• Aucun résultat trouvé

MAGNETISM AND HYPERFINE INTERACTIONS OF Fe/W(110) AND Ag-COVERED Fe/W(110)

N/A
N/A
Protected

Academic year: 2021

Partager "MAGNETISM AND HYPERFINE INTERACTIONS OF Fe/W(110) AND Ag-COVERED Fe/W(110)"

Copied!
3
0
0

Texte intégral

(1)

HAL Id: jpa-00229007

https://hal.archives-ouvertes.fr/jpa-00229007

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.

MAGNETISM AND HYPERFINE INTERACTIONS

OF Fe/W(110) AND Ag-COVERED Fe/W(110)

Soon Hong, A. Freeman, C. Fu

To cite this version:

(2)

JOURNAL DE PHYSIQUE

Colloque C8, Suppl6ment au no 12, Tome 49, dbcembre 1988

MAGNETISM AND HYPERFINE INTERACTIONS OF Fe/W(110) AND Ag-COVERED Fe/W(110)

Soon C. Hong (I), A. J. Freeman (I) and C. L.

Fu

(2' 3,

(I) Department of Physics and Astronomy, Northwestern University, Evanston, Illinois, 60208 U.S.A. (2) Department of Physics and Astronomy, Northwestern University, Evanston, Illinois, 60208 U.S.A. (3) Metals and Ceramics Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee, 37831 U.S.A.

Abstract. - Results of highly precise full potential linearized augmented plane wave (FLAPW) calculations of the magnetism and magnetic hyperfine fields for Fe/W(110) and Ag covered FefW(110) are presented and compared with recent conversion

-

electron Mossbauer spectroscopy (CEMS) measurements. Total energy studies are used to determine interlayer spacings. The effect of intedayer relaxation is to change considerably the calculated magnetic hyperfine fields and to result in excellent agreement with the CEMS experiments.

Magnetic liyperfine fields, Bhf, for both the clean and Ag-covered Fe overlayers on W(110), have been measured by conversion-electron Mossbauer spec- troscopy [I]. The measured JBhiJ for Fe/W(110) are significantly lower than that for bulk Fe despite the fact that the magnetic moment of the Fe overlayers on W(110) have essentially the same magnitude as that of bulk Fe. To understand the physical origin of the hyperfine field lowering, we have investigated the origin of the observed drastic reduction of the mag- netic hyperfine field by means of a theoretical study of the electronic and magnetic properties of a relaxed Fe monolayer on W(110) and a relaxed Ag layer on Fe/W(110) using the highly precise full potential lin- earized augmented plane wave (FLAPW) [2] method within the local spin density approximation. We ap- proximate these systems by a single slab consisting of five layers of W(110) plus a monolayer of Fe (and Ag for the Ag-covered case) on each side. The two dimen- sional lattice parameter and the W-W interatomic dis- tance are taken to be those of bulk W. The interlayer spacings of Ag-Fe and Fe-W are determined from total energy calculations.

For the unrelaxed Fe/W(110), we found the hy- bridization of Fe- and W-bands to be very important

[3] in determining the magnetism of the Fe atoms. Hence we need to first determine the Fe-W interlayer spacings of the clean FejW(110) and the Ag-covered Fe/W(110). We find that the Fe atom in the clean and Ag-covered Fe/W(110) are relaxed downward by 9.5 %

and 4 %

,

respectively, compared to the unrelaxed Fe- W interlayer spacing (determined from the average of the Fe-Fe and W-W bond lengths in their bulk).

The resuls for the relaxed clean and Ag-covered Fe/W(110) and the unrelaxed Fe/W(110) show that the W(110) substrate, unlike the effect of noble metal substrates [4-61, reduces greatly the magnetic moments (2.18 and 2.17 p~ for the relaxed clean Fe/W(110) and Ag-covered Fe/W(110), respectively) compared to that

(2.98 p ~ ) of a free monolayer Fe(ll0) with the same two dimensional lattice constant as that of Fe/W(110) and that (2.65 p ~ ) of the Fe(ll0) surface layer. By contrast, the Fe monolayer on noble metal substrates retains almost the same magnetic moment as the un- supported Fe monolayer. Hence the reduction of the Fe magnetic moment on the W substrate implies that the hybridization between the Fe-d and W-d bands plays an important role in determining the Fe mag- netism. This hybridization also causes the magnetic moment of Fe in the relaxed Fe/W(110) to be reduced greatly compared to that (2.56 p ~ ) of the unrelaxed Fe/W(110).

