HAL Id: jpa-00229078
https://hal.archives-ouvertes.fr/jpa-00229078
Submitted on 1 Jan 1988
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of
sci-entific research documents, whether they are
pub-lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diffusion de documents
scientifiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
Tb/Co MULTILAYER FILMS AND THE
RELAXATION OF THEIR INTERFACE
Y. Wang, Zhenyu Li, Q. Li, K. Sun, D. Sellmyer, J. Shen
To cite this version:
JOURNAL DE PHYSIQUE
Colloque C8, Supplkment au no 12, Tome 49, dkembre 1988
Tb/Co MULTILAYER FILMS
AND
THE RELAXATION OF THEIR INTERFACEY. J. Wang (I), Z. H. Li (I), Q. S. Li (I), K. Sun (I), D. J. Sellmyer (2) and J. X. Shen (2) (I) Institute of Physics, Chinese Academy of Sciences, Beijing 100080, China
(2) Behlen Laboratory of Physics, University of Nebraska, Lincoln, NE 68588-0111, U.S.A.
Abstract. - In Tb/Co multilayers, the saturation magnetization M, and the anisotropy constant K,, as a function of the modulation period are fitted by a mean-field model and a dipolar interaction model respectively. The Kerr rotation angles Bk relax at room temperature, which could be attributed to the existence of shear stress due to lattice mismatch at the interface.
I n t r o d u c t i o n
In recent years, the properties of the composition- ally modulated films have been widely investigated. A known fact is that a perpendicular anisotropy appears in some such multilayers. This vertical magnetization is, in general, attributed t o the action of the interface. However, the reason is not clear yet.
In the present paper, the magnetic anisotropy and Ok relaxation in TbjCo multilayer will be discussed. E x p e r i m e n t a l
By means of substrate rotation, the composition-
ally modulated Tb/Co films were sputtered on water- cooled and ambient temperature (non-water-cooled) glass substrates. The thickness of the individual Co or T b layers was chosen t o be equal, that is, the mod- ulation period is 2d. Prior to sputtering, the vacuum chamber was evacuated t o 3
x
T o r . Ar gas pres-sure was held constant a t 4 x Torr. The diameter of the targets is 65 mm.
R e s u l t s
The X-ray diffraction patterns for the water-cooled substrate TbjCo confirm that these samples basically possess the artificial superlattice structure. However, any appreciable diffraction peaks on the small' angle range have not been found for the films with the am- bient T substrates.
Figure 1 shows Ku, @k versus d for Tb/Co with dif- ferent T substrates. It is clear that for the water-cooled substrate Ok is smaller and K, is larger than that of ambient T (non-water-cooled) substrate, and Ok is tend t o increase when relaxation for the water-cooled sub- strate.
5 1 0 1 5 2 0 2 5 3 0
d ( A )
Fig. 1. - K, as the function of d. The lower insert gives Bk us. d for both water-cooled and ambient T substrates and the Bk relaxation in the water-cooled substrate. The upper insert indicates the mean value P us. d.
I
0 5 10 15 2 0 25 30 d r i r
Fig. 2. - The experimental M, us. d and the mean-field model fitting. The insert schematicaly denotes th Co con- centration distribution in each bilayer along a direction which is used for the M, summations. It is noticed that the Co concentration in Tb rich side is not given here for the large d because their contribution to the total M s are small and ignored here.
C8
-
1800 JOURNAL DE PHYSIQUEFigure 2 gives M, versus d for the ambient T sub- strates.
M,
gradually increases with increase of d. Discussion1. The magnetization Ms in amorphous RE-TM films could be described by a mean-field model (1). The total magnetization M, is expressed as
here N is the total number of atoms per unit volume, x is concentration of a sublattice 1 and S1, S2 are the av- erage spin of the two sublattices. However, the atomic arrangements in multilayers are periodic. If we cal- culate the magnetization in Tb/Co by the mean-field model, we must integrate M, (x) for the different con- centration x in Tbl-,Cox along the z direction in the bilayer. If the magnetization of each bilayer is Msb, then the total magnetization M, = nM,b. Here n is the number of the total bilayers. In the fitting pro- cedure, we calculate M, for d = 0.3 nm (x = 0.76) to be fitted to its experimental M, at first, then, M, as a function of the concentration x is determined. There- after, the distribution of Co concentrations is adjusted dong the z direction (for the simplicity) in the bilayers of Tb/Co until the above-mentioned M. summations made for each d fit the experimental curve of Ms as the function of d. This fitting gives the distribution of Co concentration in each Silaver as shown in the insert of figure 2. According t o this distribution, the average Co concentration 2 versus d is given in the right-hand insert of figure 1 and it has the form 2 = cl exp (c2d) with cl = 0.75, c2 = 0.0074. The final fits of M, and
the magnetizations of the sublattices are indicated in figure 2 by the solid lines. The fitting parameters are Sco = 1,
ST^
= 5.1, J11= (82-
2.8) x 10-l4 erg, J 1 2 = -1 x 10-l5 erg, Jz2 = 1 x 10-l6 erg.2. We discuss the origin of the magnetic anisotropy Ku in Tb/Co within the framework of the dipolar in- teraction model (2). Ku possesses the form
P in the equation (2) is a measure of the alignment of the unlike nearest-neighbor atom pairs (2). Obvi- ously in Tb/Co multilayers, the mean value
P
in the bilayer decreases with the increase of the modulation period, but Z is increasing as given by the above fit- ting procedure. As a result, we suppose that P should be related to 2 in the Ku fitting process for Tb/Co.Taking ~ / N D - ~ (the packing fraction) = 0.7, Z = 12 and substituting MI, M2, x l and 2 2 for each d to the
equation 2, the fitted K , for the ambient T substrate are denoted by the solid line in figure 1. It could be seen that the agreement is quite good. The param- eter P us. d is given in the upper insert of figure 1. It is clear from this plot that
P
dereases rapidly with increasing d. Because P (not average here) at the in- terface is always large, we conclude that the perpen- dicular anisotropy arises mainly from the interface.3. From the above discussion, it can be seen that the magnetic properties are closely related to the atom ar- rangement and/or orientation is space. For the differ- ences between two temperature substrates, we could attribute it to the same reason. In the water-cooled substrate Tb/Co, the atom arrangements are much more regular, that is, relatively perfect in periodicity, consequently, the P parameter describing the unlike atoms along the z direction should be larger for the water-cooled substrate than that of ambient one with the same d because of the thermal diffusion. Therefore, K , for the water-cooled substrate is naturally larger.
The Kerr effect is related to the material properties and determined by EX,, the off-diagonal part of the complex dielectric tensor: It is believed that EX, is de-
pendent on the state of the substance as well as on its structure and chemical composition. Therefore, that $k is different for Tb/Co with different T substrate may be true because their structure, local chemical composition or stress are different. After relaxation, these difference may gradually disappear, then their $k values are not much different, which is indicated in the lower insert of figure 1.
We suppose that the existence of the mismatch be- tween two sublattices seems to be responsible for the structural relaxation that exists through atomic ther- mal diffusion. Because this mismatch appears at the interface, we call this relaxation interface relaxation. Acknowledgments
We are grateful for financial support of the Chinese Academy of Sciences and the National Science Foun- dation under grant NSF-INT-8715441.
[I] Hasegawa, R., J. A p p l . Phys. 46 (1975) 5263. [2] Mizoguchi, T., Cargill, G. S., 111, J. A p p l . Phys.