EP0707739B1 - Cobalt/rare earth type magnetic powders containing fluorine and corresponding densified permanent magnets and process for the preparation thereof - Google Patents

Abstract

This invention pertains to the addition of fluorine to magnetic powders and to sintered permanent magnets of the Sm - Co family in order to enhance their atmospheric corrosion and oxydation resistance. These powders and magnets contain from 600 to 3000 ppm of fluorine, which is introduced in gaseous form during grinding, and/or homogenizing of said powders following grinding.

Description

The present invention relates to a method of protecting magnetic powders and permanent magnets of the transition metal - rare earth metal type against oxidation and atmospheric corrosion by the introduction of gaseous fluorine during the treatment of powders. It applies more particularly to powders and magnets of the transition metal - rare earth family, where the metal is essentially cobalt, and the rare earth essentially of samarium and / or neodymium and / or praseodymium and / or cerium. .

It is known from patent FR 2003246, a process for treating powders with permanent magnetism containing rare earths and transition metals in order to increase their coercive force; this method consists in treating said powders with an etching liquid based on an aqueous solution of hydrochloric, sulfuric or hydrofluoric acid.

Furthermore, the state of the art on these powders and magnets has been recently described in an exhaustive manner by KJ STRNAT "Rare Earth-Cobalt Permanent Magnets" in "Ferromagnetic Materials", p.131 to p.209 Vol.4, Publisher EP WOHLFARTH and KH BUSCHOW, North-Holland Amsterdam 1988.

• la famille de formule moléculaire SmCo₅ (contenant 33,8% en poids de terre rare) ne tolérant que peu de substitution du cobalt par un autre métal de transition comme le fer, mais où le samarium peut être remplacé en partie ou en totalité par le praséodyme, en partie par d'autres terres rares dont les préférées sont le néodyme, le cérium, le mish-metal, le gadolinium selon la propriété technique ou économique recherchée
• la famille de formule moléculaire Sm₂Co₁₇ (contenant 23,1% en poids de terre rare) où, de plus, le cobalt peut être substitué plus largement par d'autres métaux comme le fer, le cuivre, le vanadium, le zirconium, le niobium, le titane, le manganèse, le chrome. La substitution du Co par le fer permet en particulier d'accroître la densité d'aimantation spécifique.

There are essentially two families depending on the amount of rare earth contained:

• the family of molecular formula SmCo ₅ (containing 33.8% by weight of rare earth) which tolerates little substitution of cobalt by another transition metal such as iron, but where the samarium can be replaced in part or in whole by praseodymium, in part by other rare earths, the preferred of which are neodymium, cerium, mish-metal, gadolinium depending on the technical or economic property sought
• the family of molecular formula Sm ₂ Co ₁₇ (containing 23.1% by weight of rare earth) where, in addition, cobalt can be more widely substituted by other metals such as iron, copper, vanadium, zirconium , niobium, titanium, manganese, chromium. The substitution of Co by iron makes it possible in particular to increase the specific magnetization density.

These two families are commonly used by the electrotechnical industry because their high Curie temperature allows their use at high operating temperatures. The most flexible practice for their preparation is Powder Metallurgy where alloys are finely ground, the grains are aligned in the presence of a magnetic field, they are compressed to the required shape, they are sintered at high temperature then applying a hardening heat treatment, factory according to required specifications.

The review cited above highlights a serious problem (p. 158) of the stability of powders and magnets due to the particular affinity of rare earth for oxygen in the air, especially in the presence of moisture. The natural evolution is towards the formation of the oxide, for example Sm ₂ O ₃ and / or of the hydroxide which are non-magnetic compounds causing a reduction in the density of magnetization and being able to interfere with densification; moreover, their training consumes samarium, therefore shifts the original stoichiometry to make the formula evolve towards a composition which may be unfavorable for the properties.

génère dans le matériau fritté une fine dispersion de Sm₂ Co₁₇ qui conduit immédiatement à un effondrement de la coercitivité (résistance à la désaimantation).For example, a theoretical scheme like: $${\text{100 Sm Co}}_{\text{5}}\text{+}\frac{\text{21}}{\text{4}}{\text{O}}_{\text{2}}{\text{→ 83 Sm Co}}_{\text{5}}{\text{+ 5 Sm}}_{\text{2}}{\text{Co}}_{\text{17}}\text{+}\frac{\text{7}}{\text{2}}{\text{Sm}}_{\text{2}}{\text{O}}_{\text{3}}$$

generates in the sintered material a fine dispersion of Sm ₂ Co ₁₇ which immediately leads to a collapse of the coercivity (resistance to demagnetization).

