1

Magneto-elastic coupling in epitaxial magnetic/semiconductor heterostructures

Part I Previous Track Record

Yongbing Xu has been a new lecturer at the University of York since early last year. He was a postdoctoral research fellow from 1997 to 2000 in the Thin Film Magnetism (TFM) group (headed by Prof. Tony Bland), Cavendish Laboratory at CambridgeUniversity working on magnetic ultrathin films and patterned mesoscopic structures. He had previously worked as a PhD student on a joint Leeds-York project, “Spin-resolved photoemission”, at the CLRC Daresbury Laboratory, from 1994 to 1996, under the joint supervision of Professors Denis Greig (Leeds) and Jim Matthew (York). Before that he was a lecturer in the National Laboratory of Solid State Microstructures at Nanjing University, China working on magnetic multilayers and magneto-optics.

He has worked on two EPSRC projects; “Evolution of the magnetic properties of nanoscale structures” (PI: Prof. Bland) and “Magnetic mesostructures” (PI: Prof. Bland) in Cambridge. His work was focused on the study of the correlation between the magnetic properties and the atomic scale structure of ultrathin ferromagnetic metal (FM) films on III-V semiconductors (SC) grown by molecular beam epitaxy (MBE). He found that the magnetic properties of Fe/GaAs(100)-4x6 at room temperature evolve via three phases; a non-magnetic phase for the first three and a half monolayers, a superparamagnetic phase, and a ferromagnetic phase above about five monolayers [1]. The realisation of bulk magnetic moments at Fe/GaAs and Fe/InAs interfaces was an important step toward the development of the next generation spin-electronic devices [1, 2]. A combined in-situ MOKE, real-time RHEED, and in-situ STM study of Fe/InAs(100)-4x2 demonstrated the correlation of in-plane uniaxial magnetic anisotropy and the intrinsic atomic scale structure of the reconstructed semiconductor surface [3, 4]. Along with graduate students, he has successfully developed two techniques, namely photon-excited spin-injection [5, 6] and in-situ magneto-resistance [7] to probe the spin-dependent properties of ultrathin films. He also studied extensively the micromagnetism and spin-dependent transport in patterned dot arrays and wires [8, 9, 10]. He has designed and demonstrated a simple cross-wire structure to trap domain walls for the study of the spin-dependent transport in mesoscopic magnetic system.

His PhD thesis in Leeds was “Spin-polarized photoemission on novel magnetic materials”, which was mainly carried out in CLRC Daresbury Laboratory working with Prof. Elaine Seddon. He studied systematically for the first time the spin-resolved density of states of three amorphous magnetic alloys FeB, CoB and FeY using a synchrotron radiation source [11, 12, 13]. The Leeds-York team was the first external users of the new spin-polarized photoemission facilities including a high-energy Mott polar and a compact “micro-Mott” detector in Daresbury. This work was introduced in the key Daresbury Laboratory annual report of 1995-1996, the one which highlights a few major achievement and developments in the laboratory throughout the year. The results have also been presented at several national and international conferences as talks and invited talks. His work in Nanjing was mainly about magneto-optics in magnetic/nonmagnetic multilayers carried out in the group headed by Prof. Hongru Zhai. He had studied systematically the thickness and substrate dependence of the magneto-optical Kerr effect in bilayer and multilayer films, and proposed a new procedure to optimize the signal to noise ratio in layered magneto-optical storage media [14, 15]. Using MOKE spectra and nuclear magnetic resonance (NMR), he found the spin-polarization of nonmagnetic metals Ag and Cu in magnetic multilayers Fe/Ag and Fe/Cu [16, 17].

In short, Xu has more than ten years of research experience in the fields of magnetic nanostructures, nanofabrication using MBE growth and lithography, and surface science. He has published over 70 papers in leading academic journals and given several talks and invited talks in national and international conferences. He was awarded an EPSRC advanced research fellowship, started in October 2000, to work on “Spin-electronics in mesoscopic magnetic materials”. York magnetism group now has four permanent academic staff working on a wide range of subjects from magnetic recording media to spin-electronic materials.

