Sensors based on giant planar hall effect in dilute magnetic semiconductors

Abstract

Ferromagnetic semiconductor-based sensor devices, including sensors for detecting pressure changes and sensors for detecting magnetic fields, such as switching events in a magnetic recording medium. The pressure sensors of the present invention detect pressure changes using magnetoresistive measurement techniques, and in particular GPHE techniques. Magnetic field detection sensors such as magnetic switching detection sensors include ferromagnetic semiconductor-based materials that provide enhanced sensitivity relative to known materials and techniques. Such magnetic switching detection sensors according to the present invention are particularly useful as a read head sensor for HDD and other magnetic storage technologies.

Description

STATEMENT AS TO RIGHTS TO INVENTIONS MADE UNDER FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT

[0002] Not Applicable

BACKGROUND OF THE INVENTION

[0003] The present invention relates generally to sensors, and more particularly to ferromagnetic semiconductor-based sensing devices and methods.

[0004] Ferromagnetic semiconductors are of considerable current interest since they offer prospects for realizing semiconducting spintronics devices that have no analogs in a metallic ferromagnetic system (see, e.g., G. Prinz, Science 282, 1660 (1998).; S. A. Wolf et al., Science 294, 1488 (2001).). One recent and striking example is the electric field control of ferromagnetism (see, H. Ohno et al., Nature (London) 408, 944 (2000).). Semiconductor-based magnetic materials also offer new possibilities for attaining great improvements in performance over metallic magnetic devices. Among the examples here is conductivity matching to attain efficient spin injection into semiconductors (see, e.g., G. Schmidt et al., Phys. Rev. B 62, R4790 (2000).). Semiconductor ferromagnetism also gives rise to new physical phenomena because it is possible to engineer, and enhance, spin-orbit coupling in ways that are not possible in metallic Systems.

[0005] Advances in nanofabrication and epitaxial growth bring new levels of control and resolution to the study of magnetic domains at the microscopic level. Recent research efforts are focused upon two principal thrusts: studies of domain wall dynamics, and investigations of domain wall resistance. After decades of concerted study on metallic ferromagnetic thin films, it is generally agreed that magnetization reversal within high quality magnetic microstructures occurs first by nucleation, then propagation, of domain walls (see, e.g., J. Ferré, Topics Appl. Phys. 83, 127 (2002)). As regards the latter, even though the resistance of domain walls is difficult to measure, it has become a topic of significant current interest (see, M. Viret, et al., Phys. Rev. Lett. 85, 3962 (2000); U. Ruediger, J. Yu, S. Zhang, A. D. Kent, S. S. P. Parkin, Phys. Rev. Lett. 80, 5639 (1998); L. Klein et al., Phys. Rev. Lett. 84, 6090 (2000); R. Danneau et al., Phys. Rev. Lett. 88, 157201 (2002); U. Ebels, A. Radulescu, Y. Henry, L. Piraux, K. Ounadjela, Phys. Rev. Lett. 84, 983 (2000); T. Taniyama, I. Nakatani, T. Namikawa, Y. Yamazaki, Phys. Rev. Lett. 82, 2780 (1999); and Y. B. Xu et al., Phys. Rev. B 61, R14901 (2000)). This is due to both the important role domain wall resistance might play in advanced spintronic devices, as well as the challenges it poses for both experimental and theoretical attempts to obtain fundamental understanding of its underlying physics.

[0006] Currently, there are several experimental approaches to the study of domain wall dynamics. One of the principal techniques is time-resolved imaging via the magneto-optic Kerr effect (MOKE) (see, S. B. Choe, S. C. Shin, Phys. Rev. Lett. 86, 532 (2001); S. Lemerle et al., Phys. Rev. Lett. 80, 849 (1998); and D. A. Allwood et al., Science 296, 2003 (2002)), an optical method that can provide sufficient contrast to directly observe the growth or shrinkage of magnetic domains within a sample. Another recently developed approach is based upon electrical measurements via the giant magnetoresistance (GMR) effect, which has enabled investigations of domain wall propagation in submicron Permalloy wires (see, e.g., T. Ono et al., Science 284, 468 (1999)). In the latter work, the observed ˜1Ω GMR jumps emanate from regions that are smaller than is readily accessible to MOKE. However, this electrically-based approach is complicated by relatively small signal levels and the sample's multiple ferromagnetic layers.

[0007] At present, understanding of domain wall resistance is unsettled—there exists a conflicting body of work in the literature. Early theoretical studies (see, e.g., G. G. Cabrera, L. M. Falicov, Phys. Status Solidi B 61, 59 (1974); ibid. 62, 217 (1974) and L. Berger, J. Appl. Phys. 49, 2156 (1978)) predict that a domain wall should contribute minimally to the resistivity of a ferromagnet. This is predicated upon adiabatic electron transfer across a domain wall, based upon the reasonable assumption that the wall width is much greater than the Fermi wavelength. However, it was also pointed out that the different magnetization orientations of adjacent domains might locally perturb the current distribution and thereby yield an effect. More recent theoretical efforts strive to clarify this situation, but models yielding both positive and negative domain wall resistance have been explicated (see, P. M. Levy, S. Zhang, Phys. Rev. Lett. 79, 5110 (1997); G. Tatara, H. Fukuyama, Phys. Rev. Lett. 78, 3773 (1997); R. P. van Gorkom, A. Brataas, G. E. W. Bauer, Phys. Rev. Lett. 83, 4401 (1999); and M. Viret et al., Phys. Rev. B 53, 8464 (1996)). On the experimental side, measurements of domain wall resistance have been indirect; it is typically inferred by measuring the excess resistance arising when large ensembles of domain walls are nucleated within a sample (see, M. Viret, et al., Phys. Rev. Lett. 85, 3962 (2000); U. Ruediger, J. Yu, S. Zhang, A. D. Kent, S. S. P. Parkin, Phys. Rev. Lett. 80, 5639 (1998); and L. Klein et al., Phys. Rev. Lett. 84, 6090 (2000)). Recent progress in domain imaging and nanofabrication now allows resistance measurements on samples containing significantly reduced numbers of domain walls (see, U. Ebels, A. Radulescu, Y. Henry, L. Piraux, K. Ounadjela, Phys. Rev. Lett. 84, 983 (2000); T. Taniyama, I. Nakatani, T. Namikawa, Y. Yamazaki, Phys. Rev. Lett. 82, 2780 (1999); and Y. B. Xu et al., Phys. Rev. B 61, R14901 (2000)). However, to date, experimental results that have emerged from these various methods are also conflicting; both positive (see, M. Viret, et al., Phys. Rev. Lett. 85, 3962 (2000); L. Klein et al., Phys. Rev. Lett. 84, 6090 (2000); U. Ebels, A. Radulescu, Y. Henry, L. Piraux, K. Ounadjela, Phys. Rev. Lett. 84, 983 (2000); and Y. B. Xu et al., Phys. Rev. B 61, R14901 (2000)) and negative (see, U. Ruediger, J. Yu, S. Zhang, A. D. Kent, S. S. P. Parkin, Phys. Rev. Lett. 80, 5639 (1998) and T. Taniyama, I. Nakatani, T. Namikawa, Y. Yamazaki, Phys. Rev. Lett. 82, 2780 (1999)) domain wall resistance are reported.

