Elsevier

Thin Solid Films

Volume 446, Issue 1, 1 January 2004, Pages 6-11
Thin Solid Films

Fe3GaAs/GaAs(0 0 1): a stable and magnetic metal-semiconductor heterostructure

https://doi.org/10.1016/S0040-6090(03)01232-XGet rights and content

Abstract

We show that in agreement with the ternary Fe–Ga–As phase diagram, the solid-state interdiffusions in epitaxial Fe/GaAs(0 0 1) heterostructures lead, at a temperature of approximately 500 °C, to the formation of thermodynamically stable Fe3GaAs/GaAs(0 0 1) contacts quite similar to the well-known silicide/Si ones. The Fe3GaAs films are made of grains epitaxial on GaAs with a well-defined interface. Their magnetic and electrical properties make Fe3GaAs on GaAs an attractive metallization scheme for future magnetoelectronic devices. The results we report concern (25 or 80 nm Fe)/GaAs(0 0 1) heterostructures annealed at 480 and 500 °C for 10 min and characterized ex situ by He+ Rutherford backscattering and ion channeling, X-ray diffraction, transmission electron microscopy and alternating gradient field magnetometry.

Introduction

The growth of ferromagnetic films on semiconductor substrates has received much attention over the past decade. The combination of the magnetism of the metallic layer along with the electronic properties of the semiconductor substrate offers the opportunity for innovative magnetotransport effects and devices to be realized. However, because the two types of materials are very dissimilar in terms of physical, chemical and structural properties, integration of semiconductors and ferromagnets is generally very difficult [1], [2], [3], [4].

The (ferromagnetic metal)/GaAs systems have been widely studied in the last few years, in particular the epitaxial Fe/GaAs heterostructures [3], [4]. However, due to a strong reactivity between ferromagnetic metals and GaAs, the resulting structures are not thermodynamically stable. The purpose of this study is to obtain epitaxial and stable (magnetic compound)/GaAs structures by inducing a controlled thermal interaction between a metal film and a GaAs substrate. Many studies have been dealing with the metal/GaAs solid-state interdiffusions during annealing treatments but, as in the cases of Co/GaAs [5] and Ni/GaAs [6], [7] contacts, the final step of the interaction was always found to be a mixture of binary or ternary compounds [1], [2], [5], [6]: thermodynamically stable and epitaxial contacts equivalent to the CoSi2/Si, NiSi2/Si and CrSi2 heterostructures [8] have never been obtained for GaAs in this way. In this context, relatively few works concern the interdiffusions in Fe/GaAs structures. Rahmoune et al. [9], [10] annealed Fe/GaAs contacts prepared by ion-beam sputtering for 1 h at 500 °C and they observed a layered microstructure of the type Fe3Ga/(Fe2As+FeAs)/GaAs. However, no accurate studies of the final step of the solid-state interdiffusions in Fe/GaAs heterostructures, prepared and annealed in ultra-high vacuum (UHV) conditions, have been published.

Our starting point was the recent publications, concerning on one hand, the new magnetic ternary phases Fe3Ga2−xAsx [11], [12], [13], and on the other hand, the experimental determination of the Fe–Ga–As ternary phase diagram [14], [15]. The Fe3GaAs compound was identified for the first time by Harris et al. [11], in the form of precipitates occurring during the growth of Fe-doped GaAs single crystals using the liquid-encapsulation Czochralski technique. After that, the same authors prepared bulk Fe3GaAs samples by the same technique [12]: Fe3GaAs was identified as a NiAs-type phase (hexagonal B82-type) that exists over a range of compositions represented by the formula Fe3Ga2−xAsx (0.20⩽x⩽1.125) and presents a (2a,c) superstructure for 0.20⩽x⩽0.85. All the alloys were found to be ferromagnetic at temperatures higher than the ambient one. More recently, Fe3GaAs precipitates were also observed in Fe-implanted and annealed GaAs [16] and in Fe-doped GaAs grown by molecular beam epitaxy (MBE) [17].

Concerning the ternary Fe–Ga–As diagram, Députier et al. [14], [15] determined the solid-state phase equilibria at 600 °C and showed the occurrence of tie-lines between all the compositions of the Fe3Ga2−xAsx solid solution (in particular Fe3GaAs) and GaAs. To our knowledge, it was the first time that a ternary phase Mx(GaAs)y (M=transition metal) stoichiometric in Ga and As was found to be in thermodynamic equilibrium with GaAs. This fact suggests that the final step of the solid-state interdiffusions occurring during the annealing of a Fe/GaAs heterostructure could lead to a stable Fe3GaAs/GaAs contact. Moreover, one could expect an epitaxial growth of the Fe3GaAs layer on GaAs due to the particular structural properties of the hexagonal Fe3Ga2−xAsx alloys. Indeed, firstly, the c/a ratio of their crystalline parameters is close to √3/√2 (1.23⩽c/a⩽1.28 for 0.20⩽x⩽1.125), which allows them to be described in a pseudocubic unit cell with a parameter (apseudocubic=(3/√2)×ahexagonal), and secondly, the corresponding apseudocubic parameter is close to one and a half that of GaAs (within 2%). Such epitaxial growths by solid-state interdiffusion were previously observed for several equivalent hexagonal pseudocubic compounds [1], [2], [6], [7].

