluating a metal structure.

  In recent years, alloys for rare earth magnets, especially Nd-Fe-B alloys, have rapidly increased in production due to their high characteristics, and are used for HD (hard disk), MRI (magnetic resonance imaging), and various motors. in use. This kind of rare earth magnet alloy is usually one in which a part of Nd is replaced with another rare earth element such as Pr or Dy, or a part of Fe is replaced with another transition metal such as Co or Ni. Is generally called an RTB-based alloy including Nd-Fe-B-based alloys. Here, R is at least one of rare earth elements including Y. In addition, T is a transition metal in which Fe is essential, and a part of Fe can be substituted with Co or Ni, and additive elements such as Cu, Al, Ti, V, Cr, Ga, Mn, Nb, Ta, Mo, W, Ca, Sn, Zr, Hf, etc. may be added alone or in combination. B is boron, and a part thereof can be substituted with C or N.

The R-T-B alloy has a crystal composed of an R ₂ T ₁₄ B phase, which is a ferromagnetic phase that contributes to the magnetization action, as a main phase, and is a non-magnetic, rare earth element-concentrated low melting point R (rare earth containing Y). An alloy in which a rich phase coexists with one or more elements, and is an active metal, so that melting or casting is generally performed in a vacuum or an inert gas. Moreover, in order to produce a sintered magnet from a cast R-T-B type alloy lump by powder metallurgy, the alloy lump is pulverized to about 3 μm (FSSS: measured with a Fischer sub-sieve sizer) to obtain an alloy powder. After that, it is press-molded in a magnetic field, sintered at a high temperature of about 1000 to 1100 ° C. in a sintering furnace, then heat-treated and machined as necessary, and further plated to improve corrosion resistance and sintered. It is common to use a magnet.

In a sintered magnet made of such an R-T-B alloy, the R-rich phase plays an important role as follows.
(1) The melting point is low, and it becomes a liquid phase at the time of sintering, which contributes to increasing the density of the magnet and thus improving the magnetization.
(2) Eliminate grain boundary irregularities, reduce reverse domain nucleation sites and increase coercivity.
(3) The main phase is magnetically insulated to increase the coercive force. Therefore, if the dispersion state of the R-rich phase in the molded magnet is poor, local sintering failure and decrease in magnetism may occur. Therefore, it is important that the R-rich phase is uniformly dispersed in the molded magnet. It becomes. Here, the distribution of the R-rich phase is greatly influenced by the structure of the R-T-B type alloy ingot at the time of casting.

  Another problem that arises in the casting of an R-T-B alloy is that α-Fe is generated in the cast alloy ingot. α-Fe degrades the grinding efficiency when grinding the alloy lump, and if it remains in the magnet after sintering, it causes a reduction in the magnetic properties of the magnet. Therefore, in the conventional alloy lump, the homogenization treatment is performed at a high temperature for a long time as necessary to eliminate α-Fe.

  In order to solve the problem that α-Fe is formed in the cast R-T-B alloy ingot, a strip casting method (hereinafter abbreviated as SC method) is a method for casting the alloy ingot at a higher cooling rate. Have been developed and used in actual processes. The SC method is a method in which a molten metal is poured onto a copper roll whose inside is water-cooled, and a thin piece of about 0.1 to 1 mm is cast to rapidly cool and solidify the alloy. Precipitation of Fe can be suppressed. Furthermore, since the crystal structure of the alloy lump can be refined, an alloy having a structure in which the R-rich phase is finely dispersed can be generated. In this way, the alloy cast by the SC method has a fine dispersion of the R-rich phase inside, so the dispersibility of the R-rich phase in the magnet after pulverization and sintering is improved, and the magnet The magnetic characteristics have been successfully improved (for example, see Patent Document 1 and Patent Document 2).

  In addition, the alloy ingot cast by the SC method has excellent structure homogeneity. The homogeneity of the structure can be compared with the crystal grain size and the dispersion state of the R-rich phase. In alloy flakes produced by the SC method, chill crystals may occur on the casting roll side of the alloy flakes (hereinafter referred to as the mold surface side), but as a whole, a moderately fine and homogeneous structure brought about by rapid solidification Can be obtained.

  As described above, the RTB-based alloy cast by the SC method has an R-rich phase finely dispersed and the production of α-Fe is suppressed, so when producing a sintered magnet, The homogeneity of the R-rich phase in the final magnet is increased, and the adverse effects on pulverization and magnetism due to α-Fe can be prevented. Thus, the RTB-based alloy ingot cast by the SC method has an excellent structure for producing a sintered magnet. However, as the properties of magnets are improved, it is increasingly required to improve the structure homogeneity of the raw material alloy ingot.

Therefore, a technique for producing a sintered magnet with good magnet characteristics by setting the area ratio of the chill crystal on the mold surface side of the cast RTB-based alloy to 5% or less is disclosed. (For example, refer to Patent Document 3). Since the chill crystal part becomes a fine powder having a particle size of 1 μm or less in the pulverization step, it is considered that the particle size distribution of the alloy powder is disturbed and the magnetism is deteriorated.
JP-A-5-222488 Japanese Patent Laid-Open No. 5-295490 JP-A-10-317110

  As a result of studying the relationship between the structure of the cast R-T-B system alloy lump and the behavior during hydrogen crushing and fine pulverization, the present inventor has made the alloy powder particle size for sintered magnets uniform. In order to control, it has been found that controlling the dispersion state of the R-rich phase is more weight than the crystal grain size of the alloy lump. The volume ratio of chill crystals in the alloy lump is actually several percent or less, and the dispersion state of the R-rich phase formed on the mold surface side in the alloy lump is extremely finer than the harmful effects of chill crystals. It has been found that the region (fine R-rich phase region) has a greater influence for controlling the particle size of the magnet powder. That is, even when the amount of chill crystals in the RTB-based alloy ingot is reduced depending on the alloy ingot composition and production conditions, the volume ratio of the fine R-rich phase region may exceed 50%. It was confirmed that the rich phase region disturbed the particle size distribution of the magnet alloy powder, and it was confirmed that it was necessary to reduce the fine R rich phase region in order to improve the magnet characteristics.

The present invention has been made on the basis of the above knowledge, and its object is to generate a fine R-rich phase region in a rare earth-containing alloy lump made of a cast R-T-B alloy or the like. An object of the present invention is to provide a method for producing a rare earth-containing alloy flake capable of effectively suppressing the above, and a rare earth-containing alloy flake produced by such a production method and having a structure excellent in homogeneity.
It is another object of the present invention to provide an effective evaluation method for evaluating the metal structure of such a rare earth-containing alloy.

  The present inventor previously described an R-T-B system alloy (where R is at least one rare earth element including Y, T is a transition metal in which Fe is essential, and B is boron). As a result of investigating the relationship between the structure of the lump and the behavior during hydrogen crushing and pulverization, in order to uniformly control the particle size of the alloy powder for the sintered magnet, R is more than the crystal grain size of the alloy lump. We found that it is important to control the dispersion state of rich phase. Then, the casting rate in the SC method, particularly the surface state of the rotary roll for casting, was changed, and the volume ratios produced by the fine R-rich phase regions in the R-T-B type alloy flakes were compared. Then, it was found that there is a relationship between the surface roughness of the mold side surface of the alloy flake and the volume ratio at which the fine R-rich phase region is generated, and the alloy flake having excellent structure homogeneity with the fine R-rich phase region being 20% or less. Made possible.

In addition, the present inventor is able to homogenize the structure more effectively by giving a plurality of linear irregularities forming the surface roughness of the rotary roll for casting so as to intersect each other. confirmed.
In other words, the previous method can reduce the fine R-rich phase region, homogenize the structure, and control not only the numerical surface roughness of the rotating roll, but also the shape of the unevenness that is the origin of the surface roughness. In this case, it was confirmed that the tissue can be homogenized more effectively.

The present invention was made on the basis of the knowledge obtained by the previous research and the subsequent research, and the cast RT-T- Method for producing rare earth-containing alloy flakes having a structure excellent in homogeneity by effectively suppressing generation of fine R-rich phase region in B-based alloy ingot, and homogeneity of distribution of R-rich phase in magnet Therefore, the present invention provides a rare earth sintered magnet having high magnetic properties and excellent magnet characteristics.
The inventor has also established a metal structure evaluation method for measuring the R-rich phase interval in the cast rare earth magnet alloy and measuring the fine R-rich phase to determine the distribution state of the R-rich phase.

