SINTERED NDFEB MAGNET AND METHOD FOR MANUFACTURING THE SAME

Abstract

Disclosed is a sintered NdFeB magnet having high coercivity (H.sub.cJ) a high maximum energy product ((BH).sub.max) and a high squareness ratio (SQ) even when the sintered magnet has a thickness of 5 mm or more. The sintered NdFeB magnet is produced by diffusing Dy and/or Tb in grain boundaries in a base material of the sintered NdFeB magnet by a grain boundary diffusion process. The sintered NdFeB magnet is characterized in that the amount of rare earth in a metallic state in the base material is between 12.7 and 16.0% in atomic ratio, a rare earth-rich phase continues from the surface of the base material to a depth of 2.5 mm from the surface at the grain boundaries of the base material, and the grain boundaries in which R.sub.H has been diffused by the grain boundary diffusion process reach a depth of 2.5 mm from the surface.

Claims

1. A method for manufacturing a sintered NdFeB magnet, comprising: making a starting alloy ingot by a strip-cast method in which an amount of rare-earth in a metallic state is between 12.7% and 16.0% in atomic ratio and lamellas of rare-earth rich phases are formed at an average interval controlled to be substantially the same as a target average particle size; making a powder containing particles in which fragments of the rare-earth rich phases are attached to main phase particles by grinding the starting alloy ingot so that an average particle size becomes the target average particle size; sintering the powder to make a base material of the NdFeB magnet; and performing a grain boundary diffusion process of R.sub.H, where R.sub.H is Dy and/or Tb, to the base material.

2. The method for manufacturing a sintered NdFeB magnet according to claim 1, wherein any one of the following powders a) through e) is used in the grain boundary diffusion process: a) a powder of an alloy containing R.sub.H and an iron group transition metal with an R.sub.H content of equal to or higher than 50 atomic percent; b) a powder of a metal composed of only R.sub.H; c) a powder of a hydride of the alloy of the powder a); d) a powder of a hydride of the metal of the powder b); and e) a mixed powder of R.sub.H fluoride powder and Al powder.

3. The method for manufacturing a sintered NdFeB magnet according to claim 2, wherein the powder containing R.sub.H is applied only to magnetic pole faces of the base material in the grain boundary diffusion process.

4. The method for manufacturing a sintered NdFeB magnet according to claim 1, wherein a thickness of the NdFeB magnet is equal to or more than 5 mm.

5. The method for manufacturing a sintered NdFeB magnet according to claim 1, wherein the amount of the powder layer on the surface the magnet base material is equal to or more than 7 mg per 1 cm.sup.2.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0025] FIG. 1A is a schematic diagram showing a starting alloy ingot having lamellas of rare-earth rich phase.

[0026] FIG. 1B is a schematic diagram showing a fine powder obtained by crushing the starting alloy ingot.

[0027] FIG. 2 is a wavelength dispersive spectrometry (WDS) map at a depth of 3 mm from the pole face, measured for the present embodiment and a comparative example.

[0028] FIG. 3 shows the result of a linear analysis in which a concentration distribution of Dy was measured in one direction on a cutting surface of a sample that had undergone a grain boundary diffusion process.

BEST MODE FOR CARRYING OUT THE INVENTION

[0029] Hereinafter, an embodiment of an sintered NdFeB magnet according to the present invention and a method for manufacturing it will be described.

Embodiment

[0030] A method for manufacturing a sintered NdFeB magnet of the present invention and that of a comparative example will be described.

[0031] Initially, an alloy of a NdFeB magnet was made by using a strip cast method. Subsequently, the alloy was roughly crushed by a hydrogen pulverization method, a lubricant was added to the obtained coarse grains, and then the coarse grains were ground into fine powder in a nitrogen gas stream by a 100AFG jet-milling apparatus, produced by Hosokawa Micron Corporation, to obtain a powder of NdFeB magnet. During the process, the grain size of the fine powder created by the grinding process was controlled so that the median (D.sub.50) of the grain size distribution measured by a laser diffraction method would be 5 m. Next, a lubricant was added to this powder, and the powder was filled into a filling container to a density of 3.5 through 3.6 g/cm.sup.3. After being oriented in a magnetic field, the powder was heated at 1000 through 1020 C. in a vacuum to be sintered. Then, after being heated at 800 C. in an inactive gas atmosphere for one hour, the sintered compact was rapidly cooled. Further, the sintered compact was heated at 500 through 550 C. for two hours and was rapidly cooled. As a result, a compact (which will hereinafter be called a base material) of a sintered NdFeB magnet before the diffusion of R.sub.H was obtained.