[1]Y. B. Xu, E. T. M. Kernohan, D. J. Freeland, A. Ercole, M. Tselepi and J. A. C. Bland, Phys. Rev. B58, 890 (1998).

[2]Y. B. Xu, D. J. Freeland, E. T. M. Kernohan, M. Tselepi, C. M. Guertler, J. A. C. Bland, S. N. Holmes, and D. A. Ritchie, IEEE Trans. on Magn. 35, 3661 (1999).

[3]Y. B. Xu, E. T. M. Kernohan, M. Tselepi, J. A. C. Bland, and S. N. Holmes, Appl. Phys. Lett. 73, 399 (1998);

[4]Y. B. Xu, D. J. Freeland, M. Tselepi and J. A. C. Bland, Phys. Rev. B, B62, 1167 (2000).

[5]A.Hirohata, Y. B. Xu, C. M. Guertler and J. A. C. Bland, Phys. Rev. B63, 104425 (2001).

[6]J.A. C. Bland, A. Hirohata, C. M. Guertler, Y. B. Xu, J. Appl. Phys., 89, 6744 (2001) ( invited talk)

[7]C. M. Gurtler, Y. B. Xu, J. A. C. Bland, J. Magn. Magn. Mater, 226, 655 (2001).

[8]Y. B. Xu, C. A. F. Vaz, A. Hirohata, J. A. C. Bland, F. Rousseaux, E. Cambril and H. Launois, J. App. Phys. 85, 6178 (1999).

[9]Y. B. Xu, C. A. F. Vaz, A. Hirohata, C. C. Yao, J. A. C. Bland, E. Cambril, F. Rousseaux and H. Launois, Phys. Rev. B61, R14901 (2000).

[10]Y. B. Xu, A. Hirohata, H. T. Leung, T. Tselepi, J. A. C. Bland, E. Cambril, F. Rousseaux and H. Launois, J. Appl. Phys. 87, 7019 (2000).

[11]Y. B. Xu, C. G. H. Walker, D. Greig, E. A. Seddon, I. W. Kirkman, F. M. Quinn and J. A. D. Matthew, J.Phys. Conden. Matter. 8, 1567 (1996)

[12] Y. B. Xu, D. Greig, E. A. Seddon and J. A. D. Matthew, Phys. Rev. B55, 11442 (1997)

[13]J. A. D. Matthew, E. A. Seddon, Y. B. Xu, Journal of Electron Spectroscopy and Related Phenomena 88, 171 (1998) (invited talk).

[14]Y. B. Xu, Q. Y. Jin, Y. Zhai, M. Lu, Y. Z. Miao, Q. S. Bie and H. R. Zhai, J. Appl. Phys. 74, 3470 (1993).

[15]Y. B. Xu, H. R. Zhai and M. Lu, J. Appl. Phys. 70, 7033 (1991).

[16]Y. B. Xu, H. R. Zhai, M. Lu, and Q. Y. Jin, Phys. Lett. A, 168, 213 (1992)

[17]Q. Y. Jin, Y. B. Xu, and H. R. Zhai et al, Phys. Rev. Lett. 72, 768 (1994).

Professor M.R.J.Gibbs has held a Personal Chair within the Department of Physics and Astronomy at the University of Sheffield since 1st October 1997. Prior to that he spent 10 years in the School of Physics at the University of Bath, before moving to Sheffield with effect from January 1994. Throughout his career, Professor Gibbs has studied the magnetic properties of materials. This began with studies of bulk amorphous ferromagnets (EPSRC grants GR/F52101, GR/F52064, GR/G45878 and GR/J97618) and device applications (EPSRC grants GR/D63448 and GR/E49562). More recently interests have extended to consider bulk nanophase soft magnetic materials (EPSRC grants GR/K72407 and GR/M45535), magnetic force microscopy (EPSRC grants GR/K18306 and GR/L42094), and magnetic thin films and related devices (EPSRC grants GR/J96475, GR/K55905 and GR/R00395).

