NEW MAGNETIC ALLOYS

B. A. Vvedenskii.

I. Strongly Magnetic Alloys

Alloys of iron with nickel have long since [^1] attracted the attention of physicists and engineers thanks to their varied and remarkable properties. Indeed, depending on the composition one can, for example, obtain materials of the most diverse hardness, from very soft and ductile to very hard; one can obtain materials with a coefficient of expansion from 0.0000131 (pure nickel) to 0.0000016—“invar” with 36% nickel; “invar” thus expands 5.5 times less than platinum. One can obtain materials with a specific resistance from 7 micro-ohms (pure nickel) to 80 micro-ohms—with a 35% nickel content. Finally—and this is the proper subject of our discussion—nickel–iron alloys give us materials with extraordinarily interesting magnetic properties. It is also highly characteristic that the method of heat treatment has exceptional and decisive importance for the manifestation of these properties.

It had long been known that nickel steel with a 25-percent nickel content, depending on the heat treatment, may either possess or not possess ferromagnetic properties: when heated to the Curie point (for this steel the Curie point—the point of magnetic transformation—lies at 580° C.) and then cooled to room temperature, it remains nonmagnetic, and only cooling below zero restores magnetism to this steel.

Furthermore, in recent times Ensen [^2] drew attention to the increased—even in comparison with iron—magnetism of certain alloys; especially to the extraordinarily small magnitude of hysteresis losses in alloys with 40–80% nickel. The increased magnetism was also indicated by Wurschmidt [^3]. However, the alloy discovered by Arnold and Elmen [^4], containing 78% nickel and named by them, for its exceptional magnetism, permalloy (from permeability—permeability, and alloy—alloy), leaves all previously investigated alloys far behind in its properties.

NEW MAGNETIC ALLOYS

The most characteristic and remarkable property of this new alloy is its utterly exceptional ability to become magnetized in the weakest magnetic fields, of the order of hundredths and even thousandths of a gauss, where ordinary iron is still very weakly magnetic; in other words, this alloy possesses an extraordinarily large, in comparison with iron, initial permeability. In Fig. 1, for comparison, curves of magnetic induction are given for permalloy, on the one hand, and for very pure “Armco” iron (American Rolling Mill Company), on the other, from which it is evident that, as the external field increases, permalloy manages to become magnetized to saturation before ordinary iron has even begun to become strongly magnetized. Permalloy is magnetized to saturation already in the earth’s magnetic field! Hence arises the need for quite unusual precautions in its investigation.

Fig. 1.

The exceptional nature of the magnetic properties is presented from a somewhat different point of view in the curves of Fig. 2, which depict the course of the magnetic permeability for the new alloy and for iron: in permalloy the permeability not only reaches its greatest values at much smaller field strengths than iron, but the very magnitude of this permeability has an entirely unusual value, reaching, near 0.05 gauss, the value 85000.

Fig. 2.

However—and this is exceedingly surprising, like almost all the properties of this unusual alloy from beginning to end—the values of the induction at saturation are not only no higher than for iron, but even 40–50 percent lower. Hence the conclusion that the field of application—

and, moreover, one that is already extraordinarily successful at the present time and even more promising in the future—is the field of weak magnetic fields, the field of “weak currents” in telegraphy and telephony. Indeed, one of the problems of signaling over very long wires, namely the problem of eliminating the interfering action of the capacitance of long lines, seemed to be waiting for the discovery of permalloy in order at once to make colossal progress. The capacitance of lines had long been combated by artificially increasing their self-induction: by Pupin’s method special reactive coils are inserted into the line at definite intervals; by Krarup’s method [8] the wires are wrapped with iron wire, increasing the self-induction of the conductor. The second method is in many respects more convenient, but the chief obstacle to its application was the very low permeability of ordinary iron, and also of other previously known materials, in the weak fields that occur around a conductor carrying weak currents, for example telephone currents. It is clear that the discovery of permalloy came here at the most opportune moment; in particular, it quite literally revolutionized signaling by submarine cables.

Fig. 3.

To these unusually valuable properties of permalloy one more is added: its coercive force is extraordinarily small, the hysteresis loop is narrow, and consequently the hysteresis losses are likewise negligible [9] (Fig. 3). This makes it an unusually valuable material for all kinds of transformers and reactive coils operating in weak magnetic fields.

As we have already said, alloys with a different percentage content of nickel are more magnetic than iron. In the curve of Fig. 4 the values of the initial permeabilities are compared, obtained from the corresponding curves by extrapolation to the field value \(H = 0\), for different percentage contents of nickel. From the curve it is seen that near 80% Ni the permeability forms a very sharp maximum, so that for their discovery Arnold and Elmen had to investigate a large number of different alloys before they found permalloy. The exact composition of their alloy is as follows: nickel—78.23%; iron—21.35%; carbon—0.04%; the remainder—other impurities, such as sulfur, silicon, manganese, chromium. Impurities in general are extremely harmful, especially carbon.

NEW MAGNETIC ALLOYS

Thus, for example, in one case two alloys, one strongly magnetic and the other with ordinary properties, differed from one another only by a difference in carbon content of 0.01%.