[0032] The aforementioned operation was performed for 12 kinds of alloys having different compositions. The compositions of the obtained 12 kinds of base materials (S-1 through S-9, and C-1 through C-3) are shown in Table 1, and their magnetic properties are shown in Table 2. In Table 2, B.sub.r is a residual flux density, and MN is an abbreviation of Magic Number, which is a value defined as the sum of a value of H.sub.cJ expressed in kOe and that of (BH).sub.max expressed in MGOe. Conventionally, in the sintered NdFeB magnets manufactured under the same conditions, the values of MN are almost constant because, as previously explained, the relationship between H.sub.cJ and (BH).sub.max can be approximated by a linear function having a negative slope. The value of MN of the sintered NdFeB magnets manufactured by a conventional common method is around 59 through 64, and does not exceed 65. Also for the base materials shown in Table 2, MN is within that range.

TABLE-US-00001 TABLE 1 BASE NONMETAL MATERIAL ATOM (ppm) METAL ATOM (ATOMIC PERCENT) NUMBER O C N Nd Dy Pr Co Cu B Al Fe MR S-1 1100 685 290 26.60 0.03 4.70 0.92 0.09 1.01 0.27 Bal. 13.40 S-2 1420 830 370 26.40 0.00 4.60 0.90 0.09 1.00 0.27 Bal. 13.01 S-3 1920 950 380 26.50 0.00 4.50 0.91 0.09 1.00 0.26 Bal. 12.79 S-4 1130 810 380 26.70 0.01 4.70 0.92 0.09 1.04 0.26 Bal. 13.30 S-5 900 770 310 26.60 0.00 4.70 0.92 0.09 1.03 0.26 Bal. 13.38 S-6 1000 900 480 22.20 4.00 6.30 0.89 0.12 1.00 0.20 Bal. 13.71 S-7 1820 1000 680 22.10 4.00 6.10 0.89 0.12 0.99 0.20 Bal. 13.12 S-8 1790 950 740 22.00 4.20 6.20 0.90 0.09 1.01 0.20 Bal. 13.21 S-9 1930 1220 760 22.00 4.10 6.00 0.91 0.10 1.01 0.20 Bal. 12.86 C-1 1850 1240 880 26.55 0.01 4.70 0.90 0.09 1.00 0.27 Bal. 12.51 C-2 1980 1100 850 30.30 0.11 0.28 0.94 0.08 0.98 0.22 Bal. 12.23 C-3 1910 1340 1000 21.60 4.00 6.10 0.90 0.10 1.00 0.20 Bal. 12.45

TABLE-US-00002 TABLE 2 BASE MATERIAL B.sub.r H.sub.cJ (BH).sub.max H.sub.k SQ NUMBER (kG) (kOe) (MGOe) (kOe) (%) MN S-1 13.8 15.7 46.7 14.4 91.8 62.4 S-2 13.8 15.6 46.4 14.6 93.9 62.0 S-3 13.8 15.5 46.5 14.5 93.3 62.0 S-4 14.2 13.0 49.0 11.6 89.2 62.0 S-5 14.2 13.5 49.3 12.1 89.6 62.8 S-6 12.8 23.3 40.7 21.3 91.5 64.0 S-7 13.0 22.7 41.2 20.7 91.2 63.9 S-8 12.7 22.6 40.1 20.6 91.2 62.7 S-9 12.8 22.4 40.7 20.4 91.1 63.1 C-1 14.1 12.4 48.2 11.1 89.5 60.6 C-2 14.2 10.2 49.0 8.9 87.3 59.2 C-3 13.0 21.7 41.2 19.7 90.8 62.9

[0033] The values of the compositions shown in Table 1 were obtained by a chemical analysis of the base materials. The value of MR is the amount of rare earth in a metallic state expressed in atomic percent, and was calculated from the values obtained by the aforementioned chemical analysis. In other words, the value of MR was obtained by subtracting the amount of rare earth consumed (non-metalized) by oxygen, carbon, and nitrogen from the entire amount of rare earth of the analysis value. In this calculation, it was presumed that these impurity elements were respectively combined with rare earth R to form R.sub.2O.sub.3, RC, and RN.

