Torsion oscillation magnetometry (TOM) of Fe films on Ni(1 1 1)/W(1 1 0) substrates

https://doi.org/10.1016/S0304-8853(98)00380-1Get rights and content

Abstract

Fe films 2–20 atomic monolayers (ML) thick have been deposited on Ni(1 1 1) films (4–60 ML) prepared on W(1 1 0) under UHV conditions. The Ni(1 1 1) films grow in a Nishijama–Wassermann orientation with a 3.6% lattice expansion along Ni[2̄ 1 1] (‖ W[1̄ 1 0]). On top of these slightly distorted Ni films iron is observed to grow preferentially in two mirror orientations. The saturation magnetization of these Fe films and the anisotropies of Fe/Ni bilayers have been studied using torsion oscillation magnetometry. The magnetization of the Fe films of 2.13 μB per atom is close to the bulk value of Fe. The out-of-plane surface (interface) anisotropy constant KFeNis=(−0.65±0.15) mJ/m2 of the bilayers prefers a magnetization perpendicular to the surface. The exceptionally high value of the volume anisotropy constant of the Fe films, KFev=(+0.71±0.10) MJ/m3=(52.2±7.5) μeV/atom, was ascribed to magnetoelastic contributions.

Introduction

During the last decade much attention has been paid to the influence of the crystallographic structure of magnetic thin films on their magnetic properties [1]. Especially the behavior of FCC Fe films stabilized over a few atomic layers on single crystalline FCC substrates has attracted much experimental [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14] and theoretical [15], [16], [17], [18], [19] interest. While bulk-like FCC Fe, only formed at high temperatures T>1183 K or precipitated in a Cu matrix (aFCCFe=3.59 Å), does not show any ferromagnetic behavior, distinct ferromagnetic states have been observed in epitaxial FCC Fe films. Theoretical calculations for FCC Fe crystals predict a non-magnetic state for small lattice constants (<3.61 Å) switching to a ferromagnetic state with a high moment of 2.6 μB (high-spin state) for larger lattice constants (>3.62 Å) [16]. Therefore, it is a challenge to imprint or even to trigger the structure of thin Fe films using appropriate substrates and thus to modify their magnetic properties.

The isotropic FCC lattice considered in theory cannot be realized in epitaxial films where the cubic lattice structure generally shows a distortion. Of special interest have been Fe films on Cu substrates (aCu=3.615 Å) because of the particularly small lattice mismatch. On Cu(1 0 0), FCC Fe films (less than 4 ML thick) exhibit a high-spin phase [4], [7], whereas those deposited on Cu(1 1 1) show a low spin state (0.6 μB) [2], [14]. The dependence of the magnetic Fe moment on the lattice constant has been impressively demonstrated for thin (<3 ML) FCC Fe(1 1 1) films grown on (1 1 1) surfaces of CuAu alloys [3]. With the substrate lattice parameter increasing by changing the Au concentration of the alloy, there was a distinct rise of the mean Fe moment from about 0.6–2.6 μB. Recently, a similarly high spin state was deduced from Kerr measurements for FCC films prepared on a pure Cu(1 1 1) surface by pulsed laser deposition [14].

It is generally believed that a compression of the FCC Fe lattice results in a switch to the antiferromagnetic or non-magnetic state, see, e.g., Ref. [16]. Therefore, the application of Ni as a substrate for epitaxial Fe films is of special interest. Unlike copper, Ni exhibits two important features: (i) Ni is distinguished by the smallest lattice parameter of all FCC metals (aNi=3.524 Å), i.e., about 2.5% smaller than that of Cu, and could therefore be expected to stabilize thin FCC Fe films with a very small atomic volume (the lattice parameter of FCC Fe at room temperature (RT) was predicted [17] to be 3.55…3.58 Å). (ii) Contrary to copper, Ni substrates are ferromagnetic, i.e., there occur polarization and coupling effects between the constituents of the Fe/Ni bilayer. In polycrystalline (<32 Å)Fe/(70 Å)Ni bilayers grown on Si(1 0 0) an FCC-like structure of Fe films was deduced even up to a thickness of about 32 Å of the Fe overlayer [11]. Moreover, what is of even more interest, these Fe films show an averaged magnetic Fe moment of μFe=(0.2±0.2) μB. Therefore, these Fe films are assumed to be either antiferromagnetically ordered, or to exhibit almost vanishing Fe moments in agreement with the theoretical prediction. In contrast to that, for Fe films on Ni(1 1 1)/W(1 1 0) Sander et al. [12] proved a Kerr signal increasing with the Fe thickness.

In the present study, both the magnetic moment and the anisotropy of Fe/Ni bilayers have been studied by torsion oscillation magnetometry (TOM).