[0008] In magnetic data storage technologies, and in particular magnetic hard disk drive (HDD) technologies, one goal is to increase storage capacity by, in part, reducing key dimensions within the HDD (“scaling”). Recently, megnetoresistive (MR) and giant magetoresistive (GMR) materials have been developed for use in read heads that help facilitate HDD scaling using MR and GMR materials with improved sensitivity to magnetic fields on a disk storage medium. Increased sensitivity makes it possible to detect smaller recorded bits and to read these bits at higher data rates. Larger signals from GMR heads also help reduce electronic noise. MR and GMR materials and devices however, are limited in sensitivity and therefore in application.

[0009] It is clear that there is a need for ferromagnetic semiconductor materials that provide enhanced magnetic sensitivity as well as ferromagnetic semiconductor-based sensing systems and methods that provide enhanced measurements using domain wall resistance and domain wall dynamics as well as magnetic switching effects.

BRIEF SUMMARY OF THE INVENTION

[0010] The present invention provides sensing systems and methods based on ferromagnetic semiconductor materials. The systems and methods of the present invention provide, inter alia, exceptional magnetic switching detection and analysis as well as enhanced domain wall magnetic property measurement and analysis relative to prior sensing systems. In one aspect, the present invention provides novel pressure sensors to detect pressure changes using magnetoresistive measurement techniques, and in particular GPHE techniques. In another aspect, the present invention provides magnetic switching detection sensors using ferromagnetic semiconductor-based materials that provide enhanced sensitivity relative to known materials and techniques. Such magnetic switching detection sensors according to the present invention are particularly useful as a read head sensor for HDD and other magnetic storage technologies.

[0011] According to an aspect of the present invention, a pressure sensing device is provided that typically includes a semiconductor housing structure having an opening defined therein, with the opening having a perimeter, and a thin semiconductor membrane covering the opening so as to define an enclosed cavity within the housing structure, with the membrane defining a pressure sensing region within the perimeter. The device also typically includes a ferromagnetic semiconductor Hall bar gage structure positioned proximal at least a portion of the perimeter of the pressure sensing region. In operation, the Hall bar gage structure produces a signal responsive to a deflection of the membrane in the pressure sensing region due to a pressure difference between the interior of the cavity and the exterior of the cavity, the signal being proportional to the pressure difference. In certain aspects, the substrate includes one of GaAs and GaN, and the membrane includes one of Mn doped GaAs and Mn doped GaN. In other aspects, the substrate and membrane include a type III-V semiconductor material, doped or undoped.

[0012] According to an aspect of the present invention, a method is provided for producing a ferromagnetic semiconductor-based pressure sensor. The method typically includes providing a substrate, forming an epitaxial heterostructure having two or more layers on the substrate, and forming a cavity in the substrate such that the cavity is exposed to a portion of a first one of the two or more layers of the epitaxial heterostructure, the exposed portion of the first layer defining a sensing region having a perimeter. The method also typically includes sealing the cavity, and patterning the layer adjacent the first layer so as to form a Hall bar gage structure proximal the perimeter of the sensing region and so as to expose the sensing region of the first layer to the atmosphere. The Hall bar gage structure, in certain aspects, produces a signal responsive to a deflection of the first layer in said pressure sensing region due to a pressure difference between the interior of the cavity and the exterior of the cavity, the signal being proportional to the pressure difference.

[0013] According to yet another aspect of the present invention, a ferromagnetic semiconductor-based read head sensor is provided. The sensor is typically configured to detect magnetic domain orientations in a magnetic recording medium having a plurality of domains, each domain having a magnetization. The sensor typically includes a substrate defining a plane, a ferromagnetic semiconductor epilayer formed on the substrate, with the epilayer having a cubic hard axis, and first and second read current contacts, each contact coupled proximal an end of the epilayer, with the contacts being configured to provide an electrical current flow along the hard axis. The sensor also typically includes one or more read probes, in electrical contact with the epilayer, configured to detect transverse magnetic resistance in the epilayer. In operation, application of an in-plane magnetic field, non-aligned with the cubic hard axis, produces a transition in the transverse magnetic resistance of the epilayer, and the magnetization of each domain produces a magnetic field having a component non-aligned with the cubic hard axis when the read head is positioned proximal thereto.

[0014] According to a further aspect of the present invention, a method is provided for detecting changes in magnetic domain orientations in a magnetic recording medium using a ferromagnetic semiconductor-based read head sensor. The method typically includes positioning a read head sensor proximal a magnetic recording medium having a plurality of domains, each domain having a magnetization, wherein the read head sensor includes a ferromagnetic semiconductor epilayer structure defining a plane and having a cubic hard axis. the method also typically includes moving the read head position relative to the domains in a sequential order, and detecting changes in the transverse magnetic resistance of the epilayer structure. Application of an in-plane magnetic field, non-aligned with the cubic hard axis, produces a transition in the transverse magnetic resistance of the epilayer, and the magnetization of each domain produces a magnetic field having a component non-aligned with the cubic hard axis when the read head is positioned proximal thereto.

[0015] Reference to the remaining portions of the specification, including the drawings and claims, will realize other features and advantages of the present invention. Further features and advantages of the present invention, as well as the structure and operation of various embodiments of the present invention, are described in detail below with respect to the accompanying drawings. In the drawings, like reference numbers indicate identical or functionally similar elements.