[0034] The base materials C-1 through C-3 each have an MR value of less than 12.7%, which is out of the scope of the present invention (i.e. within that of a comparative example). On the other hand, the base materials S-1 through S-9 each have an MR value of equal to or more than 12.7%, which is within the scope of the present invention. Of these, the base materials S-1 through S-5 do not contain Dy in excess of the impurity level, whereas the base materials S-6 through S-9 contain around 4 atomic percent of Dy. The base materials S-1 through S-9 are grouped based on the following two terms. The first group is composed of the base materials S-1 through S-3, and S-6 and S-7. For these base materials, when an alloy was put into a jet mill, the initial input amount was approximately 400 g, the supply rate was approximately 30 g per minute, and the pressure of nitrogen gas was 0.6 MPa. The second group is composed of the base materials S-4, S-5, S-8, and S-9. For these base materials, the initial input amount was more than that of the first group. The initial input amount was approximately 700 g, the supply rate was approximately 40 g per minute, and the pressure of nitrogen gas was 0.6 MPa.

[0035] Next, for the twelve kinds of base materials S-1 through S-9, and C-1 through C-3, rectangular parallelepiped base materials of 7 mm in length by 7 mm in width by 5 mm or 6 mm in thickness were cut out in such a manner that the thickness direction coincided with the direction of the magnetic orientation.

[0036] Along with the manufacture of the rectangular parallelepiped base materials as previously described, a powder to be applied to the rectangular parallelepiped base materials was prepared in order to perform the grain boundary diffusion method. Table 3 shows the compositions of the powders used in the present embodiment. The average grain size of the powders A and B was 6 m. The average grain size of the DyF.sub.3 powder used for the powders C and D was approximately 3 m, and the average grain size of the Al powder used for the powder C was approximately 5 m.

TABLE-US-00003 TABLE 3 (Unit: Percent by Weight) POWDER SYMBOL Dy Ni Co DyF.sub.3 A1 A 92 4.3 0 0 3.7 B 91.6 0 4.6 0 3.8 C 0 0 0 90 10 D 0 0 0 100 0

[0037] Subsequently, the powders A through D were applied to the surface of the rectangular parallelepiped base materials in the following manner. Initially, 100 cm.sup.3 of zirconia spherules with a diameter of 1 mm was put into a plastic beaker with a capacity of 200 cm.sup.3, 0.1 through 0.5 g of liquid paraffin was added thereto, and the spherules were stirred. A rectangular parallelepiped base material was put into the plastic beaker, and the base material and spherules in the beaker were vibrated by placing the beaker in contact with a vibrator, so that an adhesive layer composed of paraffin was formed on the surface of the rectangular parallelepiped base material. Then, 8 cm.sup.3 of stainless spherules with a diameter of 1 mm were put into a glass bottle with a capacity of 10 cm.sup.3, 1 through 5 g of the powder shown in Table 2 were added, and the rectangular parallelepiped base material coated with the adhesive layer was put into the glass bottle. For the reason which will be described later, the sides of the rectangular parallelepiped base material (i.e. the surfaces other than the pole faces) were masked with a plastic plate to prevent the powder from being applied to these sides of the magnet. This glass bottle was brought into contact with the vibrator to make a sintered NdFeB magnet in which a powder containing Dy was applied only to the pole faces. The amount of applied powder was adjusted by controlling the amount of the liquid paraffin and that of the powder added in the previously described step.

[0038] The reason why the powder was applied only to the pole faces is as follows. Aiming at an application to a relatively large motor, the present invention had to prove to be an effective technology for a magnet having a relatively large pole area. However, the use of a magnetization curve measuring device (for performing a measurement by applying a pulsed magnetic field) inevitably limited the pole area. For this reason, a sample having a relatively small pole area of 7 mm square was used. To overcome this limitation, the powder was not applied to the sides of the sample so as to create a situation virtually equivalent to the case where an experiment of the grain boundary diffusion method was performed for a sample having a large pole area.