Recent experiments with CEMS [I] did not ob- serve any spontaneous magnetization for the clean Fe/W(110) at room temperature. However, their low temperature results extrapolated to 0 K shows that the magnetic moment is almost the same as that of bulk Fe - a result that is consistent with our predictions.

The electronic spin density at the nucleus gives rise to the Fermi contact hyperfine field (H,) which is sub-

stantially larger than the contributions from any un- quenched angular momentum and dipolar fields. The calculated H, values for the relaxed clean and Ag-

covered Fe/W(110), and the unrelaxed FejW(110) and their decomposition into core and conduction electron (CE) contributions are presented in table I. As found for many bulk and surface systems, the core contri- bution scales precisely with the magnetic moment. Hence, given the same magnetic moment, the contri- butions from core electrons for both relaxed clean and Ag,covered FejW(110) are almost the same as that of bulk Fe. However, the contribution from the 4s con- duction electrons is strongly affected by their environ- ment [7].

In the bulk, the contribution from CE polarization is negative due to their indirect (covalent) polarization. The hyperfine fields of Fe atoms in the surface and interface layers, however, show positive contributions

(3)

C8 - 1684 JOURNAL DE PHYSIQUE

from the CE due to the direction polarization

-

similar to the case of free atoms. The large positive CE contri- bution for clean Fe/W(110) greatly reduces the mag- nitude of the total Fermi contact term (-148 kG'auss) compared to that (-350 kGauss) of bulk-like Fe. Our calculations and the CEMS experiment confirm a strik- ing demonstration that at surfaces and interfaces, the total hyperfine field is not proportional to the magnetic moment.

The Ag-covering produces an additional negative in- direct CE polarization of the s-like conduction elec- trons which when added to the positive CE polariza- tion results in enhancement of the magnitude of the total contact hyperfine field. The amount of enhance- ment (29 kGauss) is very consistent with the CEMS experimental value (20 kGauss).

Significantly, the CE contribution does not change with the Fe-W interlayer spacing; the CE contribu- tion of the relaxed Fe/W(110) is almost the same as that of the unrelaxed Fe/W(110) despite the large re- laxation (cf. Tab. I). Now the magnetic moment was found t o decrease with the Fe-W interlayer spacing due to the strong hybridization between Fe3d and W-5d bands. Hence, the total Fermi contact term also varies linearly with the Fe-W interlayer spacing. However, as expected, the Ag-Fe interlayer spacing (and the exact site of the Ag overlayer) does not sig- nificantly affect the magnetism of the Fe, including magnetic moment (AM = 0.1 ,uB) and Fermi contact term (AH, = 4 kGauss) in spite of the relatively large change (0.7 a.u.) of the interlayer spacing.

Table I. - Theoretical layer-projected magnetic mo- ments (in p ~ ) and magnetic contact hyperfine fields (in

kG-auss) broken down into core and

CE

contributions for the relaxed Fe/W(110) and Ag-covered Fe/W(l lo),

and the unrelmed Fe/W(110).

While our calculated value of the Fermi contact term is consistent qualitatively with the hyperfine field of the CEMS experiment, its magnitude is still higher than the experiment. Hence, we need to consider the remaining positive dipolar and unquenched angular momentum contributions which would, of course, lead the theoretical results for the hyperfine fields to be in better agreement with experiment. Furthermore, both terms are expected to be enhanced at a surface or at an interface. Of course, the dipolar term depends on the surface spin anisotropy. Now, if the direction of spin quantization is in the surface plane [8], and us-

ing a crude classical point dipole approximation, the calculated dipolar term is 11 kGauss. By contrast, if the spin direction is perpendicular to the surface plane then the same crude calculation gives

-

3 kGauss for the dipolar contribution.