In a previous study we showed that the solid-state interdiffusions at 450 °C in epitaxial Fe/GaAs(0 0 1) heterostructures lead to the formation of an epitaxial reaction layer made of Fe2As patches embedded in a Ga-rich Fe3Ga1.8As0.2 ternary phase [18], [19] and that it is possible to regrow an epitaxial Fe film on such reacted layers at room temperature (RT) [20]. In this paper, we show that an annealing at a temperature higher than 450 °C allows a further interaction with the substrate and induces the formation of a Fe3GaAs/GaAs(0 0 1) contact. The results that we report concern (25 or 80 nm Fe)/GaAs heterostructures annealed at 480 and 500 °C at which the most significant phenomena of the solid-state interdiffusion were observed.

Section snippets

Experiments

The Fe/GaAs samples were prepared in a RIBER 2300 MBE system connected by an UHV modutrack system to another UHV chamber used for metal evaporations. Five hundred-nanometers-thick undoped GaAs buffer layers were first grown on n+-type GaAs(0 0 1) substrates using standard MBE conditions, and As-rich (2×4)-reconstructed surfaces were prepared. Fe depositions were done at RT using a high-temperature effusion cell with an alumina crucible at a deposition rate equal to 1 nm min−1. The UHV annealing

(80 nm Fe)/GaAs(0 0 1) heterostructure annealed at 480 °C

The random RBS spectrum shown in Fig. 1, which was obtained with a 2.2 MeV He+ beam, is well simulated by a Fe3GaAs/GaAs structure, with a ternary compound layer containing 67×1016 Fe cm−2 as the 80-nm-thick deposited Fe layer.

The XRD pattern on the unannealed sample (not shown) confirms the epitaxial growth of Fe on GaAs as previously reported [14], [18]. Only one 0 0 2 Fe reflection is observed with those of the substrate. After the annealing at 480 °C (Fig. 2a) the Fe reflection disappears,

(80 nm Fe)/GaAs(0 0 1) heterostructure annealed at 500 °C

After the annealing at 500 °C, the XRD diagram (Fig. 2b) shows that the reflections of the ternary compound are preserved, while the Fe2As ones have totally disappeared. However, the 1 0 1 and 2 0 2 reflections have now slightly shifted towards the low angles with respect to the annealing at 480 °C (2θ=31.2° and 65.2°, respectively), which suggests a decrease in the arsenic concentration in the ternary phase [15]. This is confirmed by the RBS random spectrum shown in Fig. 5, which is well simulated

Conclusion

In this study, we have demonstrated that the UHV deposition and annealing treatments in the 500 °C temperature range of Fe/GaAs(0 0 1) contacts allow the formation of stable Fe3GaAs/GaAs(0 0 1) heterostructures, quite equivalent to the well-known CoSi2/Si, NiSi2/Si and CrSi2/Si ones [8]. This result is in agreement with the ternary phase diagram [14], [15], which predicts that Fe3GaAs and GaAs are thermodynamically stable when in contact with each other. The Fe3GaAs films show an electrical

Acknowledgements

We would like to acknowledge the ‘Equipe des surfaces et des interfaces’ for their help and the many fruitful discussions we had together, Antoine Filipe and Alain Schuhl of the UMR Thales-CNRS (Orsay) for their help during the magnetic characterization of the samples and Jacques Crestou for the specimen preparation for TEM observations. This work was supported by the CNRS (‘Structure coopérative autour de l'accélérateur 2.5 MV du Groupe de Physique des Solides’ Universités Paris VI et VII).

References (22)

  • C.J. Palmstrom et al.
  • T. Sands, C.J. Palmstrom, J.P. Harbison, V.G. Keramidas, N. Tabatabaie, T.L. Cheeks, R. Ramesh, Y. Silberberg, Mat....
  • G.A. Prinz

    Science

    (1990)
  • G.A. Prinz
  • C.J. Palmstrom et al.

    J. Appl. Phys.

    (1987)
  • A. Guivarc'h et al.

    J. Appl. Phys.

    (1989)
  • B. Guenais et al.

    J. Cryst. Growth

    (1990)
  • R.T. Tung et al.

    Thin Solid Films

    (1982)
  • M. Rahmoune et al.

    Thin Solid Films

    (1996)
  • M. Rahmoune et al.

    J. Magnetism Magnetic Mater.

    (1997)
  • I.R. Harris et al.

    J. Cryst. Growth

    (1987)
  • Cited by (5)

    View full text