[0039] Then, the rectangular parallelepiped base material coated with a powder was put on a molybdenum plate, with one of the sides to which the powder was not applied facing downward, and then heated in a vacuum of 10.sup.4 Pa. The heating was performed at a temperature of 900 C. for three hours. After that, the base material was rapidly cooled down to the room temperature, heated at 500 through 550 C. for two hours, and rapidly cooled down again to the room temperature.

[0040] In the aforementioned manner, fifteen kinds of samples D-1 through D-15 were prepared. Table 4 shows: the base material of each sample; the combination of the powder and the application amount of the powder; the measurement values of coercive force H.sub.cJ, maximum energy product (BH).sub.max, MN, and squareness ratio SQ; and the measurement result of the presence of Dy at the central position in the thickness direction (2.5 mm from the surface for a sample having a thickness of 5 mm, and 3 mm from the surface for a sample having a thickness of 6 mm).

TABLE-US-00004 TABLE 4 WITHIN THE BASE THICKNESS Dy SCOPE OF SAMPLE MATERIAL OF BASE APPLIED H.sub.cJ (BH).sub.max SQ DETEC- THE PRESENT NUMBER NUMBER MATERIAL POWDER (kOe) (MGOe) MN (%) TION INVENTION? D-1 S-1 5 A 22.2 44.9 67.1 90.8 Y Y D-2 S-2 5 A 22.0 45.1 67.1 91.3 Y Y D-3 S-3 5 A 21.9 44.7 66.6 90.5 Y Y D-4 S-4 5 A 19.5 47.8 67.3 81.5 N N D-5 S-5 5 A 19.3 47.4 66.7 82.3 N N D-6 S-6 6 A 28.4 38.7 67.1 93.3 Y Y D-7 S-7 6 A 28.3 39.4 67.7 93.4 Y Y D-8 S-8 5 A 27.0 39.8 63.2 83.4 N N D-9 S-9 5 A 26.9 39.6 62.1 85.2 N N D-10 C-1 5 A 18.8 46.8 60.9 82.7 N N D-11 C-2 5 A 16.6 47.8 61.3 79.8 N N D-12 C-3 5 A 23.4 38.6 62.0 86.7 N N D-13 S-1 5 B 21.8 44.9 66.7 90.8 Y Y D-14 S-1 5 C 21.3 45.6 66.9 90.2 Y Y D-15 S-1 5 D 17.0 46.1 63.1 85.6 N N

[0041] The magnetic properties were measured with a pulse magnetization measuring system (trade name: Pulse BH Curve Tracer BHP-1000), with the largest application magnetic field of 10 T, produced by Nihon Denji Sokki Co., Ltd. Pulse magnetization measuring systems are suitable for evaluating high H.sub.cJ magnets which are a subject matter of the present invention. However, as compared to a general system for measuring magnetization by applying a direct-current magnetic field (which is also called a direct-current B-H tracer), the pulse magnetization measuring equipment is known to tend to yield a lower squareness ratio SQ of the magnetization curve. A squareness ratio SQ equal to or higher than 90% in the present embodiment is comparable to a level equal to or higher than 95% measured by a direct-current magnetization measuring system.

[0042] The presence of Dy at the central position in the thickness direction was determined in the following manner. A section which passes through the central position and which is parallel to the pole faces of the sample was cut out by a peripheral cutter, the cut surface was polished, and then Dy was detected by the WDS analysis by an electron probe microanalyzer (EPMA; JXA-8500F produced by JOEL Ltd.). As an example, FIG. 2 (upper images) shows WDS map images at a depth of 3 mm from the pole face of a sample created from the base material S-1 by applying the powder A to only one of the pole faces and performing the aforementioned grain boundary diffusion process and the subsequent heat treatment. FIG. 2 also shows WDS map images (lower images) at a depth of 3 mm of another sample created from the base material S-1 without performing the grain boundary diffusion process. In these images, the white portions in the COMPO images indicate crystal grain boundaries of the rare-earth rich phase. Since the amount of Dy originally contained in the base material S-1 is no higher than impurity levels, no Dy was found at the grain boundaries in the sample for which the grain boundary diffusion process had not been performed. By contrast, Dy was detected (at the portions indicated with the arrows in the upper images) in the sample for which the grain boundary diffusion process had been performed. FIG. 3 shows the result of a linear analysis in which the concentration distribution of Dy in one direction on the cut surface was measured for the sample for which the grain boundary diffusion process had been performed. This linear analysis also confirmed that Dy was concentrated at the grain boundaries. The determination result of Dy detection shown in Table 4 was obtained by this WDS analysis.