That is, the present invention
(1) The microscopic structure of the cross section of the R-T-B system magnet alloy is captured as a digital image, a straight line having a length (L) substantially parallel to the solidified surface of the alloy is drawn on the microscopic structure image, and R intersecting with the straight line Both phases are discriminated from the difference in luminance between the ₂ T ₁₄ B main phase and the R rich phase, and the luminance on the straight line changes from the luminance determined to be the R ₂ T ₁₄ B main phase to the luminance determined to be the R rich phase. The metal structure evaluation method of the R-T-B system magnet alloy which evaluates the value (L / N) which remove | divided the length (L) of the straight line by the frequency | count (N) to perform as an R rich phase space | interval (r).
(2) The average value and standard deviation of the R-rich phase interval are calculated from the R-rich phase interval obtained from each straight line using a plurality of the straight lines. The RTB-based magnet alloy according to (1) Metal structure evaluation method.
(3) The microscopic structure of the cross section of the R-T-B system magnet alloy is captured as a digital image, the microscopic structure image is divided by cells of an appropriate area, and the R ₂ T ₁₄ B main phase and R are determined from the difference in luminance within each cell. R− which discriminates both phases of the rich phase and determines whether or not each cell is in a fine R-rich phase region by measuring the shape, size, and number of individual R-rich phases present in each cell The metal structure evaluation method of TB magnet alloy.
(4) As the shape of the R-rich phase, any one or more of the roundness, aspect ratio, and rectangular aspect ratio of the structure is used. Metal structure evaluation of the R-T-B system alloy according to (3) Method.
(5) A cell in which the roundness of the R-rich phase is 1.0 to 1.4 and the area of the R-rich phase having an area of 5 square micrometers or less is 20 or more in 100 square micrometers. The metal structure evaluation method for an RTB-based alloy according to (4), wherein the portion is determined to be a fine R-rich phase region.

(6) Cell portion in which 20 or more R-rich phases having an aspect ratio of R-rich phase of 1.0 to 5.0 and an area of 5 square micrometers or less are contained in 100 square micrometers The metal structure evaluation method of the R-T-B type alloy according to (4), in which is determined as a fine R-rich phase region.
(7) A portion in which 20 or more R-rich phases having a rectangular aspect ratio of 1.0 to 5.0 and an area of 5 square micrometers or less are included in 100 square micrometers. The metal structure evaluation method of the R-T-B type alloy according to (4), in which is determined as a fine R-rich phase region.

(8) The cells obtained by dividing the microstructure image are divided into lattice-like cells by a group of equally spaced lines substantially parallel to the solidified surface of the alloy and a group of equally spaced lines substantially perpendicular to the solidified surface. ) To (7), when the cell determined to be a fine R-rich phase region discontinuously exists in the direction perpendicular to the solidified surface of the alloy. The metal structure evaluation method for an R-T-B alloy according to (3), wherein all the cells from the cell farthest from the solidification surface to the solidification surface are determined as the fine R-rich phase region.
(9) The cells obtained by dividing the microscopic structure image are divided into lattice-like cells by a group of equally spaced lines substantially parallel to the solidified surface of the alloy and a group of equally spaced lines substantially perpendicular to the solidified surface. When the cells determined as the fine R-rich phase region by the metallographic evaluation method according to any one of 3) to (7) discontinuously exist in a direction perpendicular to the solidified surface of the alloy, Among the cells determined to be the fine R-rich phase region, all the cells from the two or more continuous cells farthest from the solidified surface to the solidified surface are determined as the fine R-rich phase region, and the two or more consecutive cells The metal structure evaluation method for an RTB-based alloy according to (3), in which it is determined that all cells present alone in the free plane direction are not in the fine R-rich phase region.
(10) The total area of the cells determined as the fine R-rich phase region by the metal structure evaluation method according to any one of (3) to (9) is divided by the total area of all measured cells. The structure evaluation method of the R-T-B system alloy which uses the value as the area ratio of the fine R-rich phase region.
(11) The metal structure evaluation method for an RTB-based magnet alloy according to (1) or (3), wherein the digital image is a reflected electron beam image of a scanning electron microscope.
(12) The method for evaluating a metal structure of an R-T-B system magnet alloy according to (1) or (3), wherein a threshold value is provided for the luminance difference to discriminate between an R ₂ T ₁₄ B main phase and an R rich phase. .
It is.

According to the method for producing a rare earth-containing alloy flake of the present invention, a rare earth alloy in which the volume fraction of the fine R-rich region is small and the homogeneity of the dispersion state of the R-rich phase in the alloy is even better than conventional SC materials. The flakes can be manufactured with good reproducibility. Thereby, the sintered magnet manufactured from the obtained alloy flake and the bonded magnet by HDDR method express the magnet characteristic superior to the conventional one.
Moreover, according to the metal structure evaluation method of the present invention, it is possible to objectively determine the crystallization and distribution state of the R-rich phase, and it is easy to obtain a rare earth magnet alloy exhibiting an optimum metal structure for each application. Has the advantage of

FIG. 1 shows a backscattered electron image when a cross section of a thin piece of Nd—Fe—B alloy (Nd: 31.5 mass%) cast by the conventional SC method is observed with an SEM (scanning electron microscope). In FIG. 1, the left side is the mold surface side, and the right side is the free surface side.
The white portion in FIG. 1 is the Nd-rich phase (R-rich phase is called Nd-rich phase because R is Nd), from the center of the alloy flake to the free surface side (opposite to the casting surface side) (Surface on the side) forms a small pool with a lamellar shape in the thickness direction or a shape with a directionality that the lamellar is divided. However, a region in which the Nd-rich phase is extremely finer than other portions and randomly exists is formed on the mold surface side, and the present inventors have created a fine R-rich phase region (R When the main component of Nd is Nd, it is also called a fine Nd-rich phase region), and is particularly distinguished. This fine R-rich phase region usually starts from the mold surface side and extends toward the center. On the other hand, a portion where the fine R-rich phase region from the central portion to the free surface side does not exist is referred to herein as a normal portion.

  In the hydrogen crushing step of the RTB-based alloy flakes during the production of the sintered magnet, hydrogen is absorbed from the R-rich phase and expands into a brittle hydride. Therefore, in the hydrogen cracking, fine cracks along the R-rich phase or starting from the R-rich phase are introduced into the alloy. In the subsequent pulverization step, the alloy breaks due to a large amount of fine cracks generated by hydrogen cracking, so that the finer the dispersion of the R-rich phase in the alloy, the finer the particle size after pulverization tends to be. Therefore, the fine R-rich phase region has a tendency to break finer than the normal part. For example, in an alloy powder manufactured from the normal part, the average particle size is about 3 μm as measured by a FSSS (Fischer Sub-Seeb Sizer). On the other hand, the alloy powder manufactured from the fine R-rich phase region has a high ratio of containing fine powder of 1 μm or less, and therefore the particle size distribution after fine pulverization becomes wide.

  The dispersion state of the R-rich phase in the R-T-B alloy can be controlled by controlling the cooling rate after the molten metal solidifies during casting, or by heat treatment. It is described in Kaihei 10-36949. However, the cooling rate after solidification or the behavior of the change of the R-rich phase inside the fine R-rich phase region due to heat treatment is difficult to control unlike the normal part, and the dispersion of the R-rich phase is difficult to become coarse and remains fine. It is.

