Magnetic properties of amorphous NdFeB thin films

Magnetic properties of amorphous NdFeB thin films

Journal of Magnetism North-Holland and Magnetic MAGNETIC Materials PROPERTIES J.M. ALAMEDA, 83 (1990) 75-76 OF AMORPHOUS M.C. CONTRERAS, 75 ...

175KB Sizes 2 Downloads 133 Views

Journal of Magnetism North-Holland

and Magnetic





83 (1990) 75-76







a and F.H. SALAS

Deparramento de Fisica, Universrdad de Ooiedo, 33007 Ouiedo. Spain a Lab. Louis Nkl, CNRS, 166X. 38042 Grenoble Cedex, France Transverse biased initial susceptibility (TBIS) measurements have been performed in as-prepared and in UHV annealed NdFeB and FeB amorphous films. The dependence of magnetic anisotropies (local and macroscopic) on substrate temperature during deposition (T, = RT and 77 K) and subsequent annealing treatments is discussed. Structural and magnetic properties of rapidly quenched NdFeB ribbons have been extensively investigated. Depending on the quenching rate, both amorphous (H, = 10 Oe) or polycrystalline (H, = several kOe) samples were obtained. The effective quenching rate for sputtered films is higher than that which results from melt spun techniques. For this reason, as prepared and sputtered NdFeB films are usually amorphous and present low values of coercive fields. In a previous work [l] we reported some of the magnetic properties of sputtered NdFeB thin films. The aim of this study is to present a quantitative analysis of transverse biased initial susceptibility (TBIS) measurements in order to discuss the dependence of local anisotropies on substrate temperature during deposition (r,) and subsequent thermal annealings. Moreover, we discuss the individual contributions of Nd and Fe to the local anisotropy observed in these high-quality amorphous (soft) magnetic materials. Amorphous Nd,,Fe,,B, thin films (thickness: 1000 A) were grown onto glass substrates by triode sputtering ( PAr = 4 X 10e4 Torr). Two experimental set-ups (A and B) were used. In A, a multicathode system (Nd, Fe and Fe,,B,,) was used and T, was RT. In B, a single

Table 1 Magnetic


of NdFeB

and FeB amorphous

cathode (Nd,,Fe,,B,) was used, T, = 77 K and RT, and the substrates were placed in a turnable sample holder. Subsequent thermal annealings were performed in UHV at temperatures above T,, but not high enough to induce crystallizations (see table 1). The amorphousness of the films was checked by X-ray diffraction. TBIS was measured by using a magneto-optic Kerreffect set-up [l]. Although TBIS was studied in both film/air (f/a) and glass/film (g/f) interfaces, here we discuss only the g/f results, where surface-oxidation effects may be ruled out. In fig. 1, TBIS (i.e.: x,‘(p) vs. H for p = 0 and 71/2 (see ref. [2]), for three representative cases are shown. From this figure and table 1, several facts may be noted: 1) The in-plane magnetic (induced) anisotropy H, as obtained from linear extrapolations of the x;‘(p) vs. H curves, increases when r, decreases. In all cases, a low value of H, is obtained after annealing. 2) The same is valid for the perpendicular magnetic anisotropy present in the films, as deduced from x;‘(O) vs. H curves, where the contribution arising from stripe domains is evident [2]. In fig. 1, H, corresponds to the in-plane applied field value above which the magnetiza-


B Composition


r, As-prepared /annealed

RT as-prepared

H, We)

18 8.7

K” (x10-3) (erg/cm3 ) H, We) h K,,V"* (x 104) ( erg/cm3’* )

28 0.47 3.3


77 K

142 70

RT annealed

77 K



2.8 1.4

4.6 2.2

2.3 2.4

3.2 3.7



9 4.4

6.5 3.9

0.26 1.6

0.17 1.7

RT as-prepared

10 4.9

94 0.29 9.7

0304-8853/90/$03.500 Elsevier Science Publishers (North-Holland)


0.29 1.8


J.M. Alameda et al. / Propertm

F//m Number







508 Annealed

77K 4.6 Oe I




NdFeB thin films

model for amorphous magnetism, a value of K,, = 10’ erg/cm3 is deduced for 3d anisotropy in FeB films. This order of magnitude is similar to the one obtained in other 3d-based amorphous alloys [2]. 2) However, as deduced from the high value of K,,V”’ observed in the sample obtained at low T, (77 K), a higher Nd contribution to K,, may occur in this case. Taking into account the individual contributions of Nd and Fe, the experimental value of K,,V’/* may be expressed as:



of amorphous


Fig. 1. Reciprocal susceptibility x;‘( /3) vs. applied bias field H for Nd,,Fe,,B, amorphous films (series B, thickness 1000 A) (see table 1 and text).

tion lies in the film plane. For a fixed thickness, H, is directly related to the magnitude of the perpendicular anisotropy K, [3]. 3) In all cases, linear extrapolations of x; ‘( fi) vs. H curves to the value x;‘(p) = 0 in the high field range, cut the abcissa at asymmetrical points. This fact is interpreted as magnetization ripple arising from a random distribution of local anisotropies. In this case, following Hoffmann and Harte models, the ripple parameter b can be extracted from TBIS [2]. From b, K,,V”2 is obtained [4]. Here K,,, is the magnitude of local anisotropy and V is the volume where the principal axes of these anisotropies are correlated. Both b and K&“/* are given in table 1 together with the results deduced in Fe,,B,, amorphous films. From this table, we conclude that K,,V’/* is higher for films grown at lower T,. Furthermore, K,,V’/* decreases after annealing. Additional comments to these results are: 1) For a fixed T,, K,,V ‘j2 depends on the sputtering system used (A or B). Actually, the results for NdFeB and FeB films obtained in system A, suggest that the contribution to K,, of the Nd magnetic subnetwork is not larger than that of Fe. In the framework of the HPZ

K lot.I”‘* = K,,(


+ K,+(



where V, p and K are the atomic volume, volume fraction and local anisotropies of Nd and Fe, respectively. Considering VFe = (2.5 A)‘, V,, = (3.6 A)3 and as deduced from FeB, we obtain: K,, = 10’ K,, erg/cm3. 3) The effect of annealing on V, if any, must be to increase its value. Therefore from the results of K,,V’/* given in table 1, we conclude that K,, decreases in annealed samples. 4) A clear correlation between the behaviours of macroscopic (i.e., H,, H,) and microscopic (i.e., KlwV”*) anisotropies is evident from table 1. Furthermore, samples grown at the same r, but in different sputtering systems (A or B), show clear differences. It is also important to note, that in samples prepared in system B (rotating substrate), self-shadowing effects or slight anisotropic diffusion of incoming atoms may be more pronounced. In conclusion, self-shadowing may produce lower symmetry around individual atoms and slight anisotropic arrangements of atoms at a microscopic scale. This may be the origin of the magnetic anisotropies observed.

This work is supported by the EEC (EURAM tract no. MAlE-0081-C(EDB)).


References [l] J.M. Alameda, F. Briones, M.C. Contreras, J.F. Fuertes, D. Givord and A. Litnard, in: Magnetic Properties of Amorphous Metals, eds. A. Hernando et al. (North-Holland. Amsterdam. 1987) p. 262. [2] J.M. Alameda, M.C. Contreras and H. Rubio, Phys. Stat. Sol. (a) 85 (1984) 511. [3] A. Holz and H. Kronmiiller, Phys. Stat. Sol. 31 (1969) 787. [4] J.M. Alameda, M.C. Contreras, H. Rubio, F. Briones, D. Givord and A. Litnard, J. Magn. Magn. Mat. 67 (1987) 115.