The greatest precautions are therefore taken during the development of these products, when they are in the form of very fine powders (a few microns in Fischer measurement), to produce them, store them, and handle them under cover. of air and humidity, or when, shaped in the press, they must be thermally activated to densify by sintering, in the use of discontinuous or continuous ovens placed under vacuum or under inert gas.

For Sm Co ₅ , the record was obtained, and only in the laboratory, by Narasimham (5th International Workshop on Rare Earth-Cobalt Permanent Magnets and their Applications, June 1981, Ed. University of Dayton, Ohio, USA, p.629 et seq.) which, proceeding as far as possible from the air, managed to have only 1000 ppm by weight of oxygen, which nevertheless consumes 0.6% by weight of samarium and generates 0, 7% by weight of Sm ₂ O ₃ . He could obtain, under good magnetic alignment conditions, a magnetization density Br = 1060 mT, a specific energy of (BH) _max = 220 kJ / m3.

Industrial techniques cannot reach such a level of precaution, and must be satisfied, for reasonable costs, to maintain themselves for Sm Co ₅ around 6000 ppm by weight of oxygen; in particular, it is very difficult to pump out the gaseous species contained in a tablet. The best magnets in this family therefore do not exceed Br = 950 mT and (BH) _max = 175 kJ / m ³ .

The same causes produce the same effects in the Sm ₂ Co _{17 family} where cobalt is substituted, to a lesser degree, because the lower samarium content leads to a lower affinity for oxygen. However, there are on finished magnets contents close to 4000 ppm by weight of oxygen; it is suspected that small variations around this average cause harmful variations on the rectangularity of the curve J (H), because the magnetic hardening of these alloys results from a polyphase structure where the atoms preferentially segregate and whose obtaining is delicate.

After having examined various means for controlling the intakes and departures of oxygen, and in particular after having estimated the quantities chemi-sorbed definitively and physi-sorbed temporarily in the various stages of the process, the applicant concluded that it was necessary to seek to trap active sites by an atom more electronegative than oxygen making it more labile for its evacuation. It turned out that an addition of fluorine could be effective, because an atom of fluorine or oxygen blocks respectively a third of atoms of Samarium or two thirds of atoms of samarium in the molecules Sm F ₃ and Sm ₂ O ₃ .

In addition, the Applicant has found that this simple theoretical ratio could be largely exceeded by a particular method consisting in introducing during a fine grinding step, when it is carried out in a gas jet mill, a mixture of nitrogen and fluorine gas in the inert gas stream (nitrogen, argon) for example usually used in these mills. The weight proportions of the mixture N ₂ + F ₂ and its method of introduction have been determined experimentally so as not to generate safety problems for equipment and the environment. The fluorine content of the grinding gas is generally less than 250 ppm (by volume).

It has been found that an optimum lies in a range from 600 ppm to 3000 ppm by weight of fluorine in the powder or retained in the sintered magnet, this optimum being slightly different for the Sm Co ₅ family and the Sm family. ₂ Co ₁₇ ; the preferred range is from 600 to 2000 ppm and can reach 600-1500 ppm for Sm ₂ Co ₁₇ . The fluorine contents on powders and on sintered products were determined according to the method consisting in dissolving the product to be analyzed in acid solution, distilling the fluorine assisted and assaying it by specific electrode.

Fluorine gas is preferably used. With other experimental conditions, it can be replaced by other donors such as hydrofluoric acid HF or nitrogen trifluoride NF ₃ ; but the former is much more aggressive and difficult to control, the latter is much more expensive.

• en dessous de 600 ppm, il n'y a pas assez de sites actifs neutralisés quand on manipule de grosses quantités de poudres fines, le procédé est instable;
• en dessus de 3000 ppm, il y a trop de fluor actif, qui se transforme en fluorure de samarium, déplaçant donc la composition métallurgique locale par appauvrissement en samarium.

It has been found that:

• below 600 ppm, there are not enough neutralized active sites when handling large quantities of fine powders, the process is unstable;
• above 3000 ppm, there is too much active fluorine, which transforms into samarium fluoride, thus displacing the local metallurgical composition by depletion in samarium.