The unquenched orbital angular momentum contri- bution (which arises from spin orbit interaction) can be much larger than the dipolar contribution. Writing this as [9]:

unrelaxed Fe/W(110) relaxed Fe/W(110) relaxed Ag/Fe/W(110)

AHorb = 125 X Ag (rF3) (in kGrauss)

a) Reference [3]

Total -194

-14sa

-177

where

(T--~)

is in a.u. and Ag is the gshift, then with

[lo]

Ag ci 0.09 and [9] (r-3) = 3.82 a.u.; AH,,b =

44 kGauss. Taken together with the dipolar contribu- tion (+ 11 kGauss or

-

3 kGauss) and the contact value (- 148 kGauss) gives total hyperfine field of

-

93 or

-

107 kGauss which is remarkable agreement with the CEMS experiment -especially in view of the crude- ness of the various approximations.

Core/M -138 -141 -140 Acknowledgments CE 159 158 127 Mag. Mom. 2.56 2.18 2.17

Work at Northwestern University support by the Of-

fice of Naval Research (Grant No. N00014-81-K-0438 and a computing grant at the Naval Research Labc- ratory Supercomputing Center). Work at Oak Ridge national Laboratory sponsored by the Division of Material Sciences, U.S. Department of Energy, under contract DE-AC05-840R21400 with Martin Marietta Energy Systems, Inc.

Core -353 -306 -304

[I] Przybylski, M. and Gradmann, U., Phys. Rev. Lett. 59 (1987) 1152 and J. Appl. Phys. 63

(1988) 3652.

[2] Wimmer, E., Krakauer, H., Weinert, M. and Free- man, A. J., Phys. Rev. B 24 (1981) 864 and references therein.

[3] Hong, Soon C., Fu, C. L. and Freeman, A. J., Bull.

Am. Phys. Soc. 32 (1987) 583 and unpublished. [4] Fu, C. L. and Freeman, A. J., Phys. Rev. B 33

(1986) 1611.

[5] Fu, C. L., Freeman, A. J. and Oguchi, T., Phys.

Rev. Lett. 54 (1985) 2700.

[6] Richter, R., Gay, J. G. and Smith, J. R., Phys.

Rev. Lett. 54 (1985) 2704.

[7] Freeman, A. J., Fu, C. L., Weinert, M. and Ohnishi, S., Hyperfine Interactions 33 (1987) 53 and references therein.

[8] Gradmann, U., Korecki, J. and Waller, G., J.

Appl. Phys. A 39 (1986) 101.

[9] Freeman, A. J. and Watson, E. R., Magnetism Ed. G. T. Rado and H. Suhl (Academic, New York) 1965, IIA, 167.

[lo]

Heinrich, B., Urguhart, K. B., Arrott, A. S., Cochran, J. F., Myrtle, K. and Purcell, S. T.,

Références

Documents relatifs

P. THE HYPERFINE INTERACTIONS IN THE INVAR ALLOYS ON Fe-Pt BASE.. SATBAJEV State University, Moscow, U. Les spectres Mossbauer obtenus montrent une separation des raies

The temperature dependence of the iron hyperfine field (proportional to Fe sublattice magnetization) in SmFeO, has already been measured and compared with various

Results indicate that the carbon concentration has a pronounced effect on the coupling of the magnetic ions but has little effect on the magnitude of the magneticmoment in

- The variation of the hyperfine field as iron nuclei HFe, the magnetic moment on iron atoms 03BCFe and the average magnetic moment fi for iron alloys Values are

Le clustering de I'activite IHBa dans les feuilles de nickel, cause par une grande dose d'implantation, est probablement responsable du manque d'interaction

As the frac- tion of substitutionally implanted nuclei differs considerably from unity, also at rather low xenon concentrations, a determination of the hyperfine

The hyperfine fields of the saturated ferro- magnetic state were all obtained by extrapolating the high field data to zero applied field.. For comparison, a high

Une comparaison de parametres Mossbauer de Il9Sn4f dans FeF3 avec ceux obtenus par des auteurs anterieurs pour ll9Sn4+ dans Y 0,9Cao, ~ F e o , 9Sn0,103 suggkre un