A strong thread running through many of the research activities has been the study of magnetoelasticity. Early work studied fundamental aspects of magnetostriction in amorphous ferromagnets [1,2] and its measurement [3], together with the associated E effect [4]. Applications in magnetometry [5,6] and acoustic transducers [7] have also been considered. Since 1992 work has been undertaken to study magnetostriction in monolithic and multilayer films. These have been grown by magnetron sputtering in-house. Amorphous ferromagnetic films have been characterised [8,9] building on knowledge gained from bulk ribbons and wires. The control of induced anisotropy has been established [10], offering the possibility of applications in micro-electromechanical systems (MEMS) [11,12]. Magnetostriction in multilayers has also been studied [13,14], and a phenomenological interpretation of the data developed [15,16]. The measurement of magnetostriction when the sample is a thin film on a rigid substrate has also been explored, and a significant contribution made to the understanding of the cantilever deflection method [17]. This area forms a further part of the current proposal.

[1] / A.P.Thomas, M.R.J.Gibbs, J.H.Vincent and S.J.Ritchie, IEEE Trans.Mag. 27, 5247(1991).
[2] / A.P.Thomas and M.R.J.Gibbs, J.Magn.Magn.Mat. 103, 97 (1992).
[3] / P.T.Squire and M.R.J.Gibbs, J.Phys.E: Sci.Instrum. 20, 499 (1987).
[4] / P.T.Squire and M.R.J.Gibbs, IEEE Trans.Mag. 25, 3614 (1989).
[5] / P.T.Squire and M.R.J.Gibbs, IEEE Trans.Mag. 24, 1755 (1988).
[6] / D.Brugel, M.R.J.Gibbs and P.T.Squire, J.Appl.Phys. 63, 4249 (1988).
[7] / D.W.Rees, M.R.J.Gibbs and N.G.Pace, IEEE Trans.Ultrasonics, Ferroelectrics & Frequency Control 36, 332 (1989)
[8] / A.D.Mattingley, C.Shearwood and M.R.J.Gibbs, IEEE Trans.Mag. 30, 4806 (1994)
[9] / C.Shearwood, A.D.Mattingley and M.R.J.Gibbs, J.Magn.Magn.Mat. 162, 147 (1996)
[10] / M.Ali, R.Watts, W.J.Karl and M.R.J.Gibbs, J.Magn.Magn.Mat. 190, 199 (1998)
[11] / M.R.J.Gibbs, C.Shearwood, J.L.Dancaster, P.E.M.Frere and A.J.Jacobs-Cook, IEEE Trans.Mag. 32, 4950 (1996)
[12] / W.J.Karl, R.Watts, A.L.Powell, C.R.Whitehouse, R.B.Yates and M.R.J.Gibbs, Electro.Chem.Soc.Proc. 98-20, 354 (1999)
[13] / T.A.Lafford, M.R.J.Gibbs, R.Zuberek and C.Shearwood, J.Appl.Phys. 76, 6534 (1994)
[14] / T.A.Lafford, R.Zuberek and M.R.J.Gibbs, J.Magn.Magn.Mat. 140-144, 577 (1995)
[15] / T.A.Lafford and M.R.J.Gibbs, IEEE Trans.Mag. 31, 4094 (1995)
[16] / H.J.Hatton and M.R.J.Gibbs, J.Magn.Magn.Mat. 156, 67 (1996)
[17] / R.Watts, M.R.J.Gibbs, W.J.Karl and H.Szymczak, Appl.Phys.Lett. 70, 2607 (1997)

Part II Proposed project and its context

A.Introduction and background

The increasing miniaturization of storage media and electronic devices calls for magnetic materials having well defined and controlled magnetic anisotropy, coercivity and saturation fields at thicknesses down to the nanometer/atomic scale. Magnetic ultrathin films, with thicknesses on the nanometer/atomic scale, have become increasingly important for high-density magnetic storage [1] and spin-electronic devices [2, 3]. Patterned ultrathin dot arrays, for example, have the possibility of overcoming the superparamagnetic limit (or have a superparamagnetic limit at very small lateral dimensions), as the stray field is minimized, and the combination of anisotropies and the exchange interaction may keep the spins aligned parallel along a certain direction [1, 4, 5]. The use of ultra-thin ferromagnetic pads in spin-electronic devices [6-8] may reduce the dipole interaction and allow controllable switching of the pads, as well as minimizing the stray field in adjacent semiconductors.