BRIEF DESCRIPTION OF THE DRAWINGS

[0016] FIG. 1(a) shows a Hall device, including a measurement setup, according to an embodiment of the present invention. FIG. 1(a) also shows an electron micrograph of a Hall device.

[0017] FIG. 1(B) shows a typical GPHE resistance vs. applied magnetic field (R-H) loop for the device of FIG. 1(a).

[0018] FIG. 2 shows the temporal evolution of domain wall propagation using the device of FIG. 1(a), measured via a series of GPHE measurements at 5° K, for applied in-plane magnetic fields stepped between 74 Oe and 88 Oe. Inset FIG. 2(a) shows a comparison of resistance switching curves for T=0.35K, 5K, and 20K. The horizontal axis value is normalized to time-of-flight between adjacent probes, whereas the vertical axis is normalized by half the asymptotic jump height. Inset FIG. 2(b) shows a magnified view that demonstrates the fast dynamics occurring at 88 Oe. FIG. 2(c) shows an estimated domain wall profile corresponding to inset FIG. 2(a).

[0019] FIG. 3 shows the measured dependence of domain wall velocity on applied in-plane field for temperatures of 5°, 10° and 20° K, displayed in linear and semi-logarithmic formats (inset). The external field is orientated 15° away from [110]. In the linear plot, solid lines represent fits to the high field velocities. In the semi logarithmic plot (inset), solid lines are linear fits to the low field region.

[0020] FIG. 4 shows time-resolved magnetoresistance measured across a single domain wall at 5K.

[0021] FIG. 4(a) shows GPHE signals from two pairs of transverse probes that monitor the arrival and departure of a single domain wall from the “measurement region” between probes.

[0022] FIG. 4(b) shows longitudinal resistances across the wall measured simultaneously from top and bottom of the device.

[0023] FIG. 4(c) shows the resistance differences satisfy a sum rule:

δR=RU−RD=RL−RR.

[0024] FIG. 4(d) shows the average, <R>=(RU+RD)/2, which effectively separates the magnetoresistance of an individual domain from obfuscating GPHE signals. The central portion of this trace, obtained when the domain wall is completely localized between the outermost transverse probes, displays a small, but distinct, offset from the interpolation between asymptotic resistance values.

[0025] FIG. 4(e) shows the temperature dependence of the averaged longitudinal resistance, scaled by the asymptotic resistance jump height, RAS. As shown, the resistance offset at the center of the traces, arising from the presence of a single domain between probes, does not change markedly with temperature. However, above ˜10K, increased “tails” (FIG. 4(a)) make it increasingly difficult to evaluate this offset.

[0026] FIGS. 5(a)-5(c) show families of data taken from Hall bars spanning in width from 6 μm to macroscopic (1 mm) dimensions. These are obtained for orientation φH=20° away from [110].

[0027] FIG. 5(d) shows the field-dependent sheet resistance of a 100 μm (wide) Hall bar.

[0028] FIG. 5(e) shows the relative orientations of the sensing current I, the external field H, and the magnetization M in a Hall bar device similar to the device shown in FIG. 1. A SEM micrograph of a 6-μm-wide device is also shown.

[0029] FIG. 5(f) shows the Barkhausen jumps that are evident in 6-μm-wide devices near the resistance transitions.

[0030] FIG. 6 shows the angular dependence of the GPHE.

[0031] FIG. 6(a) shows the dependence of R-H loops upon field orientation angle φH as it is varied from −30° to 30° in the plane of the device.

[0032] FIG. 6(b) summarizes the signatures of the coercive fields manifested in the electrical transport measurements.

[0033] FIG. 6(c) shows the planar Hall resistance measurements for a series of fixed-magnitude clockwise and counterclockwise sweeps of magnetic field orientation at different (fixed) magnitudes.

[0034] FIG. 7(a) shows the temperature dependence of the GPHE for a 10 μm Hall bar, measured under conditions of careful temperature regulation, stepped downward from 50 to 0.32 K, for fixed-orientation, swept-magnitude, applied magnetic fields.

[0035] FIG. 7(b) shows semilog plots of the planar Hall resistance jump, ΔRH, and sheet resistance, R, vs temperature.

[0036] FIG. 7(c) shows a comparison between the ratio, ΔRH/R□, measured on a 10-μm-wide Hall device, and the sample magnetization, M, measured by SQUID magnetometry on a macroscopic [(3×3) mm2] sample.

[0037] FIG. 8 shows an example of freezing of a single domain wall.

[0038] FIG. 9 shows a general starting heterostructure for use in forming devices according to the present invention.

[0039] FIG. 10 shows an example of a magnet system for generating an applied magnetic field in any direction (3-D) without disturbing the sample (internal).

[0040] FIG. 11 shows a representative graph of domain wall velocity vs. magnetic field orientation (v-H curves). Angles are measured with respect to the longitudinal axis of the device. Inset: Domain wall velocity vs. field component along [110]. In the inset, data points of different angles condense onto on a single universal curve in the low field regime.

[0041] FIG. 12 illustrates a micro-machined pressure sensor according to one embodiment of the present invention.

[0042] FIG. 13 illustrates a rectangular diaphragm type sensor according to an embodiment of the present invention. (Ga,Mn)As Hall bar placement in a rectangular diaphragm sensor. (a top view). To achieve higher signal to noise ratio and reduce drift, Hall voltages from a strained Hall cross as well as a unstrained Hall cross are measured simultaneously and differentially amplified.

[0043] FIGS. 14(a) and (b) show the resistance response to a sweeping in-plane magnetic field from both the sensor Hall crosses. (a,b) Sensor and reference planar Hall resistance response. (c) the output of differential amplifier. The width of the plateaus is proportional to the pressure applied in the linear response region.

[0044] FIGS. 15(a)-(d) illustrate a process of fabricating a sensor device according to an embodiment of the present invention.

[0045] FIG. 16 shows a GPHE read head sensor according to an embodiment of the present invention.

[0046] FIG. 17 shows a merged read/write head configuration including a GPHE sensor.