The volume ratio of the fine R-rich phase region can be measured by the following method. FIG. 3 is a reflected electron beam image having the same field of view as FIG. 1, but a line is drawn at the boundary between the fine R-rich phase region and the normal part. The boundary between both regions can be easily determined from the dispersion state of the R-rich phase, and the area ratio of the fine R-rich phase region can be calculated. The area ratio in the cross section of the visual field corresponds to the volume ratio in the alloy. Further, if an image analysis apparatus is used, it can be obtained quickly and uniquely.
This point will be described in detail later as a metallographic evaluation method, and an outline thereof will be described here.
The R-rich phase in the fine R-rich phase region is different from the normal part in shape, size, and density. That is, if the density of the R-rich phase that is fine and close to a spherical shape is more than a certain level, it can be determined as a fine R-rich phase region. The threshold varies depending on the alloy composition and the image quality of the reflected electron beam image used for image analysis. However, for each R-rich phase, the inventors have an R-rich phase density of 20 to 100 square microns with a roundness of 1 to 1.4 and an area of 5 square microns or less. If there is a part that does not meet the reference on the mold surface side of the part that meets or exceeds this part, or a part that does not meet the reference, it is determined that it is a fine R-rich phase region including that part, and reproducibility is good. It has been confirmed that a fine R-rich phase region can be measured. Here, the roundness is a value obtained by dividing (the square of the perimeter of the target figure) by (4π × area), and becomes 1 for a circle, and increases as the shape becomes elongated.
Furthermore, it was confirmed that the volume fraction of the fine R-rich phase region calculated in this way can well explain the characteristics of the alloy structure, the particle size distribution after pulverization, and the magnetic properties of the sintered magnet. ing.
In the measurement of the volume fraction of the fine R-rich phase region, the change in the amount of the fine R-rich phase region is large even between the thin pieces and within the same flake, even if the alloy thin pieces are cast at the same time. Therefore, the volume ratio of the fine R-rich phase region of the entire alloy can be calculated by measuring about 5 to 10 or more thin pieces and taking the average.

FIG. 2 shows a reflected electron beam image of a cross section of the RTB-based alloy flake (Nd 31.5 mass%) of the present invention. In FIG. 2, the left side is the mold surface side, and the right side is the free surface side. A feature of the alloy flake manufacturing method of the present invention is that the surface roughness of the rotary roll for casting is brought about by a plurality of linear irregularities formed on the casting surface in the method of manufacturing a rare earth-containing alloy flake by the strip casting method. Many of them are to use a rotating roll for casting having an angle of a certain level or more with respect to the rotating direction of the casting roll. The alloy flakes shown in FIG. 2 are produced using a casting rotary roll having a casting surface having the above-described configuration and a surface roughness of 3.5 μm. As shown in FIG. 2, there is no fine R-rich phase region on the mold surface side (left side in FIG. 2), and the dispersion state of the R-rich phase is extremely homogeneous from the mold surface to the free surface.
On the other hand, the alloy flakes in FIG. 1 are produced using a casting rotary roll in which linear irregularities are formed on the casting surface so as to extend substantially parallel to the roll rotation direction. The volume ratio of the fine R-rich phase region on the left side of the figure is larger than that of the alloy flakes shown in FIG.

The relationship between the direction of linear irregularities on the surface of the casting rotary roll and the fine R-rich phase region of the alloy flakes in the strip casting method can be explained as follows.
When the surface of the casting roll is smooth and the wettability with the molten alloy is good, the heat transfer from the molten metal to the mold is very good (the heat transfer coefficient is large), and the mold surface side of the alloy is excessively quenched. The The fine R-rich phase region has a large heat transfer coefficient between the mold and the molten metal, and it is considered that the fine R-rich phase region tends to be generated when the mold surface side of the alloy is excessively cooled.

  On the other hand, when fine irregularities are formed on the surface of the casting rotary roll, the surface roughness of the casting rotary roll surface increases. Then, because of the viscosity of the molten alloy, the molten alloy cannot completely enter the fine irregularities on the surface of the rotary roll for casting, resulting in an uncontacted portion and a reduced heat transfer coefficient. As a result, it is considered that the mold surface side of the alloy is not excessively cooled, and generation of a fine R-rich phase region can be suppressed.

  Here, paying attention to the shape of the unevenness, when it is linear, the contact portion and non-contact portion of the molten metal and the roll tend to exist continuously along the linear unevenness. Therefore, the internal structure also tends to have continuity along the unevenness. In that case, if a fine R-rich phase region is generated for some reason in a certain linear uneven portion, the probability that a fine R-rich phase region is generated over the entire linear portion becomes extremely high.

However, even with the same linear unevenness, the extending direction of the unevenness intersects the rotational direction of the casting rotary roll as in the rotating roll for casting applied to the manufacturing method of the present invention. When it was formed, it was confirmed that the continuity of the internal structure of the alloy flakes along the extending direction of the unevenness can be reduced, and at the same time, the generation of the fine R-rich phase region is suppressed.
The relationship between the direction of linear irregularities formed on the casting surface and the internal structure formed in the alloy flakes is considered as follows. That is, when the extending direction of the unevenness is parallel or close to the rotating direction of the rotating roll, when the molten metal comes into contact with the surface of the rotating roll and enters the uneven space, the atmosphere gas in the recessed portion is pushed out in the rotating direction. As a result, the contact area between the surface of the rotating roll and the molten metal increases.
On the other hand, when the extending direction of the unevenness is in a direction intersecting with the roll rotation direction, the atmosphere gas in the recesses on the surface of the rotary roll tends to be trapped by the molten metal as the angle between the two increases. Becomes stronger, and excessive contact between the roll surface and the molten metal can be suppressed. Therefore, even if the surface roughness of the casting surface is the same, the generation of the fine R-rich phase region tends to be suppressed as the angle between the extending direction of the linear unevenness and the roll rotation direction increases. . In addition, the continuity of the internal structure along the extending direction of the unevenness is also lower than the case where the rotation direction and the extending direction of the unevenness match, so that even if a fine R-rich phase region is generated, it can be expanded. The nature becomes low.

In the present invention, the surface roughness of the casting roll surface is 10 μm or less in terms of 10-point average roughness (Rz), and the surface roughness of the casting surface is caused by a plurality of substantially linear irregularities formed on the casting surface. 30% or more of the substantially linear unevenness is formed so as to extend in a direction that forms an angle of 30 ° or more with the roll rotation direction. .
Preferably, 30% or more of the linear projections and depressions are formed so as to extend in a direction that forms an angle of 45 ° or more with the roll rotation direction, or 50% or more of the linear projections and depressions. A casting rotary roll is used in which the projections and depressions extend in a direction that forms an angle of 30 ° or more with the roll rotation direction.
More preferably, a casting rotary roll is used in which 50% or more of the linear irregularities extend in a direction that forms an angle of 45 ° or more with the roll rotation direction.
Here, in the present invention, the angle between the extending direction of the unevenness and the roll rotation direction is 0 ° when the extending direction of the unevenness is parallel to the roll rotating direction, and the extending direction of the unevenness is parallel to the roll width direction. Is defined as 90 °.
As a result, even if the surface roughness is smaller than when the linear unevenness is parallel to the roll rotation direction, the effect of homogenizing the structure by reducing the fine R-rich phase can be obtained. This makes it possible to reduce the amount of cutting when adjusting the roll surface and increase the life of the casting roll. Further, since it is difficult to be affected by the difference in surface roughness, it is possible to simplify the management standard of the roll surface state.

Even in the conventional SC method, alloy flakes having a homogeneous structure as shown in FIG. 2 are included to some extent, but flakes containing a large amount of fine R-rich phase regions as shown in FIG. 1 are also generated at the same time. As a result, there was a problem in the homogeneity of the structure throughout the alloy. Such variations in the alloy structure produced by the conventional SC method are considered to be caused by the difference in the contact state between the roll surface and the molten metal, such as the subtle surface condition of the rotating roll for casting, the supply state of the molten metal, and the atmosphere. It is done.
The unevenness of an appropriate size on the surface of the casting rotary roll can suppress excessive heat transfer when the molten metal solidifies, and can suppress generation of a fine R-rich phase region with good reproducibility. Furthermore, in this invention, the extending direction of the linear unevenness | corrugation of the roll surface is formed with an angle with respect to the rotation direction of the rotary roll for casting, and the fine R rich phase area | region production | generation suppression effect is heightened. Thus, the effect can be obtained even with a relatively small surface roughness. As a result, the yield of the alloy flakes having a homogeneous structure as shown in FIG. 2 can be further increased.
In addition, the linear unevenness | corrugation of this invention does not necessarily need to be continuous, and may be intermittent. Further, it may be curved or linear. Usually, the roll material is preferably pure copper or a copper alloy. Further, the present invention is possible even when the casting surface layer is coated.