Section snippets

Experimental

The samples have been prepared and characterized in situ in a UHV torsion oscillation magnetometer [20]. The base pressure was less than 6×10−11 Torr. The UHV chamber was supplied with a combined low energy electron diffraction (LEED)/Auger electron spectroscopy (AES) system to characterize both the substrate surface and the growth of thin films on top of it. BeO crucibles were used for the thermal sublimation of Ni and Fe. The applied deposition rates (about 0.6 ML/min for Ni and 0.7 ML/min for

Structure of the films

The structures of the deposited films have been analyzed by LEED. In agreement with previous investigations on Ni/W(1 1 0) [12], [21], [22] a pseudomorphic growth of Ni has been observed up to about 0.4 ML. Thicker Ni films are known to grow in a Nishiyama–Wassermann (NW) orientation (Ni[0  1]‖W[0 0 1] and Ni[2̄ 1 1]‖W[1̄ 1 0]). A result typical of Ni films thicker than 3 ML is shown in Fig. 1, where the LEED spots of a W(1 1 0) surface are directly compared with those of a 6.0 ML Ni film. For this

Torsion oscillation magnetometry

To characterize the magnetic properties of thin Fe/Ni bilayers torsion oscillation magnetometry (TOM) was performed at RT. The main features of TOM have been described previously in detail [20], [25]. Therefore, solely some basics and the main components of the applied magnetometer are briefly mentioned here. In TOM experiments, small amplitude oscillations of the sample suspended on a thin torsion filament are studied to determine the magnetic torque constant R as a function of the strength of

Magnetic results

A typical result of our TOM experiments is shown in Fig. 5. Immediately after the Ni deposition (here 20.6 ML) we determined the magnetic torque constant R(H) for several fields. To simplify matters we plotted R/H as a function of H, with R/H being a rough measure of the magnetic moment of the sample (see Eq. (5)). The magnetic moment mNis(=JNisVNi) and the anisotropy field HNianis resulting from a least squares fit are given in the figure. Then, on top of the Ni film we deposited a Fe film of

Discussion

The magnetic moment of the bilayer samples determined by TOM linearly increases with their Fe film thickness. The absolute value of the atomic Fe moment is close to the RT bulk value of μFe(bulk)=2.17 μB [28]. Thus, in the present study of Fe/Ni bilayers on W(1 1 0) there is no indication of a lowered magnetic moment expected to appear in thin Fe films of a reduced atomic volume as observed in Ref. [11] for polycrystalline Fe(<30 Å)/Ni(70 Å) bilayers on Si(1 0 0). The atomic Fe volume could not be

Summary

In the present paper the magnetic properties of Fe films (>2 ML) grown on bulk like Ni(1 1 1) films deposited on W(1 1 0) have been studied by UHV torsion oscillation magnetometry. Due to the reduced symmetry of the slightly distorted Ni(1 1 1) films (NW orientation, +3.6% misfit along Ni[2̄ 1 1]‖W[1̄ 1 0]) the subsequently deposited Fe films do not grow pseudomorphically on the Ni(1 1 1) film but in two equivalent structural domains showing mirror planes of symmetry along Ni[2̄ 1 1] and Ni[0  1]. It has not

Acknowledgements

The support by the Deutsche Forschungsgemeinschaft of the present work, as the result of a several months’ working stay of H.H. at the University of Mainz, is gratefully acknowledged. The authors would like to thank Prof. U. Gradmann and Prof. J. Kirschner for careful reading the manuscript.

References (28)

  • U. Gradmann et al.

    J. Magn. Magn. Mater.

    (1980)
  • M. Wuttig et al.

    Surf. Sci.

    (1995)
  • M. Zharnikov et al.

    J. Magn. Magn. Mater.

    (1997)
  • O.K. Andersen et al.

    Physica B

    (1977)
  • J. Kolaczkiewicz et al.

    Surf. Sci.

    (1984)
  • K. Wagner et al.

    J. Magn. Magn. Mater.

    (1997)
  • U. Gradmann in: K.H.J. Buschow (Ed.), Handbook of Magnetic Materials, vol. 7, Elsevier, Amsterdam, 1993, and references...
  • W. Kümmerle et al.

    Phys. Stat. Sol. (a)

    (1978)
  • R.D. Ellerbrock et al.

    Phys. Rev. Lett.

    (1988)
  • D. Li et al.

    Phys. Rev. Lett.

    (1994)
  • J. Thomassen et al.

    Phys. Rev. Lett.

    (1994)
  • J. Fassbender et al.

    Phys. Rev. Lett.

    (1995)
  • M. Straub et al.

    Phys. Rev. Lett.

    (1996)
  • Y. Li et al.

    Phys. Rev. B

    (1996)
  • Cited by (10)

    • Growth of thin films

      1999, Journal of Physics Condensed Matter
    View all citing articles on Scopus
    View full text