DETAILED DESCRIPTION OF THE INVENTION

[0047] FIG. 1(a) shows a Hall device 10, including a measurement setup, according to an embodiment of the present invention. FIG. 1(a) also shows an electron micrograph of a typical Hall device 10. As shown, device 10 is arranged similar to a standard Hall bar having electrical contacts 1 and 5 positioned at opposite ends of the longitudinal axis (as shown, the [110] direction). Three pairs of transverse (Hall) voltage probes are provided. For example, probes 2 and 8 make up a pair, probes 3 and 7 make up another pair, and probes 4 and 6 make up the third pair. Although three sets of probes are shown, it is appreciated that two sets may be used or that more than three sets may be used in certain aspects.

[0048] Device 10 is preferably formed using a ferromagnetic based semiconductor material. In one embodiment, epilayers of (Ga, Mn)As are formed on a GaAs substrate and patterned to form the device. In one aspect, for example, device 10 is fabricated by growing a Ga1−xMnxAs epilayer, e.g., 150 nm-thick Ga0.948Mn0.052As epilayer, on top of an insulating GaAs buffer layer (substrate) by low temperature molecular beam epitaxy (MBE). Such epilayers are subsequently patterned into one or more standard Hall bars, each with the longitudinal axis (i.e., the direction of current flow) preferably oriented along the [110] direction (a cubic hard axis), and each including a desired number of voltage probes. For example, for mass production of such devices, the epilayers may be formed on a wafer and then patterned into the desired number of devices, with the wafer subsequently diced to separate the devices. Details of both the fabrication methods and the GPHE-based analysis of the magnetic properties of such ferromagnetic semiconductor devices, and in particular (Ga,Mn)As epilayer devices, are described later. It is appreciated, however, that different materials and material concentrations and thicknesses may be used without departing from the scope of the invention.

[0049] Returning to FIG. 1(a), in one aspect, the width of the Hall bar is approximately 100 μm, and each pair of voltage probes are separated by about 100 μm along the longitudinal axis. It should be appreciated that other dimensions and spacings may be implemented, and that the current dimensions and sizes depicted represent one particular embodiment and are used to simplify the description and analysis that follows.

[0050] To avoid geometrical pinning of domain walls in their vicinity (see, J. Wunderlich et al., IEEE Trans. Magn. 37, 2104 (2001)), the probe widths are preferably minimized with respect to that of the channel (e.g., 100 μm). In one aspect, therefore, the probe widths are each approximately 3 μm. Dynamical measurements are carried out with a battery-supplied, constant dc drive current (e.g., 10 μA). Induced transverse voltages detected by the three voltage probe pairs are coupled (e.g., through triaxial cables) to three electrically-isolated, dc-coupled low noise differential amplifiers 9 providing, e.g., a 1 MHz signal bandwidth. A multichannel digital oscilloscope 11, or other signal processing and display system, enables simultaneous acquisition of voltage time records from the signal channels. During measurement, device temperature is preferably regulated, e.g., at values between about 0.3 K and about 45 K. After a domain wall is nucleated at one side of the bar, the wall propagates sequentially across the transverse probe sets, successively generating GPHE signal voltages. FIG. 1(b) shows a typical GPHE resistance (R) vs. magnetic field (H) loop for the device of FIG. 1(a). In one aspect, a 3-coil superconducting magnet system is used to generate the applied magnetic field (H), which can be rotated electronically in any direction under computer control. The magnet system, in one aspect, includes three orthogonal superconducting coils that allows for directing an applied magnetic field in any direction without physically disturbing the sample, e.g., device 10. FIG. 10 shows an example of such a magnet system. Such a 3-axis magnet device can be obtained from Oxford Instruments of Cambridge, UK. It will be appreciated that any system or mechanism capable of generating an applied magnetic field of desired magnitude(s) and direction(s) may be used.

[0051] It is known that (Ga,Mn)As films are intrinsically magnetized in-plane due to a combination of compressive lattice-mismatch-induced strain and demagnetization effects (see, e.g., H. Ohno, Science 281, 951 (1998)). A representative R-H loop (here R is the giant planar Hall resistance) as shown in FIG. 1(b) is obtained at T=5° K with an applied magnetic field (H) oriented 15° away from [110] and ramped at a rate of 15 Oe/s. The first resistance jump corresponds to a magnetization transition from [0{overscore (1)}0] to [100]; the second jump completes the reversal by switching from [100] to [010]. The square hysteresis loops obtained at low temperatures are indicative of magnetization switching dominated by wall motion, rather than domain nucleation. In (Ga,Mn)As, except for those directions precisely collinear with the easy axes, the magnetization transitions evolve via the formation of a 90°-domain wall. Here the focus is on time-domain studies of the first transition, which occurs sharply at a rather low switching field and involves initial and final states very close to those at equilibrium in zero field.

[0052] In certain aspects, a metastable magnetization state is induced within the sample, for example to facilitate domain wall propagation studies. This is achieved in certain aspects by first applying a strong in-plane magnetic field in a selected direction to saturate the magnetization, then smoothly ramping to a specific field magnitude with orientation antialigned to the initial saturation field. At the temperatures studied (e.g., between about 0.3° K and about 45° K), domain wall nucleation occurs infrequently through intrinsic stochastic processes. Once nucleated, the constant in-plane field drives growth of the domain possessing magnetization most closely aligned with the applied field. It is noted that domain wall motion induced in this manner always involves propagation from a wide current contact pad into the channel (see, e.g., K. Shigeto, T. Shinjo, T. Ono, Appl. Phys. Lett. 75, 2815 (1999)). With this protocol, completely reproducible signals are detected.

[0053] FIG. 2 shows the temporal evolution of domain wall propagation using the device of FIG. 1(a), measured via a series of GPHE measurements at 5° K, for in-plane magnetic fields stepped between 74 Oe and 88 Oe. The three simultaneously-obtained time records (of GPHE voltage) reveal passage of a single domain wall sequentially along each of the transverse probe sets. At a given field, the three temporal waveforms have substantially identical magnitudes and transition (i.e., rise) times, but occur sequentially. The time delays between their onsets at successive channels are substantially identical given the constant inter-probe spacing (e.g., 100 μm). These data are consistent with the picture that domain walls propagate along the device while retaining a fixed shape. Domain wall velocity can therefore be extracted by measuring the time-of-flight between transverse probe sets.