Further details of the present invention will be described.
(1) Strip casting method The present invention relates to a rare earth-containing alloy flake. Here, casting of the R-T-B alloy by the strip casting method will be described.
FIG. 4 shows a schematic view of an apparatus for casting by strip casting. Usually, the RTB-based alloy is melted with the refractory crucible 1 in a vacuum or an inert gas atmosphere because of its active properties. The molten alloy melt is kept at a temperature of 1350-1500 ° C. for a predetermined time, and if necessary, the casting rotary roll 3 is cooled with water through a tundish 2 provided with a rectifying mechanism and a slag removing mechanism. To be supplied. The supply speed of the molten metal and the number of rotations of the rotating roll are appropriately controlled according to the desired thickness of the alloy. Generally, the rotational speed of the rotating roll is about 0.5 to 3 m / s in terms of peripheral speed. The material of the rotary roll for casting is suitably copper or a copper alloy from the viewpoint of good thermal conductivity and easy availability. Depending on the material of the rotating roll and the surface condition of the roll, metal easily adheres to the surface of the rotating roll for casting. Therefore, if a cleaning device is installed as necessary, the quality of the cast R-T-B alloy is stable. To do. The alloy 4 solidified on the rotating roll separates from the roll on the opposite side of the tundish and is collected in the collection container 5. By providing a heating and cooling mechanism in the collection container, the state of the R-rich phase tissue in the normal part can be controlled.

  The thickness of the alloy flakes produced by the production method according to the present invention is preferably 0.1 mm or more and 0.5 mm or less. If the thickness of the alloy flakes is less than 0.1 mm, the solidification rate increases excessively, the crystal grain size becomes too fine, and it becomes close to the finely pulverized grain size in the magnetizing process. There is a problem of inviting. On the other hand, if the thickness of the alloy flakes is larger than 0.5 mm, problems such as a decrease in R-rich phase dispersibility due to a decrease in solidification rate and precipitation of α-Fe are caused.

(2) Surface Roughness of Casting Surface of Casting Rotating Roll In the present invention, when casting an R-T-B magnet alloy by the strip casting method, the surface roughness of the casting surface of the rotating roll for casting is sufficiently increased. The point average roughness (Rz) is 3 μm or more and 60 μm or less.
Here, the surface roughness is measured under the conditions shown in JIS B 0601 “Definition and Display of Surface Roughness”, and ten-point average roughness (Rz) is also defined therein. Specifically, first, a curve (roughness curve) obtained by removing a surface waviness component longer than a predetermined wavelength with a phase compensation type high-pass filter or the like from a cut surface (cross-sectional curve) obtained by cutting along a plane perpendicular to the measurement surface. Ask. From the roughness curve, a reference length is extracted in the direction of the average line, and from the average line of the extracted part, the average value of the absolute values of the altitude (Yp) of the highest peak from the highest peak to the fifth peak is the lowest. The sum of the absolute values of the elevations (Yv) of the valley bottom from the valley bottom to the fifth is called ten-point average roughness (Rz). For the measurement parameters such as the reference length, standard values for the surface roughness are specified in the above JIS.
The surface roughness may vary greatly, and at least 5 points should be measured and the average value should be used.

(3) Shape of irregularities on the roll casting surface The surface roughness is mainly caused by a plurality of linear irregularities formed on the roll surface, and many of the linear irregularities are somewhat in the direction of rotation of the casting roll. It is formed to extend in the direction that forms the angle. Specifically, 30% or more of the linear irregularities have an angle of 30 ° or more from the roll rotation direction. Preferably, 30% or more of the linear irregularities have an angle of 45 ° or more from the roll rotation direction, or 50% or more of the linear irregularities have an angle of 30 ° or more from the roll rotation direction. It is to have. More preferably, 50% or more of the linear irregularities have an angle of 45 ° or more from the roll rotation direction.
When the extending direction of the linear irregularities is parallel to or close to the rotation direction, the molten metal comes into contact with the roll surface, and when it enters the gaps between the irregularities on the surface, it becomes easy to push out the atmospheric gas in the concave parts, and the roll surface Increases the contact area. However, as the angle between the extending direction of the unevenness and the rotation direction increases, the tendency that the atmosphere gas in the recess is trapped by the molten metal increases, and excessive contact with the roll surface can be suppressed.
Therefore, even if the surface roughness is the same, as the angle between the extending direction of the linear unevenness and the roll rotating direction becomes larger, or the linear unevenness having a large angle with the roll rotating direction as such. As the number increases, the generation of the fine R-rich phase region is suppressed. In addition, since the continuity of the texture in the uneven direction is also lower than when the rotation direction and the extending direction of the unevenness coincide with each other, the possibility that the fine R-rich phase region is expanded is reduced.

More than 30% of the plurality of linear irregularities formed on the casting surface has an angle of 30 ° or more from the roll rotation direction, thereby effectively suppressing the fine R-rich region described above. Can do. Less than 30% is less effective.
The linear irregularities on the surface of the casting rotary roll according to the present invention are not necessarily continuous, and may be intermittent. Further, it may be curved or linear.
Such linear irregularities can be applied to a polishing direction and a roll rotation direction even in a device having a rotating polishing paper, a rotating wire brush, or a belt polishing device in which the polishing paper moves linearly. Can be formed by providing an angle.

In the present invention, the effect of homogenizing the structure can be obtained even if the surface roughness is relatively small due to the effect of linear irregularities extending at an angle with respect to the rotation direction of the casting roll.
However, if the surface roughness is 3 μm or less, the effect of unevenness on the surface of the rotary roll for casting cannot be obtained, and the heat transfer coefficient is large because of good contact with the molten metal. As a result, it becomes easy to generate a fine R-rich phase region in the alloy.
Further, when the surface roughness of the casting rotary roll exceeds 60 μm, the alloy flakes bite into the roll surface, making it difficult to peel off, which may cause troubles such as destroying the tundish. Therefore, the surface roughness of the casting rotary roll is 60 μm or less.

(4) Volume ratio of fine R-rich phase region in alloy When producing an R-T-B type alloy by the production method of the present invention, the volume fraction of fine R-rich phase region in the alloy should be 20% or less. Is preferred. As a result, since the particle size distribution of the finely pulverized alloy powder is narrowly aligned in the sintered magnet process, a homogeneous sintered magnet with no variation in characteristics can be obtained.

(5) Rare earth sintered magnet alloy powder, method for producing rare earth sintered magnet Rare earth-containing alloy flakes made of R-T-B alloy used as a raw material for magnets cast according to the present invention are pulverized, molded, Through a sintering process, an anisotropic sintered magnet having high characteristics can be manufactured.

The alloy flakes are usually pulverized in the order of hydrogen pulverization and fine pulverization to produce an alloy powder of about 3 μm (FSSS).
Here, hydrogen cracking is divided into a hydrogen storage process in the previous process and a dehydrogenation process in the subsequent process. In the hydrogen storage step, hydrogen is mainly stored in the R-rich phase of the alloy flakes in a hydrogen gas atmosphere at a pressure of 266 hPa to 0.3 MPa · G, and the R-rich phase is formed by the R-hydride generated at this time. Utilizing the volume expansion, the alloy flakes themselves are finely divided or innumerable fine cracks are generated. This hydrogen occlusion is carried out in the range of room temperature to about 600 ° C. In order to increase the volume expansion of the R-rich phase and efficiently divide it, the pressure in the hydrogen gas atmosphere is increased and the room temperature is about 100 ° C. It is preferable to implement in the range. A preferred treatment time is 1 hour or more. Since the R-hydride generated by this hydrogen storage process is unstable in the atmosphere and easily oxidized, dehydration is performed by holding the alloy flakes in a vacuum of about 1.33 hPa or less at about 200 to 600 ° C. after the hydrogen storage treatment. It is preferable to perform a raw treatment. By this treatment, it can be changed to R-hydride stable in the atmosphere. A preferable treatment time for the dehydrogenation treatment is 30 minutes or more. In the case where atmosphere management for preventing oxidation is performed in each process from hydrogen storage to sintering, dehydrogenation treatment can be omitted.