The fact that nickel alloys possess outstanding properties becomes still more astonishing if it is pointed out that an alloy just taken from the furnace and cooled does not yet possess any special qualities. Only after a very peculiar heat treatment do these properties appear: permalloy—in the form of thin ribbons—is heated to \(900^\circ\) for one hour and then slowly cooled; next it is again heated to \(600^\circ\) and rapidly cooled. In this, not only the temperatures but also the time of their action is very important; moreover, changes in the composition of the alloy also require changes in the method of treatment (cf. below, Ensen’s method).

To these unusual properties there is added still a whole series of others. It is known that magnetic properties change under mechanical action, for example, under drawing. In exactly the same way, the permeability of permalloy changes—and, in fact, decreases—for the same reason. But the remarkable thing is that this change may attain an enormous magnitude: under the action of drawing, the permeability decreases to 10% of its original value. At the same time, so long as in drawing we do not pass the elastic limit, the changes in permeability are reversible, just like mechanical deformations.

Fig. 4.

Subsequent investigations by Buckley \([^{10}]\) introduced still much that was new. It turned out that alloys with contents of 65% and 84% nickel give mutually opposite effects of tension on the magnetic properties. An alloy with 81% is almost completely indifferent with respect to tension. However, incorrect heat treatment may completely mask these characteristic differences. Further, it turns out that the hysteresis of the alloy decreases under tension and can be brought down to one hundredth of the hysteresis (measured by the area of the magnetization curve) of the best silicon steel.

It is quite natural to pose the converse question: concerning the influence of a magnetic field on the dimensions of a body, for example, of a wire made of permalloy (magnetostriction). In this direction the question, so far as is known, has not yet been fully clarified: Arnold and Elmen themselves found that the striction phenomena in permalloy are greater than ordinary ones; on the contrary, Honda and Kido \([^{11}]\), who investigated a similar alloy as early as 1919,

(they did not suspect its exceptional properties, which appear only after heat treatment), found an absence of strictional phenomena. But in both cases it is extremely interesting to compare the anomaly of magnetostriction with the extremely small magnitude of hysteresis (Fig. 3) in permalloy. In the event that it should turn out that a small value of hysteresis is also accompanied by a small value of magnetostriction, this would lead to an entirely new view of the nature of hysteresis in general, since until now the phenomena of hysteresis and magnetostriction have not been connected with one another. Mac Keegan \[12\] expresses the view that the stresses inside the atoms, arising as the result of the action of an external field (“atomic magnetostriction”), are the chief cause of the “magnetic hardness” of steel, i.e. of hysteresis. Strictional phenomena in iron and nickel in weak fields have opposite signs: the length of a nickel wire continually decreases, whereas in the case of iron, in weak fields, an elongation is observed. Therefore, with a properly chosen percentage content of nickel, the interatomic stresses must mutually cancel one another, which should lead—and in fact does lead—to the disappearance of hysteresis.

The next interesting fact is the comparatively very considerable change in the electrical conductivity of permalloy both under the action of a magnetic field and under the action of mechanical stretching. A field of 1 gauss decreases the conductivity by 2%. In small fields and under stretching forces acting simultaneously, the two effects add together; and if the change, for example under the action of the magnetic field, has reached the greatest possible value, then stretching adds nothing further to the effect. This, apparently, unquestionably indicates that the mechanism of both phenomena—the changes due to the field and due to stretching—is the same.

The change in conductivity under the influence of the field is so great that by means of this phenomenon it proves entirely possible to make an accurate measurement of magnetic fields as weak as the earth’s field.

As we see, the properties of permalloy are sufficiently enigmatic and present a very broad field for investigation. The matter, one may say, is finally complicated by the circumstance that no other properties of permalloy, apart from those indicated, are in any way unusual: the crystal lattice is the same as in nickel (i.e. cubic with face-centered faces) and remains so as long as the percentage content of nickel does not fall below 35%. The mechanical properties are also normal, and—what is most surprising—heat treatment, which has such decisive significance for the magnetic properties, has not the slightest influence on the mechanical properties.

Very recently permalloy has acquired a competitor in the form of a 50% alloy proposed and investigated by Ence—

NEW MAGNETIC ALLOYS

... by him [13]; however, the initial permeability of this alloy is 3–4 times less than that of permalloy, but the maximum permeability—at 0.1 gauss—reaches approximately the same colossal value. To this latter circumstance is added the fact that the saturation induction, owing to the higher iron content, is approximately 40% higher than in permalloy. If one also takes into account its comparative cheapness—the lower content of expensive nickel—then it must be acknowledged that, in the region of comparatively strong fields, Ensen’s alloy may possess an undoubted superiority over permalloy; but the region of extremely weak fields remains, undoubtedly, with permalloy.