[0054] During each of the transitions, the transverse voltage varies almost linearly with time (FIG. 2(a)). This is evident at both limits, i.e., for very fast (88 Oe, FIG. 2(b)) and for very slow (74 Oe) domain wall motion. However, above T=4° K, extended tails are evident at the initial and final stages of the reversal. These tails become progressively suppressed as temperature is decreased, and almost entirely disappear at the lowest measurement temperature (˜0.35K). There, the ratio of the transition time and the propagation time between neighboring transverse probes (the delay between their respective signal onsets) is about 1.4±0.2. This indicates that the domain wall extends longitudinally 140 μm±20 μm along the Hall bar. The persistence of a nearly linear temporal evolution over the range 4K<T<25K suggests domain wall shape does not change markedly at elevated temperatures. FIG. 2(c) shows an estimated domain wall profile corresponding to inset FIG. 2(a).

[0055] Below 25° K, the data indicate that magnetization reversal proceeds through infrequent nucleation of single domain walls within one of the (large) current contacts, followed by uniform propagation through the device channel. Above 25° K, however, multiple local domain wall nucleation appears to dominate the magnetization reversal process, and the time of flight method is not applicable. (Data not shown) FIG. 3 shows the measured dependence of domain wall velocity on applied in-plane field for temperatures of 5°, 10° and 20° K, displayed in linear and semi-logarithmic formats (inset); over this range the velocity spans four decades. Measurements were taken with the external applied field (H) oriented 15° away from [110]. In the linear plot, solid lines represent fits to the high field velocities. In the semi logarithmic plot, solid lines are linear fits to the low field region.

[0056] The v-H curves exhibit highly nonlinear dependence upon in-plane magnetic field. This behavior can be divided into two distinct regimes. For driving fields H larger than a specific threshold H0 (essentially the intrinsic coercive field), the influence of disorder-induced domain wall pinning can be ignored. In this case, a linear v˜H relation is obtained that is characteristic of viscous flow. In this regime the wall velocity can be expressed as:

v =μ(H−H0)  (1)

[0057] where μ is the domain wall mobility (see, e.g., T. H. O'Dell, Ferromagnetodynamics. the dynamics of magnetic bubbles, domains, and domain walls (John Wiley & Sons, New York, 1981)). This mobility varies strongly with temperature; the data shows the mobility increasing from 1.4 mm/s·Oe at 5° K to about 14 mm/s·Oe at 20° K. These results are summarized in Table I. below.

TABLE I
Characteristic parameters determined for (Ga, Mn)
As in two field regimes.
Temperature (K)	Mobility (mm/s · Oe)	H0(Oe)	α(Oe−1)	1$\begin{array}{c}{V}_N^{1/3}\\ \left(\mathrm{nm}\right)\end{array}\&AutoRightMatch;$
5	1.39	79.8	0.185	18.2
10	6.11	52.1	0.260	26.5
20	14.2	24.8	0.342	37.4

[0058] The temperature dependence of the domain wall velocity is of importance because it can yield information about the source of spin damping in the magnetic system. On reducing the magnetic field well below H0, domain wall velocity in the system decays dramatically. The data in this regime are best fit by a thermally activated depinning model (see, e.g., M. Laurune, S. Andrieu, F. Rio, P. Bernstein, J. Magn. Magn. Mater. 80, 211 (1989)),

v∝exp [(H·ΔMVN−Ep)/kBT]∝exp[α(H cos φH−H0)]  (2)

[0059] Here Ep is the activation energy for domain wall propagation, VN is the activation volume, ΔM=M2−M1 (with M1 representing the initial magnetization (close to [0{overscore (1)}0]) and M2 the magnetization after switching (close to [100]) ), φH is the magnetic field orientation and α=MSVN/kBT (here MS represents the saturation magnetization of (Ga,Mn)As). From the low-field slopes of the semi-logarithmic curves (FIG. 3, inset) the activation volumes for different temperatures (displayed as a linear dimension, 2$V_N^{1/3},$

[0060] in Table I.) are estimated.

[0061] Investigating the v-H curves along various sweep angles further confirms the validity of Eq. 2. The dependence of domain wall velocity on applied field, φH, swept along 5°, 15°, 22° and 30° orientations has been measured. FIG. 11 shows a representative graph of domain wall velocity vs. magnetic field orientation (v-H curves). The resulting data coalesce onto a single universal curve (FIG. 11 inset) in the low field region upon scaling the magnetic field by cos φH. This demonstrates that it is solely the field component along [110] that is effective for driving domain wall motion.

[0062] It is useful to compare the present approach to other methods for investigating domain wall dynamics. Existing GMR-based techniques, for example, involve resistance measurements which reflect global changes in the growth (i.e. volume expansion) of domains, integrated over the entire sample (see, e.g., T. Ono et al., Science 284, 468 (1999)). By such an average approach it is not possible to track the spatial propagation of an individual domain wall “wavefront”, as is demonstrated herein. By contrast, use of separate transverse probes advantageously allows for the local evolution of the domain wall to be followed in real time, with high spatial resolution. Furthermore, extremely high sensitivity is achieved through the (four-probe) circumvention of what would otherwise be a large (two-probe) background resistance.

[0063] According to one aspect, individual domain walls are manipulated, directly and reproducibly, by controlling the external field. For example, once a domain wall is detected between probes, it can be “frozen” inside the channel by quickly removing the external field. The domain wall then decelerates virtually instantaneously, and remains stationary thereafter for as long as needed. Domain wall motion can subsequently be resumed by reapplication of an in-plane field. This ability to entrap a single domain wall between probes enables the directly characterization of the electrical resistivity of an individual entity. FIG. 8 shows an example of freezing of a single domain wall. The domain wall is driven at a slow speed and the planar Hall resistance of the central probes is monitored to reflect the passage of the domain wall. The crossing-zero of this signal triggers the removal of external magnetic field. The domain wall is found to remain stationary. In one aspect, this is achieved quasistatically. First, the in-plane driving field is adjusted so that, once nucleated, a solitary domain wall propagates very slowly across the sample (˜100s delay between successive probes). Then, during its traversal high resolution measurements of two distinct longitudinal resistances, termed RU and RD, are obtained via a.c. lock-in techniques (e.g., 79 Hz drive, 30 ms integration time). These involve either a pair of voltage probes on the top side of the device, RU=R15,24 or a pair on the bottom, RD=R15,86. (Here, conventional four-probe notation is used; Rij,kl corresponds to driving current from terminal i to j, and measuring the potential induced from k to l.) Simultaneously, the Hall resistances obtained from the left and right transverse probes (RL=R15,28, RR=R15,46) are measured to monitor the arrival and departure of the domain wall from the “measurement region” between these probes. FIG. 4(a) shows GPHE signals from two pairs of transverse probes that monitor the arrival and departure of a single domain wall from the “measurement region” between the probes. In certain aspects, the concomitant time records of these four resistances (RU, RD, RL and RR) are obtained by digitally sampling the averaged analog output from four separate lock-in amplifiers, collectively synchronized to a common reference. In one aspect, bridge circuits are utilized to null the large static background in the two longitudinal resistance measurements.