  The RTB-based alloy flakes produced by the strip casting method of the present invention are characterized in that the R-rich phase is uniformly dispersed. A preferable average value of the R-rich phase interval depends on the pulverized particle size in the magnet manufacturing process, but is generally 3 μm to 8 μm. In hydrogen cracking, cracks are introduced along the R-rich phase or starting from the R-rich phase. Therefore, by crushing after hydrogen cracking, it is possible to maximize the effect of the R-rich phase uniformly and finely dispersed in the alloy, and the alloy powder with a very narrow particle size distribution can be efficiently obtained. It is possible to produce. When a sintered magnet is manufactured without performing this hydrogen crushing step, the characteristics of the manufactured sintered magnet are inferior. (M. Sagawa et al. Proceeding of the 5th International Conference on Advanced Materials, Beijing China (1999)).

The fine pulverization is to pulverize the RTB-based alloy flakes to about 3 μm (FSSS). As a pulverizing apparatus for fine pulverization, a jet mill apparatus is most suitable because of good productivity and a narrow particle size distribution. By using the alloy flakes having a small fine R-rich phase region of the present invention, an alloy powder having a narrow particle size distribution can be produced with high efficiency and good stability.
The atmosphere when pulverizing is an inert gas atmosphere such as argon gas or nitrogen gas. 2% by mass or less, preferably 1% by mass or less of oxygen may be mixed into these inert gases. As a result, the grinding efficiency is improved, the oxygen concentration of the ground alloy powder can be 1000 to 10,000 ppm, and the alloy powder can be appropriately stabilized. At the same time, abnormal growth of crystal grains when the magnet is sintered can be suppressed.

  When molding the above alloy powder in a magnetic field, the powder is lubricated with zinc stearate or the like to reduce the friction between the alloy powder and the inner wall of the mold and to improve the orientation by reducing the friction between the powders. It is preferable to add an agent. A preferable addition amount is 0.01 to 1% by mass. The lubricant may be added before or after fine pulverization, but before forming in a magnetic field, it is preferable to sufficiently mix using a V-type blender or the like in an inert gas atmosphere such as argon gas or nitrogen gas.

The alloy powder pulverized to about 3 μm (FSSS) is press-molded with a molding machine in a magnetic field. The mold is manufactured by combining a magnetic material and a non-magnetic material in consideration of the direction of the magnetic field in the cavity. The molding pressure is preferably 0.5 to 2 t / cm2. The magnetic field in the cavity at the time of molding is preferably 5 to 20 kOe. Further, the atmosphere during molding is preferably an inert gas atmosphere such as argon gas or nitrogen gas, but in the case of the above-mentioned oxidation-resistant powder, it can also be performed in the air.
The molding can also be performed by cold isostatic pressing (CIP) or pseudo isostatic pressing (RIP) using a rubber mold. Since CIP and RIP are compressed hydrostatically, there is less disorder in the orientation during molding, the orientation ratio can be increased as compared with mold molding, and the maximum magnetic energy product can be increased.

  Sintering of the molded body is performed at 1000 to 1100 ° C. As a sintering atmosphere, an argon gas atmosphere or a vacuum atmosphere of 1.33 × 10 −2 hPa or less is preferable. The holding time at the sintering temperature is preferably 1 hour or longer. In sintering, it is necessary to remove as much as possible hydrogen contained in the lubricant and alloy powder in the molded body before reaching the sintering temperature. A preferable removal condition of the lubricant is to hold at 300 to 500 ° C. for 30 minutes or more in a vacuum of 1.33 × 10 −2 hPa or less or in a reduced Ar flow atmosphere. Moreover, the preferable removal conditions of hydrogen are hold | maintaining for 30 minutes or more at 700-900 degreeC in the vacuum of 1.33 * 10 <-2> hPa or less.

  After the sintering is completed, heat treatment can be performed at 500 to 650 ° C. as necessary to improve the coercive force of the sintered magnet. A preferable atmosphere in this case is an argon gas atmosphere or a vacuum atmosphere, and a preferable holding time is 30 minutes or more.

  Moreover, the R-T-B type alloy flakes for rare earth magnets that suppress the production of the fine R-rich region produced by the production method of the present invention can be suitably used for producing bonded magnets in addition to sintered magnets. it can. Below, the case where a bonded magnet is produced from the alloy flake produced with the manufacturing method of the alloy flake for rare earth magnets of this invention is demonstrated.

  The RTB-based alloy flakes of the present invention are first heat treated as necessary. The purpose of the heat treatment is to remove α-Fe from the alloy and to coarsen the grains. HDDR (Hydrogenation Deposition Recombination Recombination) processing is used to manufacture alloy powders for bond magnet manufacturing, but α-Fe present in the alloy cannot be erased in the HDDR processing process, causing the magnetism to decrease It becomes. Therefore, α-Fe needs to be erased before the HDDR process is performed.

  Moreover, the average particle diameter of the alloy powder for bonded magnets is 50 to 300 μm, which is very large compared to the alloy powder for sintered magnets. In the HDDR method, the crystal orientation of the original alloy coincides with the orientation of the recombined submicron crystal grains with a certain distribution. Therefore, when two or more crystal grains having different crystal orientations in the alloy flakes of the raw material are included in one bond magnet alloy powder, the alloy powder includes a region having greatly different crystal orientations. The orientation rate of the magnet decreases, and the maximum magnetic energy product decreases. In order to avoid this, it is convenient that the crystal grain size in the alloy flakes is large. In an alloy cast by a rapid solidification method such as a strip casting method, the crystal grain size tends to be relatively small, so that the coarsening of the crystal grain by heat treatment is effective in improving the magnet characteristics.

  There are many reports on a method for producing an alloy powder for bonded magnets by the HDDR method (for example, T. Takeshita et al, Proc. 10th Int. Worksshop on RE magnets and therair application, Kyoto, Vol. 1 p551 (1989)). ). Production of the alloy powder by the HDDR method is performed as follows.

  When the raw R-T-B alloy flakes are heated in a hydrogen atmosphere, the R2T14B phase of the magnetic phase is decomposed into three phases of α-Fe, RH2, and Fe2B at about 700 to 850 ° C. Next, when hydrogen is removed by switching to an inert gas atmosphere or a vacuum atmosphere at the same temperature, the decomposed phase recombines with the R2T14B phase having a crystal grain size of about submicron. At this time, if the alloy composition and processing conditions are appropriately controlled, the recombination easy axis of each R2T14B phase (R2T14B phase C axis) is substantially parallel to the C axis of the R2T14B phase in the raw alloy before decomposition, An anisotropic magnet powder in which the easy axis directions of magnetization of each fine crystal grain are aligned can be obtained.

  The HDDR-treated alloy is pulverized to about 50 to 300 μm to form an alloy powder, which is then mixed with a resin and subjected to compression molding, injection molding, or the like to obtain a bonded magnet.

  The fine R-rich phase region has a strong tendency to be pulverized during the HDDR treatment as in the above-described hydrogen cracking treatment. The characteristics of magnetic powder by the HDDR method decrease as the particle size decreases. Therefore, the RTB-based alloy that suppresses the formation of the fine R-rich phase of the present invention can be suitably used for the production of magnetic powder for bonded magnets in HDDR processing.

Recently, it has been reported that the homogeneity of the structure is improved by making the surface roughness of the outer peripheral surface of the roll within a certain range in the production of rare earth alloys (Japanese Patent Laid-Open No. 2002-59245, special feature). (Kaihei 9-1296). However, they are intended to suppress the change of the structure in the width direction of the strip and to prevent the cooling rate from decreasing at the end of the strip. Further, the shape of the unevenness that brings about the surface roughness is not particularly defined.
On the other hand, the present invention suppresses the structural change from the thickness direction of the alloy flakes, that is, from the roll surface side to the free surface side, and homogenizes the structure, and introduces a fine R-rich phase region as a measure of homogeneity. However, a specific volume ratio is clearly shown, which is completely different from Japanese Patent Application Laid-Open No. 2002-59245 or Japanese Patent Application Laid-Open No. 9-1296.
Japanese Patent Laid-Open No. 4-28457 discloses that the roughness of the roll surface is set to 0.05 to 1.5 μm in Ra to reduce the fluctuation of the crystal grain size. However, it relates to ultra-rapid cooling with a high roll speed, and the surface roughness is different in the standard of measurement, but it has been confirmed that Ra of the cast roll surface of the present invention is larger than that.