Of interest is the heat-treatment prescription given by Ensen for his alloy: annealing in vacuum for one hour at 900°, slow cooling to 625° (the Curie point); holding at this temperature for 15 minutes. During this time the furnace is filled with nitrogen. Then the specimen is quickly removed from the furnace and cooled in air. What is common with the method of Arnold and Elmen here is the slow cooling from a temperature above the Curie point. Ensen believes that in this way structural features present at the Curie point are preserved in the alloy even at room temperature; it is apparently these latter features that impart to various magnetic materials, especially iron, a strong increase in magnetism in the immediate vicinity of the exact Curie point.

II. A New Steel for Permanent Magnets

In the brilliance of permalloy there remained little noticed another alloy that appeared recently—this time of iron with cobalt and a number of other elements—distinguished by an enormous coercive force (the direct opposite of permalloy) and therefore especially suitable for the manufacture of permanent magnets. This new material was proposed by Gumlich [14].

Gumlich arrived at the discovery of this steel from the following facts. It had long been known that steel containing manganese in considerable quantities (up to 12%) in the hardened state possesses a very large coercive force of 130 gauss, but, unfortunately, the residual induction in this case falls to a very small value (of the order of 1000). However, on the other hand, Weiss and Preuss [15] found that an alloy of 35% cobalt with iron gives a saturation magnetization 10% higher than that of pure iron. From this arose the idea of increasing the coercive force by adding to iron both manganese and cobalt at the same time, without detriment to the magnitude of the induction. After a whole series of trials of various alloys (these trials were carried out at Krupp’s in Essen), there was finally found the composition of an alloy (1.10% C, 3.5% Mn, 36% Co, and 4.8% Cr), which, at a suitably chosen temperature...

after hardening gave an unusually large coercive force of 227 gauss (in the best magnetic steels this value did not exceed 90 gauss), with a very good residual induction of 9300 gauss.

According to Gumlich, the most advantageous hardening temperature is 850° C., and the hardening itself is carried out in machine oil cooled to 0°.

For technology the new steel offers a twofold advantage: first, magnets made of the new steel are far more stable with respect to temperature fluctuations, shocks, the action of external fields, etc.—this makes them especially suitable for the manufacture of precision measuring instruments; second, a magnet with a large coercive force resists much better the demagnetizing action of its ends, owing to which a magnet that is strongly open-circuited (i.e., with a large demagnetizing factor), when made of the new steel, is much stronger than one made of ordinary steel. For example, a rod 6 cm long and 0.7 cm in cross-section, used in the Kohlrausch–Holborn magnetometer, made of cobalt steel, has a 130% higher residual magnetization than the same rod made of chromium steel. Thus the new steel makes it possible to economize on material and, what is still far more important, on weight and volume.

The discovery of the remarkable properties of permalloy and other nickel alloys, as well as cobalt alloys, is, besides being of purely technical importance, also of enormous scientific interest, despite—or, better said, precisely because of—the enigmatic character of these properties. This discovery will undoubtedly give a powerful impetus to the elucidation of such fundamental, but still very obscure, questions as the nature of ferromagnetism and of metallic conductivity, and their dependence on other—for example, mechanical—properties of bodies. In particular, the problem of alloys, which has already yielded so much for magnetism (magnetic steels, Heusler alloys, etc.), in connection with rational (thermal, mechanical, and perhaps also magnetic) treatment promises in the future—there is every reason to think so—a series of new rich possibilities.

LITERATURE

1. Hopkinson, Proc. Roy. Soc., Dec. 1889; see J. Ewing, “Magn. Ind....”
2. Bargess a. Aston. Metall. Chem. Eng. 8, p. 23, 1910.
3. T. D. Jensen. Trans. Amer. Inst. El. Eng. 39, I p. 791, 1920.
4. L. Würschmidt. Phys. Zs. 23, 499, 1922 and Zs. f. Phys. 12, p. 128, 1922.
5. T. D. Jensen. Trans. Am. Inst. El. Eng. 39, p. 369, 1920.
6. H. D. Arnold a. G. W. Elmen. Journ. Frankl. Inst. 195, p. 621, 1923; Electrician, June 22, p. 672, 1923.
7. K. Zschiesche. Zs. f. Phys. 12, p. 201, 1922.
8. A. E. Kennelly. Journ. Frankl. Inst. 197, p. 623, 1924.
9. See, for example, “Kramp Cable.” Electrician, 91, p. 692, 1923.
10. L. W. Mc. Keechan a. P. P. Cioffi. Phys. Rev. 23, p. 305, 1924.

1. O. E. Buckley. Phys. Rev. 23, p. 783, 1924.
2. K. Honda and K. Kido. Sci. Rep. Tohoku Univ. 9, p. 221, 1920.
3. L. W. McKeehan. Phys. Rev. 23, p. 783, 1924; on this author’s views on the mechanism of ferromagnetism, see Journ. Frankl. Inst., 197, pp. 601 and 757, 1924.
4. T. D. Jensen. Journ. Frankl. Inst., 199, p. 333, 1925.
5. E. Gumlich. Zs. f. Phys. 14, p. 241, 1923.
6. Preuss. Diss. Zurich, 1912; cf. P. Weiss, Rev. Gén. des Sciences, 15 Jan. 1914.