[0064] FIG. 4(b) displays the change in longitudinal resistance resulting from entrapment of a single domain wall in the measurement region at 5° K. A perturbation as large as 0.6% (˜30Ω) is manifested. However this should not be literally interpreted as the intrinsic contribution from an individual domain wall. Because of the presence of the domain wall, both RU and RD comprise an admixture of longitudinal and Hall resistance. The latter, i.e. the GPHE signal, is predominant, and contributes a term with sign that is dependent on the direction of wall propagation. Precise extraction of the domain wall resistance from the overall magnetoresistance requires self-consistent simulation of the local current distribution. This is problematic; the exact shape of domain wall and the potential profile around it, are unknown. Instead, in certain aspects, an operational approach is taken, such as for example, averaging the measured longitudinal resistances, <R>=(RU+RD)/2, in effect, to compensate the GPHE contribution to first order. FIG. 4(d) shows the average, <R>=(RU+RD)/2. These compensated data indicate that the domain wall induces non-uniform eddy-like currents upon entering or leaving the measurement region, resulting in an observable perturbation, of order of 0.2% of the resistance background. These eddy-like currents are local, static variations in current density that arise from the resistivity discontinuity across the domain wall (see, L. Berger, J. Appl. Phys. 49, 2156 (1978)); they do not originate from the dynamics of domain propagation. However, when the domain wall is completely localized between these probes, <R> clearly settles to reflect what appears to be the true contribution to the resistivity from a single domain wall. This is manifested in the central, linear feature within the trace of FIG. 4(d), which has two salient elements. The first, a linear evolution of the sheet resistance in this region, is interpreted as reflecting the fractional contributions from the two domains present—these change smoothly as the domain wall traverses the measurement region. (These two domains possess different resistivity, represented in the asymptotic values attained by <R>). The second, a small negative offset from the interpolated value in this locale, roughly −0.44Ω, is significant and appears to directly reflect the contribution from the single domain wall localized between the probes. Although completely reproducible, and well within measurement resolution (˜0.1Ω), this offset is comparable to the accuracy of the background interpolation, which is estimated to be about 1 part in 10,000 of <R>, or ˜0.5Ω. These conclusions are supported by data at other temperatures as shown in FIG. 4(e).

[0065] Irrespective of its precise value, the resistance of an individual domain wall in (Ga,Mn)As epilayers is exceedingly small. If it is finite, its value appears to be negative and less than 0.01% of the sheet resistance of the (Ga,Mn)As epilayers. Measurements of single domain wall propagation and magnetoresistance within patterned (Ga,Mn)As conductors possessing in-plane magnetization have been performed. The techniques of the present invention are applicable at temperatures down to the millikelvin range, where thermal fluctuations cannot overcome domain wall pinning and macroscopic quantum tunneling (MQT) of spatially-extended domain walls should be manifested.

[0066] Fabrication and Analysis

[0067] Semiconductor ferromagnetism also gives rise to new physical phenomena because it is possible to engineer, and enhance, spin-orbit coupling in ways that are not possible in metallic systems. According to the present invention, a giant planar “Hall” effect (GPHE) in ferromagnetic semiconductor devices, e.g., epitaxial (Ga,Mn)As thin film devices, is one such engineered effect. The GPHE is manifested as a spontaneous transverse voltage that develops, because of spin-orbit coupling, in response to longitudinal current flow in the absence of an applied field. Analogous effects studied in metallic ferromagnets are exceedingly small—typically of order mΩ (see, e.g., K. Hong and N. Giordano, Phys. Rev. B 51, 9855 (1995).). Related phenomena have recently been investigated in ferromagnetic semiconductors (see, e.g., H. Ohno et al., Science 281, 951 (1998).), but herein the previously unrecognized, and quite remarkable, response of the GPHE to an applied in-plane magnetic field is disclosed. Within the high quality, single domain ferromagnetic semiconductors of the present invention, a switchable effect is produced that is about 4 orders of magnitude stronger than found in metallic ferromagnets. Below, measurements that take advantage of this strong GPHE provide insight, and unprecedented resolution, into the mechanism of magnetic switching within these materials. These data, in turn, enable complete characterization of the magnetic anisotropy of the ferromagnetic semiconductor structure, e.g., (Ga,Mn)As epilayers. Without being limited to one particular theory, it is believed that this phenomenon stems from the combined effects of significant spin-orbit coupling in the valence band of the zinc blende crystal structure, and the large spin polarization of holes in (Ga,Mn)As. The temperature dependence of the magnetization and the coercivity determined by electrical measurement should provide additional insight into the underlying physical mechanisms.

[0068] In one aspect, GaMnAs epilayers are formed on a GaAs substrate. The epilayers are thereafter, or concurrently, patterned using conventional techniques to form the device of interest, e.g., Hall bar device 10 of FIG. 1. Voltage probes are also formed as desired using conventional fabrication techniques. In one preferred aspect, molecular beam epitaxy at 250° C. is used to deposit 150-nm-thick Ga0:948Mn0:052As epilayers on top of an insulating GaAs(001) substrate with a buffer layer. It should be appreciated that other formation techniques may be used, such as for example, chemical vapor deposition (CVD) techniques. Various thicknesses and concentrations of materials have been investigated. In certain aspects, samples are annealed at elevated temperatures. The devices described herein are patterned from a single wafer (UCSB-001115A, Curie temperature Tc˜45 K); however, it is notable that all devices studied (fabricated from a variety of different epilayers) exhibit consistent behavior. It should also be appreciated that other concentrations and thicknesses of materials may be used, and that other materials may be used. For example, in certain preferred aspects, GaMnAs epilayers having a Mn concentration of between about 1% to 8%, and more preferably between 1% and 7%, and having a preferred thickness of between about 10 nm and about 350 nm, and more preferably between about 100 nm and about 350 nm are used. Additionally, any III-V type semiconductor materials and others may be used, such as, for example, Mn doped GaN (e.g., Ga1−xMnxN) epilayers on a GaN substrate.