Next, the metal structure evaluation method of the present invention will be described.
The metal structure evaluation method of the present invention measures the R-rich phase interval in the rare earth magnet alloy and measures the fine R-rich phase region to determine the distribution state of the R-rich phase.
First, a metal structure of a material to be used for evaluation is observed with an electron microscope, and a reflected electron beam image is taken into a computer as a digital image. For example, it is suitable that the digital image has 1470 vertical x 1960 horizontal x 255 gradations of brightness. Image quality (brightness, contrast) is the most important for tissue evaluation.
After that, operate on the computer screen and measure.

[Measurement of R-rich phase interval]
To measure the distance between the R-rich phase, draw a straight line L ₁ of the length L on a digital image of the reflected electron image as shown in FIG. Since the cast structure has directionality, the straight line L ₁ is preferably drawn parallel to the solidified surface 10. Then, count the number N of intersections P of the straight line L ₁ and the R-rich phase 12 in the metal structure. Since the R-rich phase appears bright on the screen, a threshold is set for the brightness, and points that change from a point below the threshold to a point above the threshold are counted.
Thereafter, a value r obtained by dividing the length L of the straight line by the number N of points crossing the R-rich phase is defined as the R-rich phase interval. That is, r = L / N.
The R-rich phase usually extends in a needle shape from the solidified surface toward the free surface, and is finely distributed in the vicinity of the solidified surface as the distance from the solidified surface increases. Naturally, since the interval of the R-rich phase also varies depending on the location, a plurality of (L ₁ , L ₂ , L ₃ ,... L _N ) lines with a length L are drawn to obtain a large number of r values. The average value and standard deviation of the interval can also be obtained, and more accurate organization evaluation can be performed. A plurality of straight lines (L ₁ , L ₂ , L ₃ ,... L _N ) used at this time are also preferably drawn parallel to the solidified surface 10 as shown in FIG.

[Measurement of fine R-rich phase]
Next, measurement of the fine R-rich phase region will be described. The fine R-rich phase is a structure containing many fine and nearly spherical metal structures rich in rare earth elements (R), which are generated by excessive rapid solidification near the solidified surface. A digital image in the vicinity of the solidified surface is divided by a mesh, and the shape, size, and number of R-rich phases contained in each cell are measured, and the form and distribution state of the R-rich phase are evaluated. It is determined whether or not the cell is a fine R-rich phase region.
First, a digital image is divided into a large number of cells by dividing the digital image with a mesh composed of vertical and horizontal lines (x _n , y _n ). For example, in FIG. 6, straight lines (y ₁ , y ₂ ,... Y ₂₀ ) are drawn by dividing the screen into 20 equal parts. On the other hand, the straight line (x ₁ , x ₂ ,... X ₂₀ ) is drawn by dividing the measurement region into 20 equal parts because the region range to be measured varies from the solidified surface. The cell is divided into a total of 400 cells, and the shape, size, and number of R-rich phases in each cell are measured.
Here, the vertical and horizontal lines (x _n , y _n ) constituting the mesh are divided into lattice-like cells by a group of equally spaced lines substantially parallel to the solidified surface and a group of equally spaced lines substantially perpendicular to the solidified surface. Is preferred.

As shown in FIG. 7, since the R-rich phase 12 appears bright in the matrix of the dark main phase 11 on the digital image, whether or not each pixel of the digital image is in the R-rich phase by setting a threshold value for brightness. To determine. As shown in FIG. 6, the R-rich phase pixels in each cell divided by the mesh are counted, and the shape, size, and number of R-rich phases are measured.
Next, R rich phases satisfying the following conditions among those determined as R rich phases from the brightness are counted.
Condition 1): The R-rich phase is approximated to a circle, and the R-rich phase having a roundness of 1 to 1.4 is counted.
Here, the roundness is defined as (perimeter length ² ÷ (4π × area)).
Condition 2): Count the R-rich phase with an area of 5 square micrometers or less.
Condition 3): Count the R-rich phase that satisfies both condition 1) and condition 2).

Actually, the shape of the R-rich phase may be closer to a needle shape or a square shape than a circular shape. Therefore, it is also effective to use an aspect ratio or a rectangular aspect ratio instead of roundness.
For example, when the aspect ratio is used, the R-rich phase has an aspect ratio of 1.0 to 5.0 and an area of 5 square micrometers or less, and there are 20 or more R-rich phases in 100 square micrometers. The included cell portion is determined to be a fine R-rich phase region.
In addition, when the quadratic approximation is used, an R-rich phase having an R-rich phase square aspect ratio of 1.0 to 5.0 and an area of 5 square micrometers or less is 20 in 100 square micrometers. A portion including at least one is determined as a fine R-rich phase region.
In FIG. 6, the hatched portion of the cell is a cell determined to be a fine R-rich phase region.

[Judgment of fine R-rich phase region]
When the count number of the R-rich phase extracted under the condition 3) is 20 or more in 100 square micrometers, the cell is determined to be a fine R-rich phase region. In this manner, whether or not the 400 cells in FIG. 6 are in the fine R-rich phase region is determined and mapped. _{S 1,} S 2 of the hatched in FIG. 6, · · · _{S n} is the fine R-rich phase region.
For example, in the portion adjacent to the solidified surface in FIG. 6, there is a region that does not contain a fine and nearly spherical metal structure due to excessive quenching, and such a region is not determined to be a fine R-rich phase region.
However, the fine R-rich phase region is generated due to excessive rapid solidification and is generated from the cooling roll side. Therefore, if a fine R-rich phase region exists in a portion away from the solidified surface, the solidified surface side should naturally be a fine R-rich phase region.

One of the determination methods, for example as in the S ₆ from S ₁ shown in FIG. 6, when the cell where it is determined that the fine R-rich phase region is present discontinuously in a direction perpendicular to the solidification surface, solidification surface All cells from the cell farthest to the solidified surface should be determined as the fine R-rich phase region. That is, S ₁ ′, S ₂ ′, S ₃ ′ and S ₄ ′ shown in FIG. 6 which are not determined as the fine R-rich phase region are also regarded as the fine R-rich phase region, and the regions from S ₁ ′ to S ₆ are considered. The fine R-rich phase region is reasonable.
As another determination method, for example as in S ₁₂ from S ₇ shown in FIG. 6, when the cell where it is determined that the fine R-rich phase region is present discontinuously in a direction perpendicular to the solidification surface, Among the cells determined to be the fine R-rich phase region, all cells from the two or more continuous cells farthest away from the solidification surface to the solidification surface are judged as the fine R-rich phase region, and two or more continuous There is a method of determining that all the cells scattered in the free plane direction alone are not in the fine R-rich phase region. That is, S ₅ ′ and S ₆ ′ not determined as the fine R-rich phase region in FIG. 6 are regarded as the fine R-rich phase region, and S ₁₁ and S ₁₂ are not regarded as the fine R-rich phase region. Therefore, there is an advantage of preventing the fine R-rich phase region from being overestimated particularly in a fine structure including high-order dendrites of the R-rich phase. Such a structure includes a fine streak-like R-rich phase, but a criterion for determining an R-rich phase in which a fine higher-order dendrite partially agglomerated and torn is determined as a fine R-rich phase region. This is because rarely exists.

By employing any of the determination methods described above, the range enclosed by the thick line of S _n from S ₁₃ shown in FIG. 6, also comprise a cell which has not determined that even if the fine R-rich phase region, fine R-rich A cell that is regarded as a phase region and scattered outside the range surrounded by a thick line is not regarded as a fine R-rich phase region.
Further, a value obtained by dividing the total area ΣSn of the cells determined to be the fine R-rich phase region by the total area S of all the measured cells is defined as an area ratio Q of the fine R-rich phase region. That is, Q = ΣSn / S.
In this way, it is possible to evaluate how the fine R-rich phase region is distributed in the vicinity of the solidified surface of the rare earth magnet alloy.
Next, the present invention will be specifically described with reference to Examples and Comparative Examples.

