We obtain control of magnetic anisotropy in epitaxial (Ga,Mn)As by anisotropic strain relaxation in patterned structures. The strain in the structures is characterized using sophisticated X-ray techniques. The magnetic anisotropy before patterning of the layer, which shows biaxial easy axes along [100] and [010], is replaced by a hard axis in the direction of large elastic strain relaxation and a uniaxial easy axis in the direction where pseudomorphic conditions are retained. This strong anisotropy can not be explained by shape anisotropy and is attributed solely to lattice strain relaxation. Upon increasing the uniaxial strain anisotropy in the (Ga,Mn)As stripes, we also observe an increase in magnetic anisotropy.PACS numbers: 75.50. Pp, 75.30.Gw The (Ga,Mn)As material system has been the focus of many studies over the last years. As the understanding of its complex transport and magnetic properties increases, the focus of interest shifts from basic research towards its application in devices. For this, it is necessary to understand how different parameters influence the ferromagnetic material in structures at the device level. In this letter we present a systematic study of the role of strain relaxation as the dominating factor contributing to the magnetic anisotropy in (Ga,Mn)As nanostructures. A (Ga,Mn)As layer grown epitaxially on a GaAs substrate is subject to compressive strain in the plane of the sample and typically exhibits biaxial in-plane easy axes along [100] and [010], at temperatures around 4 K [1, 2]. In earlier studies, control of the magnetic anisotropy has been achieved by modifying the strain in the layer. Tensile strain can be imposed on the (Ga,Mn)As by growing it on a thick, plastically relaxed (In,Ga)As buffer with a larger lattice parameter, and results in an out-of-plane easy axis [3,4].Here, we follow an alternative approach in modifying the lattice strain of (Ga,Mn)As on GaAs by lithography, which allows us to locally control the magnetic anisotropy of the material. By structuring a fully pseudomorphic 70 nm (Ga,Mn)As layer into thin, elongated stripes, we allow anisotropic, elastic strain relaxation perpendicular to the long axis of the stripe. To increase the strain in the structure compared to the case of (Ga,Mn)As on GaAs, a second sample is processed which includes a highly compressively strained layer acting as an extra stressor to the overlying (Ga,Mn)As layer. The uniaxial strain relaxation in the structures is investigated by grazing incidence X-ray diffraction (GIXRD) and high-resolution X-ray diffraction (HRXRD). To determine the influence of patterning on the magnetic anisotropy, a series of magnetometric and magnetotransport studies are performed. We also present finite-element simulations of anisotropic strain relaxation and k ·p calculations which confirm the relationship between the structural and magnetic behavior observed in our samples.The samples are grown in a dedicated III-V MBE chamber with effusion cells for Ga, In, Mn, and a valved As 4 cell. A 200 nm thic...

The focus of studies on ferromagnetic semiconductors is moving from material issues to device functionalities based on novel phenomena often associated with the anisotropy properties of these materials. This is driving a need for a method to locally control the anisotropy in order to allow the elaboration of devices. Here we present a method which provides patterning induced anisotropy which not only can be applied locally, but also dominates over the intrinsic material anisotropy at all temperatures.The coupling of transport and magnetic properties in ferromagnetic semiconductors gives rises to many interesting anisotropy related transport phenomena such as strong anisotropic magnetoresistance (AMR), inplane hall [1], tunneling anisotropic magnetoresistance (TAMR) [2,3] and Coulomb blockade AMR [4]. Studies on all of these effects so far have primarily made use of the intrinsic anisotropy present in the host (Ga,Mn)As layer. Before they can be harnessed to their full potential, a means of engineering the anisotropy locally is needed, such that multiple elements with different anisotropies can be integrated, and their interactions can be properly investigated.One successful approach to local anisotropy control in metallic ferromagnets has been to make use of shape anisotropy. The same approach has been tried in the prototypical ferromagnetic semiconductor (Ga,Mn)As with lackluster results. In Ref. 5, the authors reported the observation of shape induced anisotropy in (Ga,Mn)As wires of 100 nm thickness x 1.5 x 200 µm 2 , but only over a limited temperature range. Moreover, our own experience in attempting to use wires of similar dimensions have yielded sporadic results with the wires having irreproducible anisotropy, with either biaxial or uniaxial easy axes in inconsistent directions.Furthermore, a simple calculation of the expected shape anisotropy term in such wires indicates that it should not play a significant role. While the infinite rod model used in [5] does predict an appreciable shape anisotropy field given by µ 0 M S /2, where M S is the sample magnetization, it is not applicable to structures which are much thinner than their lateral dimensions. A more exact rectangular prism calculation [6] gives a 5 times weaker shape anisotropy with an anisotropy energy density of 80 J/m 3 which is much too small to compete with the typical crystalline anisotropy of 3000 J/m 3 [5,7] in this material.Growth strain reduces the cubic symmetry of the (Ga,Mn)As zinc-blend crystal structure creating a uniaxial anisotropy with an easy/hard magnetic axis in growth direction when the (Ga,Mn)As layer is tensile/compressively strained.This growth strain is known[8] to influence the strength of the perpendicular component of the anisotropy of the whole layer. Here we discuss (Ga,Mn)As grown on a (001) oriented GaAs substrate, whose out-of-plane hard magnetic axis confines the magnetization in the plane. Phenomenologically, the net in-plane magnetic anisotropy is known to result from a competition of two primary contributions...