Finally, the Applicant has found that the results could still be slightly improved if the blending homogenization step, which normally follows any grinding in a gas jet mill for the purpose of reducing grinding segregation, was carried out under the same procedure. atmosphere of nitrogen plus fluorine mixture.

In particular, storage in humid air shows a slower evolution of oxygen recovery.

To support its invention, the applicant relies on the following figures and examples.

Figure 1 shows the known ranges of oxygen uptake in the classic manufacture of Sm Co ₅ magnets.

The oxygen contents are established on the LECO commercial device equipped with the appropriate oven. The process steps are as follows: (1) raw materials; (2) coarse pre-grinding (50 µm); (3) fine grinding (5 µm); (4) intermediate storage; (5) after sintering.

FIG. 2 gives the results of the magnetic characteristics (remanence Br, induction coercivity H _cB , limiting stiffness field Hk) and the density p as a function of the rare earth content of the powder mixture and of the sintering temperature ( Δ 1115 ° C; x: 1125 ° C; o: 1135 ° C) according to the prior art.

Figure 3 shows schematically an example of technical embodiment of the invention.

FIG. 4 illustrates, by comparison with FIG. 2, the significant slip obtained according to the invention towards higher optima lying towards lower rare earth contents, in particular below 36% for the compositions of Sm Co ₅ type. .

FIG. 5 illustrates a similar gain obtained in the family of Sm ₂ Co ₁₇ , in particular on the rectangularity of the curve (gain in field of stiffness limit Hk), Hk corresponding to the demagnetizing field reducing the induction to 90% of Br. The curve (1) is representative of the prior art, the curve (2) is representative of the invention.

Example 1 (according to the prior art )

• la base à 34,2% de terre rare (analyse fournisseur).
• l'additif amenant l'excès de terre rare à 41,2% (analyse fournisseur).

Two alloys have been supplied from outside to constitute:

• the base at 34.2% rare earth (supplier analysis).
• the additive bringing the excess of rare earth to 41.2% (supplier analysis).

These alloys were mechanically ground to around 50 µm (passing sieve) and then mixed to provide the contents plotted on the abscissa of Figure 2. These mixtures were finely ground (4 to 5 µm Fischer), compressed under magnetic field, sintered under vacuum at the temperatures indicated in Figure 2, then returned to 875 ° C for 4 hours and vigorously quenched, as prescribed by conventional treatment of this material (see KJ STRNAT).

One of the compositions was followed in oxygen content at the various stages, to give the average values of FIG. 1. On sintered product, apart from measurement uncertainties, we place ourselves in the current level of 6000 to 6500 ppm by weight of oxygen.

The interesting physical quantities are shown in FIG. 2. Even though the alloys used are not of the maximum quality that the suppliers deliver from time to time, we find the usual behavior of an extremely critical situation. Density, remanence and coercivity of induction peak for an average content of 36.5% of Sm and are a little sensitive to the sintering temperature; on the other hand, at this content, the limit stiffness field H _k , guarantee of thermal stability, is already collapsed. And the four sizes have collapsed to 36%. We thus realize the criticality of the manufacturing, which will be aligned on an average rate of 37%, but which is at the mercy of an untimely taking of 800 ppm by weight of oxygen.

The classic explanation from the literature is that the oxygen present in the material is labile and captures the samarium of the phases present to stabilize in refractory Sm ₂ O ₃ . There is then no longer sufficient residual free samarium to assist the grain-to-grain densification of the main phase, said coalescence of the grains being able to even be hampered by intergranular oxides.

Example 2 (according to the invention) .

• le produit final avait capté beaucoup moins d'oxygène, 3500 ppm environ.
• ceci entraînait un glissement des fonctions de réponse vers les basses teneurs en terre rare (optimum à 35% en poids), avec un gain conséquent de 20 à 30 mT sur la rémanence.
• et surtout, une bien plus grande stabilité du procédé en ce qui concerne les autres grandeurs.

The same alloys were treated according to the claimed procedure. After coarse grinding, the mixtures were treated in continuous feed in a nitrogen gas jet mill where, according to the diagram in Figure 3, a small amount of d was introduced into the main nitrogen delivered at 100 m3 / h nitrogen loaded with 10% fluorine by means of a calibrated nozzle. Some preliminary experiments are necessary so that the final product contains from 600 to 2000 ppm by weight of fluorine. The continuation of the operations is identical to that of Example 1, and in the results presented in FIG. 4, there was a level of 1000 ppm of fluorine in the finished product. It was surprising to find that:

• the final product had captured much less oxygen, about 3500 ppm.
• this resulted in a shift in the response functions towards low rare earth contents (optimum at 35% by weight), with a consequent gain of 20 to 30 mT on the afterglow.
• and above all, a much greater stability of the process with regard to the other quantities.