The alloy composition is Nd: 31.5% by mass, B: 1.00% by mass, Co: 1.0% by mass, Al: 0.30% by mass, Cu: 0.10% by mass, and the balance iron. The raw materials blended with metal neodymium, ferroboron, cobalt, aluminum, copper, and iron are melted in a high-frequency melting furnace in an atmosphere of 1 atm with argon gas using an alumina crucible, and the molten metal is strip cast. The alloy flake was produced by casting.
The diameter of the casting rotary roll was 300 mm, the material was pure copper, the inside was water-cooled, and the surface roughness of the casting surface was adjusted to 4.0 μm in terms of 10-point average roughness (Rz). The surface roughness at this time was mainly caused by linear irregularities imparted to the casting surface, and 50% or more thereof had an angle of 45 ° or more from the roll rotation direction. The peripheral speed of the roll during casting was 1.0 m / s, and alloy flakes having an average thickness of 0.30 mm were produced.

After thirty of the obtained alloy flakes were embedded in a resin and polished, a backscattered electron beam image (BEI) at a magnification of 350 times was taken with a scanning electron microscope (SEM) for each alloy flake cross section. The taken photograph was taken in as a digital image, and the average R-rich phase interval and the fine R-rich phase region were measured with an image analyzer. Specifically, each measurement was calculated by the following method.
First, a method for measuring the R-rich phase interval will be described.
For each cross-sectional structure, straight lines approximately parallel to the alloy surface are drawn at 10 μm intervals, and the R-rich phase interval is determined for each straight line. At that time, the area within 10 μm from both surfaces was removed from the analysis because of polishing sagging. Therefore, it was possible to obtain the R-rich phase interval data for each cross-sectional structure from one tenth of the thickness to a few fewer.
The average R-rich phase interval was 4.9 μm as the average value of all R-rich phase intervals obtained by analysis of each cross-sectional structure.

Next, the variation of the R-rich phase interval was evaluated. Two types of methods were used to evaluate the variation.
One method was the standard deviation of all R-rich phase intervals used in the calculation of the average R-rich phase interval, which was 0.78 μm. This represents a variation in microscopic R-rich phase interval within each slab of the alloy. Here, this is called microstructural change.
In another method, first, an average value of the R-rich phase interval is obtained for each cross-sectional structure. In this example, the average R-rich phase interval is calculated for each of the 30 cross-sectional structures. Since the structure of each slab changes depending on the solidification rate or the cooling rate after solidification, the average R-rich phase interval between the slabs is also different. For this reason, the 30 average R-rich phase intervals also have some variation, which represents a macro-like variation in the R-rich phase interval between slabs and is referred to as a macroscopic structural change. The standard deviation of the 30 average R-rich phase intervals was 0.40 μm.

In the fine R-rich phase region, each cross-sectional structure was divided into 20 × 20 cells by a group of equally spaced lines substantially parallel to the surface and a group of equally spaced lines perpendicular to them. When the roundness of the R-rich phase existing in each cell is 1.0 to 1.4 and the area is 5 square micrometers or less, 20 or more are included in 100 square micrometers The mesh was defined as a fine R-rich phase region. The fine R-rich phase region was slightly present in the vicinity of the solidified surface, and the volume ratio was 1.7%.
In the determination of the fine R-rich phase region, a method in which 20 or more R-rich phases having an aspect of 1.0 to 5.0 and an area of 5 square micrometers or less are included in 100 square micrometers is used. The volume fraction of the fine R-rich phase region was 1.9%. Further, a fine R-rich phase region in a method in which 20 or more R-rich phases having a square aspect ratio of 1.0 to 5.0 and an area of 5 square micrometers or less are included in 100 square micrometers. The volume ratio was 1.6%.

(Comparative Example 1)
Raw materials were blended in the same composition as in Example 1, and dissolution and casting by the SC method were performed in the same manner as in Example 1. However, the surface roughness of the casting rotary roll surface is 4.0 μm in terms of ten-point average roughness (Rz), has linear irregularities substantially parallel to the roll rotation direction, and is inclined with respect to the rotation direction. Concavities and convexities were virtually absent.
As a result of evaluating the R-rich phase interval of the obtained alloy flakes in the same manner as in Example 1, the average R-rich phase interval was 5.8 μm, the microstructural change was 1.33 μm, and the macrostructural change was 0.88 μm. It was.
On the other hand, the fine R-rich phase region was also evaluated in the same manner as in Example 1. However, not only the solidified surface side but also the internal cells were determined to be the fine R-rich phase region. In addition, since the cells that are partially determined to be the fine R-rich phase region existed discontinuously, all the cells from the two or more consecutive cells farthest from the solidified surface to the solidified surface are treated as the fine R-rich phase region. The volume ratio was 25%.

(Comparative Example 2)
Raw materials were blended in the same composition as in Example 1, and dissolution and casting by the SC method were performed in the same manner as in Example 1. However, the surface roughness of the rotary roll for casting is 100 μm in terms of ten-point average roughness (Rz), and the surface roughness is mainly due to linear irregularities imparted to the casting surface as in Example 1. More than 50% of which had an angle of 45 ° or more from the roll rotation direction.
Then, a part of the metal did not detach from the roll during the casting, and contacted the tundish and destroyed its front end, so the casting operation was stopped.

Next, an example in which a sintered magnet was produced will be described.
The alloy flakes obtained in Example 1 were cracked with hydrogen and pulverized with a jet mill. The conditions of the hydrogen occlusion process, which is the previous process of the hydrogen crushing process, were maintained at 100% hydrogen atmosphere and 2 atm for 1 hour. The temperature of the metal piece at the start of the hydrogen storage reaction was 25 ° C. The conditions for the dehydrogenation process, which is a subsequent process, were maintained at 500 ° C. for 1 hour in a vacuum of 0.133 hPa. To this powder, 0.07% by mass of zinc stearate powder was added, thoroughly mixed with a V-type blender in a 100% nitrogen atmosphere, and then finely pulverized with a jet mill apparatus. The atmosphere during pulverization was a nitrogen atmosphere mixed with 4000 ppm of oxygen. Thereafter, the mixture was sufficiently mixed again with a V-type blender in a 100% nitrogen atmosphere. The oxygen concentration of the obtained powder was 2500 ppm, and it was calculated from the analysis of the carbon concentration of the powder that the zinc stearate powder mixed in the powder was 0.05% by mass. The average particle size D50 was 5.11 μm, D10 was 1.90 μm, and D90 was 8.60 μm as a result of measurement with a laser diffraction particle size distribution analyzer.

  Next, the obtained powder was press-molded in a transverse magnetic field molding machine in a 100% nitrogen atmosphere. The molding pressure was 1.2 t / cm 2 and the magnetic field in the mold cavity was 15 kOe. The obtained molded body was kept at 500 ° C. in a vacuum of 1.33 × 10 −5 hPa for 1 hour, and then kept at 800 ° C. in a vacuum of 1.33 × 10 −5 hPa for 2 hours. Sintering was performed at 1050 ° C. for 2 hours in a vacuum of 33 × 10 −5 hPa. The sintered density was 7.5 g / cm 3 or more, which was a sufficiently large density. Furthermore, this sintered body was heat-treated at 560 ° C. for 1 hour in an argon atmosphere to produce a sintered magnet.

  Table 1 shows the results of measuring the magnetic properties of the sintered magnet with a DC BH curve tracer. Table 1 also shows the oxygen concentration and particle size of the fine powder of the sintered magnet raw material.

(Comparative Example 3)
The alloy flakes obtained in Comparative Example 1 were pulverized in the same manner as in Example 2 to obtain fine powder.
Further, through the same molding and sintering steps as in Example 2, a sintered magnet was produced.

  Table 1 shows the results obtained by measuring the magnetic characteristics of the sintered magnet produced in Comparative Example 3 with a direct current BH curve tracer. Table 1 also shows the oxygen concentration and particle size of the fine powder of the sintered magnet raw material of Comparative Example 3.