[0069] In one aspect, a general starting heterostructure is shown in FIG. 9. In certain aspects, a ferromagnetic epilayer 30 is formed on a substrate layer 40 using conventional techniques such as low temperature molecular beam epitaxy (MBE), chemical vapor deposition (CVD), etc. In one aspect, when using (Ga,Mn)As, because (Ga,Mn)As is heavily p-doped, a Be doped GaAs layer (e.g., buffer layer) is used as p-type conducting channel. In preferred aspects, substrate 40 is doped to form the buffer layer, however, the buffer layer may be formed using alternate techniques such as by deposition or epitaxy. Examples of useful Be doping levels include between 1×1019 cm−2 and 1×10−2 cm−2. The contacts between (Ga,Mn)As and GaAs:Be are ohmic for this range of concentrations. Compared to n-type semiconductors (i.e., those with electronic conduction), ohmic contact to p+GaAs:Be is easy to achieve and alloying is usually not needed. Alternatively, a separate buffer layer may be formed on the substrate 40, using conventional techniques appropriate for the desired buffer layer material(s).

[0070] After cleaving and etching the heterostructure using techniques as are well known, Ti/Pd/Au/Ti metal layers are deposited, e.g., via electron beam lithography and lift-off techniques, to produce voltage probes and to define one or more micro-scale bars (e.g., device/bar 10) at various separations. For example, a single continuous deposition process may be used to deposit TI-Pd-Au-Ti layers in succession. The remaining description will reference two or more micromagnet structures. The widths and aspect ratio of these bars may be made different to yield micromagnets with slightly different coercivities after subsequent ion-mill etching (e.g., using metallic layers as self aligned mask). Alignment marks are also put down to assist the fabrication of the remaining layers. In certain aspects, an etch is performed. The etch is calibrated to completely remove all the top (Ga,Mn)As epilayers and additional 50 nm-100 nm into the p-type paramagnetic GaAs layers. Therefore, except for those isolated areas under metal masking layers, no continuous ferromagnetic epilayers exist on the chip, and the p-type conducting GaAs layers are exposed. Using another electron beam lithography and lift-off step, a thick Ti layer is defined and employed as an etching mask to form—p−GaAs conduction channel that connects both micromagnets. Conventional photolithography processes allow for the deposition of same thickness Ti masking layer for large contacts and for leads connecting to the p−GaAs channel in the center. Mesa isolation may also be achieved by ion-beam etching. Because of its excellent repeatability, ion-beam etching is preferred for (Ga,Mn)As device fabrication. Note that for this material, ion beam induced damage will be of less concern. The Ti masking layer is then removed, e.g., using diluted HF. The bottom thin titanium layer adhering to magnets is basically untouched since it reacts with Pd and tends to heavily dope the GaAs in that vicinity. In one aspect, a final composite Pd/Ti/Pd/Au layer may be patterned to make contacts to both p+GaAs and micromagnets (with Ti/Pd/Au on top), e.g., through combined electron beam lithography (EBL) and photolithography (PL). In this aspect, the whole process involves three EBL and one PL steps, four thin film depositions steps, and two ion-beam etching processes.

[0071] In certain aspects, magnet widths of between about 1,000 μm (or greater) and about 0.1 μm (or smaller), preferably between about 100 μm and about 0.5 μm, and more preferably between about 0.75 μm and about 0.5 μm are made. The resistance between the pads of the ferromagnet and p−GaAs pads are in the range of 10 kΩ to 50 kΩ, in a few cases above 100 kΩ. Most of the resistance is present at the metal/(Ga,Mn)As interface, whereas the resistance across (Ga,Mn)As/GaAs interface is very small. Further optimization of the fabrication may be required to obtain consistent small contact resistances between metal and p+GaAs.

[0072] Magnetoresistance measurements have been carried out on families of Hall bars (e.g., widths ranging from 6 μm to 1 mm) and square van der Pauw devices (e.g., (3×3) mm2 devices). The former are aligned along the [110] direction by a combination of photo- and electron-beam lithography. Voltage probes on the Hall bars are carefully designed to minimize their perturbation upon current flow within the devices (see, e.g., scanning electron micrograph (SEM), FIG. 5(e)). Standard four-probe lock-in measurements are performed by a 10 nA ac sensing current at 14 Hz; excitation is intentionally kept quite low to obviate electron heating. Magnetic fields are generated using a three-coil superconducting magnet that allows 3D field orientation without physically disturbing the sample. Two classes of experiments were conducted. In the first, the applied magnetic field orientation is fixed in-plane along a specific direction, φH, with respect to the longitudinal axis of the Hall bars, while the field magnitude is swept linearly between ±1000 Oe. In the second, the magnitude of the applied in-plane field is fixed, while stepping its orientation clockwise or counterclockwise. Prior to each sweep, an in-plane field of approximately 6000 Oe is applied to saturate the sample magnetization, M.

[0073] At all angles, except for those along [110] directions, two abrupt jumps are observed in GPHE measurements. Data taken for various Hall bar devices spanning from 6 μm to macroscopic (1 mm) dimensions (width) are shown in FIGS. 5(a)-5(c). These are obtained for orientation φH=20° away from [110]. For comparison, the field-dependent sheet resistance of a 100 μm Hall bar is also displayed in FIG. 5(d).

[0074] Four distinct features emerge. First, large switching events at distinct magnetic fields are observed in the Hall resistance; these are accompanied by small jumps (relative to the background) in the longitudinal resistance. Second, between these switching fields, the planar Hall resistance remains substantially constant at approximately 37Ω. The signal polarity reverses at each switching event. Third, the switching fields are substantially independent of sample size and geometry. Measurements on samples with square, van der Pauw geometry, as large as (3×3) mm2, exhibit identical switching behavior as those of the smaller, micron-scale devices—even though the magnitude of the Hall resistance is reduced in the former, presumably due to nonuniform current distribution. Fourth, when the width of the Hall device is reduced to about 6 μm, small Barkhausen jumps are observed. These occur in close proximity to the switching transitions as shown in FIG. 5(f), and demonstrate that the propagation of domain walls is constrained by geometry (see, e.g., J. Wunderlich et al., IEEE Trans. Magn. 37, 2104 (2001); T. Ono et al., Science 284, 468 (1999).).