[0043] The result shown in Table 4 demonstrates that only the sintered NdFeB magnets in which the value of MR in a metallic state contained in the base material of the sintered NdFeB magnet was equal to or higher than 12.7 atomic percent and the concentration of Dy in the crystal boundaries was detected at a depth of equal to or more than 2.5 mm from the surface of the sintered compact, have a high H.sub.cJ, high (BH).sub.max, and a high SQ value. The samples D-4, D-5, D-8, and D-9, which were prepared by using the base materials S-4, S-5, S-8, and S-9 (which were the base materials of the second group) having a relatively high MR value, had no concentration of Dy at the grain boundaries at the central portion of the sample for the reason which will be described later. Such samples all do not have a high H.sub.cJ high (BH).sub.max, or high SQ value. Only the sintered NdFeB magnet of a sample which satisfies the following two conditions has an MN value exceeding 66 and an SQ value equal to or higher than 90: the MR value is equal to or higher than 12.7 atomic percent and the concentration of Dy at the crystal grain boundaries is detected at a depth of equal to or more than 2.5 mm from the surface of the sintered compact. Every sample was made by using the base materials of the first group.

[0044] The difference between the samples prepared from the base materials of the first group and the samples prepared from the base materials of the second group will be described. For the first group and the second group, an alloy powder before being formed into a base material (sintered compact) was observed with an electron microscope and the ratio of the grains with the rare-earth rich phases attached thereon to the whole grains was obtained. As a result, the ratio was equal to or higher than 80% for the first group, whereas the ratio was not higher than 70% for the second group. Such a difference probably occurred due to the difference of the conditions of the previously described process of preparing fine powders. It is known that, in the 100AFG jet milling apparatus, the crushing energy tends to be larger as the amount of crushing object accumulated in the apparatus becomes larger and as the gas pressure becomes higher. In a strip cast alloy before crushing, plate-like lamellas of rare-earth rich phase are distributed at regular intervals. Hence, the higher the crushing energy becomes (i.e. more for the second group than for the first group), the more easily the rare-earth rich phases are separated. If a rare-earth rich phase is separated from the main phase, a point where a rare-earth rich phase does not exist appears in the grain boundaries after the sintering, causing a discontinuity of the rare-earth rich phases. At such a chasm, when the base material is heated in the grain boundary diffusion process, the grain boundaries will not be melted. In the grain boundary diffusion process, R.sub.H diffuses within the base material (sintered compact) through melted grain boundaries as a passage, and therefore does not reach the portion deeper than the chasm of the rare-earth rich phases. Consequently, in the position deeper than equal to or more than 2.5 mm from the surface of the sintered compact, Dy does not exist for the second group, whereas Dy exists for the first group.

[0045] A sintered NdFeB magnet used for a high-tech product such as a large motor for a hybrid or electric car is required to have a large H.sub.cJ and (BH).sub.max, and therefore large MN, in addition to a large SQ value. Further, a magnet to be used in such large motors normally has a relatively large thickness of equal to or more than 5 mm. Conventionally, no magnet with such a thickness has the aforementioned characteristics. The sintered NdFeB magnet according to the present invention is a long-awaited magnet which has all the aforementioned characteristics and can be used as a high-performance magnet of the highest quality.

[0046] In the present embodiment, the explanation is made for the case where Dy is used as R.sub.H. However, if Tb (which is more expensive than Dy) is used in place of Dy, the value of H.sub.cJ can be further increased.

EXPLANATION OF NUMERALS

[0047] 10 . . . Starting Alloy Ingot [0048] 11 . . . Main Phase [0049] 12 . . . Rare-Earth Rich Phase Lamella [0050] 13 . . . Fine Powder Grain [0051] 14 . . . Part of the Rare-Earth Rich Phase Lamella