Progress in (Ga,Mn)As lithography has recently allowed us to realize structures where unique magnetic anisotropy properties can be imposed locally in various regions of a given device. We make use of this technology to fabricate a device in which we study transport through a constriction separating two regions whose magnetization direction differs by 90• . We find that the resistance of the constriction depends on the flow of the magnetic field lines in the constriction region and demonstrate that such a structure constitutes a non-volatile memory device.

The rich magnetic anisotropy of compressively strained (Ga,Mn)As has attracted great interest recently. Here we discuss a sensitive method to visualize and quantify the individual components of the magnetic anisotropy using transport. A set of high resolution transport measurements is compiled into color coded resistance polar plots, which constitute a fingerprint of the symmetry components of the anisotropy. As a demonstration of the sensitivity of the method, we show that these typically reveal the presence of both the [110] and the [010] uniaxial magnetic anisotropy component in (Ga,Mn)As layers, even when most other techniques reveal only one of these components.

We observe the occurrence of an Efros-Shklovskii gap in (Ga,Mn)As based tunnel junctions. The occurrence of the gap is controlled by the extent of the hole wave-function on the Mn acceptor atoms. Using k · p-type calculations we show that this extent depends crucially on the direction of the magnetization in the (Ga,Mn)As (which has two almost equivalent easy axes). This implies one can reversibly tune the system into the insulating or metallic state by changing the magnetization.PACS numbers: 71.30.+h, 75.30.Hx, 75.50.Pp A very direct way to observe the Efros-Shklovskii (ES) gap, the soft gap induced by Coulomb correlations near the Fermi level of a Mott insulator [1,2], is by means of tunnel spectroscopy. Such experiments were, e.g., performed on the (three-dimensional) nonmetallic doped semiconductor Si:B [3, 4] and on thin (two-dimensional) Be films [5]. While both of these experiments employed large area tunnel junctions and a metallic counter electrode, a more recent study employed Ge:As break junctions[6]. This latter approach avoids possible screening of the Coulomb correlations, but the mesoscopic character of the contact may complicate extraction of bulk Coulomb gap behaviour [7].We have recently investigated the physics of a novel type of magnetoresistance, dubbed TAMR (tunneling anisotropic magnetoresistance) [8,9]. TAMR results from the dependence of the density of states (DOS) in strongly spin-orbit coupled ferromagnetic semiconductors, such as (Ga,Mn)As, on the direction of the magnetization of the material. In [9] we reported a drastic (> 10 4 ) increase of the spin-valve signal in a (Ga,Mn)As/GaAs/(Ga,Mn)As tunnel structure on lowering the sample temperature from 4.2 to 1.7 K, and speculated that this behaviour might result from the opening of an ES gap. Here, we provide evidence that the high resistance state of the sample indeed corresponds to a soft-gapped Mott insulator. In these samples, the metalto-insulator transition (MIT) is driven by a large variation of the Bohr radius of a hole bound to a Mn-impurity when the magnetization of the layer is switched from one easy axis to the other. This assignment is supported by a k · p-type calculation of a hydrogen-like impurity in a ferromagnetic GaAs host, extending the successful mean field model for (Ga,Mn)As [11,12].Our (Ga,Mn)As tunnel structure is shown in Fig. 1b. From bottom to top, the Ga 0.94 Mn 0.06 As (100 nm)/ GaAs (2 nm) /Ga 0.94 Mn 0.06 As (10 nm) trilayer stack has been grown by low temperature molecular beam epitaxy (LT-MBE) on a semi-insulating GaAs substrate and a 120 nm undoped GaAs buffer layer. Both (Ga,Mn)As layers are ferromagnetic with an as-grown Curie temperature of ∼ 65 K and highly p-type due to the intrinsic doping arising from the Mn atoms.As seen in the optical micrograph of Fig. 1a, the layer stack is patterned into a square mesa of 100×100 µm 2 by positive optical lithography, metal evaporation, liftoff and wet etching. The top contact is in-situ Ti/Au. Contact to the lower (Ga,Mn)As layer is established by a W/...