  As shown in Table 1, since D10 is smaller in Comparative Example 3 than in Example 2, it can be seen that the proportion of very fine powder smaller than about 1 μm is large. Such very fine particles are easily oxidized, and the oxygen concentration of the fine powder is slightly higher in Comparative Example 3 than in Example 2. The reason why the magnetic characteristics of the magnet of Comparative Example 3 are lower than that of Example 2 is considered to be mainly due to an increase in oxygen concentration and inhomogeneity of the crystal structure.

Next, an example in which a bonded magnet was produced will be described.
The raw materials were blended so that the alloy composition was Nd 28.5%, B: 1.00% by mass, Co: 10.0% by mass, Ga: 0.5% by mass, and the balance iron. An alloy flake was cast by the SC method under conditions.
As a result of evaluating the obtained alloy flakes in the same manner as in Example 1, the volume ratio of the fine R-rich phase region was 3% or less, and α-Fe was not contained.

The alloy flakes described above were held in 1 atmosphere of hydrogen at 820 ° C. for 1 hour and then subjected to HDDR treatment in which the alloy flakes were held at the same temperature in vacuum for 1 hour. The obtained alloy powder was pulverized to 150 μm or less with a brown mill, 2.5% by mass of an epoxy resin was added, and a magnetic field of 1.5T was applied to perform compression molding to obtain a bonded magnet. Table 1 shows the magnetic properties of the obtained bonded magnet.
Shown in

(Comparative Example 4)
Raw materials were blended in the same composition as in Example 3, and dissolution and casting by the SC method were performed as in Comparative Example 1. As a result of evaluating the obtained alloy flakes in the same manner as in Example 1, the volume ratio of the fine R-rich phase region was 30%.

  Next, a bonded magnet was produced in the same manner as in Example 3 using the alloy flakes obtained in Comparative Example 4. Table 1 shows the magnetic properties of the obtained bonded magnet.

  From Table 1, it can be seen that the bonded magnets of Example 3 and Comparative Example 4 have excellent magnetic characteristics of Example 3. In Comparative Example 4, since the volume ratio of the fine R-rich region is high and the amount of relatively fine particles of 50 μm or less after the HDDR treatment or pulverization is large, it can be estimated that the magnetism is low.

According to the method for producing a rare earth-containing alloy flake of the present invention, the volume fraction of the fine R-rich phase region is small, and the homogeneity of the dispersion state of the R-rich phase in the alloy is even better than the conventional SC material. Alloy flakes can be manufactured with good reproducibility. Thereby, the sintered magnet manufactured from the obtained alloy flake and the bonded magnet by HDDR method express the magnet characteristic superior to the conventional one. In this way, the development of high-performance magnets contributes greatly to industrial development.
In addition, according to the metal structure evaluation method of the present invention, it is extremely easy to objectively determine a more preferable rare earth magnet, which can greatly contribute to the magnet industry.

It is a figure which shows the cross-sectional structure | tissue of the alloy flakes for rare earth magnets containing the fine R rich phase area | region manufactured by the conventional SC method. It is a figure which shows the cross-sectional structure | tissue of the alloy flakes for rare earth magnets concerning this invention. FIG. 2 is a diagram in which a line is drawn at the boundary between a fine R-rich region and a normal part in the cross-sectional structure of FIG. 1. It is a schematic diagram of the casting apparatus of a strip casting method. It is a figure explaining the measuring method of R rich phase space | interval. It is a figure explaining the measuring method of a fine R rich phase field. It is a figure explaining the reflected electron beam image of the alloy for rare earth magnets.

Explanation of symbols

1 .... Refractory crucible, 2 .... Tundish, 3 .... Rotating roll for casting, 4 .... Alloy, 5 .... Capture Collection container, 10 ... Solidified surface, 11 ... Main phase, 12 ... R-rich phase

Claims (12)

R ₂ T _{14 that} captures a microscopic structure of a cross section of an R-T-B magnet alloy as a digital image, draws a straight line having a length (L) substantially parallel to the solidified surface of the alloy on the microscopic structure image, and intersects the straight line. Both phases are discriminated from the difference in luminance between the B main phase and the R rich phase, and the number of times the luminance on the straight line changes from the luminance determined as the R ₂ T ₁₄ B main phase to the luminance determined as the R rich phase ( N), a value (L / N) obtained by dividing the length (L) of the straight line is evaluated as an R-rich phase interval (r).   2. The R-T-B system magnet according to claim 1, wherein an average value and a standard deviation of the R-rich phase interval are calculated from an R-rich phase interval obtained from each straight line using a plurality of the straight lines. Method for evaluating the microstructure of an alloy. The microscopic structure of the cross section of the R-T-B magnet alloy is handled as a digital image, the microscopic structure image is divided by cells of an appropriate area, and the R ₂ T ₁₄ B main phase and the R rich phase are determined from the difference in luminance within each cell. It is characterized in that both phases are discriminated and the shape, size, and number of individual R-rich phases existing in each cell are measured to determine whether or not each cell is in a fine R-rich phase region. The metal structure evaluation method of a RTB system magnet alloy.   The metal of the R-T-B system alloy according to claim 3, wherein any one or more of the roundness of the structure, the aspect ratio, and the rectangular aspect ratio is used as the shape of the R-rich phase. Organization evaluation method.   The cell portion in which the roundness of the R-rich phase is 1.0 to 1.4 and the area of the R-rich phase having an area of 5 square micrometers or less is included in 20 or more in 100 square micrometers is fine. The metal structure evaluation method for an R-T-B alloy according to claim 4, wherein the metal structure is determined to be an R-rich phase region.   A cell portion in which 20 or more R-rich phases having an aspect ratio of the R-rich phase of 1.0 to 5.0 and an area of 5 square micrometers or less are contained in 100 square micrometers is fine R The metal structure evaluation method for an R-T-B alloy according to claim 4, wherein the metal structure is determined to be a rich phase region.   A portion in which 20 or more R-rich phases having a rectangular aspect ratio of the R-rich phase of 1.0 to 5.0 and an area of 5 square micrometers or less are included in 100 square micrometers is fine R The metal structure evaluation method for an R-T-B alloy according to claim 4, wherein the metal phase is determined to be a rich phase region.   The cells dividing the microscopic structure image are divided into lattice cells by a group of equally spaced lines substantially parallel to the solidified surface of the alloy and a group of equally spaced lines substantially perpendicular to the solidified surface. When the cell determined to be a fine R-rich phase region by the metallographic evaluation method according to any one of Items 7 is discontinuously present in a direction perpendicular to the solidified surface of the alloy, the solidified surface of the alloy 4. The method for evaluating a metal structure of an R-T-B alloy according to claim 3, wherein all the cells from the cell farthest to the solidified surface are determined as a fine R-rich phase region.   The cells dividing the microscopic structure image are divided into lattice-like cells by a group of equally spaced lines substantially parallel to the solidified surface of the alloy and a group of equally spaced lines substantially perpendicular to the solidified surface. When the cell determined to be a fine R-rich phase region by the metallographic evaluation method according to any one of Claims 7 exists discontinuously in a direction perpendicular to the solidified surface of the alloy, the fine R-rich Among the cells determined to be the phase region, all cells from the two or more consecutive cells farthest from the solidified surface to the solidified surface are determined to be fine R-rich phase regions and are more free than the two or more consecutive cells. 4. The method for evaluating a metal structure of an RTB-based alloy according to claim 3, wherein all the cells present independently in the plane direction are determined not to be in the fine R-rich phase region.   A value obtained by dividing the total area of the cells determined as the fine R-rich phase region by the metal structure evaluation method according to any one of claims 3 to 9 by the total area of all the measured cells. A method for evaluating the structure of an R-T-B alloy, wherein the area ratio of the R-rich phase region is used.   The metal structure evaluation method for an RTB-based magnet alloy according to claim 1 or 3, wherein the digital image is a reflected electron beam image of a scanning electron microscope. 4. The metal structure of the R-T-B magnet alloy according to claim 1, wherein a threshold value is provided for the difference in luminance to discriminate between an R ₂ T ₁₄ B main phase and an R-rich phase. 5. Evaluation methods.

Images