An explanatory hypothesis can be advanced in the temporary blocking of active samarium sites exposed to grinding by the fixation of a fluorine atom. This then prevents the approach and physi-sorbtion of a molecule with two oxygen, which could cause the harmful oxide Sm ₂ O ₃ to evolve. On the contrary, the simultaneous presence of samarium, fluorine and oxygen would favor, under vacuum and high temperature around 900-1000 ° C, the evolution towards a stable oxyfluoride Sm OF, the formation of which retains mobility in the samarium in the sintering operation.

Example 3 (according to the invention)

Before being compressed, the finely ground products of Example 2 were homogenized by stirring under an N ₂ atmosphere containing 200 ppm of fluorine. A slight gain on the oxygen content was noted, since its content fell to 3000 ppm. This is not enough to have a significant effect on the final properties. However, the wet air resistance could be increased from a few hours to a few days.

Example 4 (according to the invention)

Of all the rare earths, the praseodymium in Pr Co ₅ has the strongest magnetization (1200 mT) against 1100 mT for Sm Co ₅ . But attempts to use this more abundant element have always failed, probably in connection with increased oxidizability. So we only use it as a partial substitution when we want to slightly increase the magnetization, at the cost of increased technical instability.

As for Example 2, the preparation was repeated from an alloy base Sm Co ₅ 33% substituted by praseodymium. The same advantages have been obtained, as Table I illustrates for the remanence (in isostatic compression). TABLE 1 Prior art According to the invention Basic alloy Sm Co5 Br = 0.92 - 0.95 T 0.96 - 0.98 T Basic alloy (Sm2 / 3 Pr 1/3 Co5) Br = 0.99 - 1.01 T 1.02 - 1.05 T

These modest gains in absolute terms are very important for applications. Much more important are the reliability and flexibility provided in the manufacture of products, with less critical heat treatments, including the quenching speeds, especially when one wants to use higher praseodymium contents.

Example 5 (according to the invention)

The process was also applied to alloys of the Sm ₂ Co _{17 family} . For this, we chose a composition rich in Fe and poor in Sm with the hope of increasing the afterglow.

For composition by weight%
Sm = 25 Fe = 22 Cu = 4.5 Zr = 2.4 Co = balance

The usual process makes it possible to obtain the excellent performances of FIG. 5, with Br = 1160 mT, H _cJ = 1070 kA / m, H _cB = 770 kA / m, (BH) _max = 240 kJ / m3. But this product is handicapped by the lack of rectangularity of the curve, resulting in an H _k of 620 kA / m only. Such a magnet contains approximately 4000 ppm of oxygen.

• la montée de la rémanence à 1180 - 1190 mT
• la meilleure rectangularité à H_k = 850 kA/m

With the process according to the invention, and at a level of 700 ppm of fluorine, it it was possible to reduce the oxygen content to 2000 ppm and to lower the samarium to 24.7% by weight. Two beneficial effects have been obtained:

• the rise in remanence to 1180 - 1190 mT
• the best rectangularity at H _k = 850 kA / m

Claims (6)

1. Magnetic powder or permanent magnet of the SmCo₅ or Sm₂Co₁₇ type Sm Co series, characterised in that it contains between 600 and 3,000 ppm by weight of fluorine.
2. Magnetic powder or permanent magnet according to claim 1, characterised in that it contains between 600 and 2,000 ppm by weight of fluorine.
3. Magnetic powder or permanent magnet according to claim 1 or 2, belonging to the Sm₂Co₁₇ series, characterised in that it contains between 600 and 1,500 ppm by weight of fluorine.
4. Magnetic powder and permanent magnet according to one of claims 1 to 2, containing less than 36% by weight of rare earths in SmCo₅ type alloys.
5. Magnetic powder and permanent magnet according to claim 3, characterised in that it contains less than 25% by weight of rare earths.
6. Process for obtaining magnetic powder or a permanent magnet according to one of claims 1 to 5, characterised in that a mixture of gases containing less than 250 ppm by volume of gaseous fluorine is introduced into the grinder and/or into the powder homogenisation stage.

Images