There is growing evidence that the magnetic properties of ultrathin films may become dominated by interface phenomena such as lattice mismatch. One of the key intrinsic magneticparameters in this may be the magneto-elastic (ME) interaction, which links the structure and the magnetic properties of materials. The magneto-elastic coupling caused by even a small strain can change dramatically the magnetic anisotropy and the magnetization process of a magnetic thin film. For example, a 1% lattice mismatch in a cubic Fe system will introduce a magneto-elastic energy comparable to the magneto-crystalline energy. Recent studies [9-14] reveal that both the magnitude and the sign of the magneto-elastic coupling coefficient in ultrathin films could be distinctly different from that of their respective bulk materials. Sander et al [9, 10] found that the magneto-elastic coupling coefficient of Fe films grown on W depends on the film thickness, and even the sign is changed at around a thickness of 3nm. Even in amorphous alloys the magneto-elastic coupling was found to differ by more than a factor of three near the surface region [13]. These results suggest strongly that one cannot take for granted the bulk magneto-elastic constants in explaining the magnetic anisotropy observed in ultrathin films. However, understanding of the magneto-elastic interaction in ultrathin films is still poor, and very few systematic studies have been made.

Magnetic/semiconductor heterostructures are of growing interest for the study of fundamental magnetism of ultrathin films and for the development of spin-electronic devices. Within the context of spin-electronics, the electrons’ spins, not just their electrical charge, are controlled for the operation in information circuits. As the conventional solid-state electronic devices are based on semiconductors, the injection and manipulation of spin electrons in magnetic/semiconductor heterostructures may lead to the development of next generation spin-electronic devices for data storage and processing at the same time [2, 3, 14].The patterned ultrathin magnetic particles on semiconductors might also be promising materials for high-density magnetic storage media [1, 4, 5]. The stabilization of epitaxial bcc Fe on GaAs has been achieved in several groups [15-20]. Recently, Xu et al [19] demonstrated that high quality Fe films without interface magnetic dead layers can be achieved in the Fe/GaAs system. A new epitaxial magnetic/semiconductor, namely Fe/InAs, has also been synthesized despite the large lattice mismatch (5.6%) [21]. One of the important issues concerning the magnetic properties of the Fe/III-V semiconductor system is the origin of a uniaxial magnetic anisotropy, which can then be controlled and exploited in future devices. Take Fe/GaAs(100), for example: there is a strong uniaxial anisotropy, unexpected from the cubic anisotropy of bulk bcc Fe, in the ultrathin films below about 50ML [15-20]. The cubic anisotropy of the bulk phase can even disappear completely at a thickness around 7ML. The measurement of the magneto-elastic constant, which determines the balance of long range strain effects and short-range chemical effects (section E.a), in this project will give a crucial insight into this issue. Furthermore, the ferromagnetic metal/III-V semiconductor heterostructures provide a unique “laboratory” to study directly for the first time fundamental aspects (sections E.b, E.c, and E.d) concerning the magneto-elastic coupling in magnetic nanostructures because of the novel “tunable” lattice mismatch in this system.

B. Project objectives

  1. Determine unambiguously the origin of the uniaxial magnetic anisotropy by measuring the thickness dependence of the magneto-elastic constants in Fe/GaAs and Fe/InAs.
  2. Establish the relationship between the strain, the thickness, and the magneto-elastic coupling by exploring the unique tunable lattice mismatch between Fe and certain III-V semiconductors.
  3. Study the magneto-elastic coupling in metastable magnetic phases such as bcc Co films.
  4. Study the possible effect of nonmagnetic capping layers to the magnetic-elastic coupling in ultrathin magnetic films and it’s implication to the operation of practical devices.

C. Timeliness and novelty

A key aspect of this proposal is the new and close collaboration between York and Sheffield magnetism research groups with complementary expertise in MBE growth (Dr. Xu in York) and magneto-elastic measurements (Prof. Gibbs in Sheffield). There is growing evidence that the magneto-elastic constants in nanoscale magnetic materials are in general different from their bulk values in not only the magnitude but also possibly the sign. There has been little progress, however in understanding the basic physics of magnetoelasticity in nano scale materials as well as in tailoring the magnetoelastic interaction in devices. Ferromagnetic metal/semiconductor heterostructures are important materials for the development of spin electronic devices and the understanding of the magneto-electronic interaction in this system is crucial to control their magnetic properties. This proposal represents an opportunity to establish a world-class facility, ultimately with in-situ measurement capacity, for the study of magneto-elastic interaction in epitaxial magnetic thin films. In comparison with reported studies in the Fe/W system [9], ferromagnetic metal/semiconductor heterostructures and specifically Fe/GaIn1-xAsx provide a unique opportunity to establish directly the strain dependence of the magnetoelastic coupling and to probe for the first time the intrinsic size effect in nano scale magnetic materials. This proposal also provides support for a newly appointed academic (Xu) to further develop his research base. This is his first EPSRC proposal apart from the fellowship one.

D. Experimental methodology

a. Growth of ferromagnetic metal/III-V semiconductor (FM/SC) heterostructures

With funds from EPSRC, the university and the Department, Xu has recently built up a new UHV chamber in York especially for the growth of magnetic/semiconductor heterostructures. The growth system has three e-beam sources, which can be used for magnetic metals, such as Fe, Co, and FeNi, and nonmagnetic metals such as Au and Cr. The sample manipulator has incorporated an e-beam heater for the annealing of substrates and the removal of As capping layers to ensure a clean surface for growth. There is a shutter fitted with a linear driver for step growth. A set of samples of up to about ten different thicknesses can be fabricated in one growth run to ensure the same growth conditions. For ex-situ magnetic and magneto-elastic constant measurements, the magnetic films will be capped with a thin Au or Cr layer of about 3nm. With funds from this application we will purchase a RHEED system for monitoring the growth and structural analysis, and an Ar sputtering gun for further substrate cleaning when necessary. The RHEED system, in combination with magneto-elastic measurements will allow for a full understanding of the relation between structureand magnetic properties in ultrathin films.

b. Ex-situ and in-situ magneto-elastic measurements.

For the last ten years Professor Gibbs has been studying the magneto-elastic properties of films on a wide range of substrates. A number of techniques have been evaluated including small-angle-magnetisation-rotation, cantilever and bent substrate [22-24]. Only the bent substrate technique is appropriate in this context. The present facilities in Sheffield have been developed for the study of soft magnetic films (NiFe and FeCo) with anisotropy fields no higher than 5kAm-1. The technique will now be adapted to handle the field requirements of Fe/GaAs and Fe/InAs films by building a system around an electromagnet. Work on NiFe has shown excellent signal to noise ratio down to 3nm thickness of NiFe (see Fig.1), and thus the technique requires only the field adaptation to be viable. As the MOKE signal from Fe and Co is about three times larger than that from FeNi, good signal to noise ratio down to a few monolayers will be achieved in ultrathin Fe and Co films. Sheffield also has AFM/MFM facilities, which will be used to evaluate surface roughness contributions to the surface magnetism.It is important to work ex-situ (even though there will be a capping layer) not only to evaluate the technique and ensure correct implementation in UHV, but also to study extensively the effects of capping layers.

Ina parallel phase of the project, an appropriate form of the ex-situ Villari measurement method developed in Sheffield will be combined with York UHV growth chamber to develop an in-situ magneto-elastic measurement system. The key components for in-situ measurements are a magnet compatible with UHV and a sample holder capable of bending the sample in UHV. The magnet with field up to 240kAm-1 has been built up already. Fig. 2 shows a schematic diagram of the proposed in-situ measurement system with a purposely-designed sample holder based on the current low field ex-situ measurement apparatus in Sheffield.