[0075] Hall bars fabricated along many other directions besides the [110] crystalline axis were also investigated. It was found that that the switching fields do not depend on the orientation of the Hall bars, whereas the magnitude of planar Hall resistance jumps is systematically reduced as the orientation moves away from the (110) directions.

[0076] FIG. 6(a) shows the dependence of R-H loops upon field orientation angle φH as it is varied from −30° to 30° in the plane of the device. In the field range of these experiments, only one jump occurs along the (110) directions. Away from these special orientations, a two-jump reversal is always observed. The first switching field Hc1 is almost constant, while the second switching field Hc2 decreases dramatically and approaches Hc1 at around ±30°.

[0077] Without being limited to one particular theory, the jumps in the Hall resistance are described as follows. The electric field within a single domain ferromagnetic film with in-plane magnetization can be written as:

E x =jρ⊥+j (ρ||−ρ⊥)cos2 φ  (1)

E y =j (ρ||−ρ⊥)sin φ cos 100 .  (2)

[0078] where the current density j is assumed to be uniformly distributed along the Hall bar, x and y are the longitudinal and transverse axes, and φ is the angle between the magnetization and current density j. ρ|| and ρ⊥ are the resistivities for current oriented parallel and perpendicular to the magnetization. (see, J. P. Pan, Solid State Physics, edited by F. Seitz and D. Turnbull (Academic, NewYork, 1957),Vol. 5, p. 1-96.) The anisotropic magnetoresistance phenomenon is described by Eq. (1). The transverse resistance, i.e., the planar Hall resistance, is expressed in Eq. (2), which exhibits extrema at φ=45° and its cubic equivalents. To verify this angular dependence of the planar Hall resistance, an in-plane field of constant magnitude 6000 Oe is applied to saturate the magnetization, and its orientation is swept through 360°. FIG. 6(c) shows the planar Hall resistance measurements for a series of fixed-magnitude sweeps of magnetic field orientation at different (fixed) magnitudes. In accordance with Eq. (2), the measured Hall resistance exhibits extrema for applied field orientations of about 45°, about 135°, about 225°, and about 315°. Note that the first maximum of planar Hall resistance appears at 135° instead of 45°, indicating ρ||−ρ⊥<0[−73Ω from Eq. (2)]. This property of (Ga,Mn)As is distinct from that in conventional ferromagnetic metals, where ρ||−ρ⊥>0. It may originate from the different manner in which holes and electrons contribute to the spin-orbit interaction in ferromagnetic materials.

[0079] The anomalous switching behavior of the Hall resistance shown in FIGS. 5(a)-5(c) can be explained by a two-jump sequence of magnetization: [100] (φ˜−45°)→[010] (φ˜45°)→[100] (φ˜135°). This evolution also accounts for the accompanying small longitudinal resistance jumps shown in FIG. 5(d). Between the switching events, the sample remains in what appears to be a macroscopic single domain state. In this regime the magnetization evidently rotates coherently according to the Stoner-Wohlfarth model (see, E. C. Stoner and E. P. Wohlfarth, Philos. Trans. R. Soc. London A 240, 74 (1948).); hence, the planar Hall resistance continues to evolve to a small degree with the field. Scanning SQUID microscopy of (Ga,Mn)As epilayers magnetized in-plane have provided evidence for the existence of macroscopic single domains on length scales extending to hundreds of microns (see, e.g., T. Fukumura et al., Physica (Amsterdam) 10E, 135 (2001).). Domain states within the sample exist only in the vicinity of the switching field, and the associated domain wall scattering evidently generates the remarkable resistance spikes shown in FIG. 5(d).

[0080] FIG. 6(b) summarizes the signatures of the coercive fields manifested in the electrical transport measurements. The field loci delineating the resistance transitions are shown in polar coordinates. The Hc1, lines form a rectangular shape, whereas the Hc2 lines are more complicated. The latter follow the extrapolation of Hc1 lines at low field but evolve towards the (110) axes in higher fields. Eventually, at a field around 2500 Oe, the second jump becomes smeared and reversible. These measurements clearly elucidate behavior that is generic in these (Ga,Mn)As epilayers: the in-plane magnetic anisotropy is nearly cubic, but it is biased by a small twofold preference along [110].

[0081] Unusual multiple switching, somewhat analogous to that demonstrated herein, has also been observed in ultrathin epitaxial Fe films, through the magneto-optic Kerr effect. A switching pattern analogous to that of FIG. 6(b) was measured in a Ag/Fe/Ag(001) system by Cowburn et al. (see, R. P. Cowburn et al., J. Appl. Phys. 78, 7210 (1995).), although with significantly less resolution in their metallic system. To explain their results, a simple model is invoked incorporating a well-defined domain wall pinning energy into a complex, anisotropic magnetocrystalline energy surface. (A weak in-plane uniaxial anisotropy is superimposed along one easy axis of a strong cubic anisotropy.) The present experimental data can be explained via similar domain reversal energetics, but in the present case the in-plane uniaxial easy axis is collinear with a hard axis of the cubic anisotropy. The corresponding free energy density of such a single domain magnet can be written as E=Ku sin2 φ+(K14)cos2 2φ−MH cos(φ−φH) Here, Ku and K1 are in-plane uniaxial and cubic anisotropy constants. The equilibrium state is defined by the conditions, ∂E/∂φ=0 and ∂2E/∂2>0. The former gives

K u sin 2φ−K1 sin 4φ+MH sin(φ−φH)=0  (3)

[0082] At zero field, four distinct magnetization states, corresponding to four local energy minima, can exist: 3${\phi }_{1,2}^0=\pm \left(\pi /4=\delta \right),{\phi }_{3,4}^0=\pm \left(3\pi /4+\delta \right),$

[0083] with δ=sin−1(Ku/K1). Domains can exist over short length scales in a demagnetized thin film. Upon application of an in-plane field, these small-scale domains quickly become suppressed, and the whole sample evolves into a macroscopic single domain state with φ close to 1 of the zero fields minima 4${\phi }_{1-4}^0.$

[0084] When the external field is reversed, magnetization reversal is achieved via an intermediate state corresponding to the sample magnetization oriented almost orthogonally (90°) to the initial and final directions of the magnetization. Domain states mediate the transitions from one energy minimum to another. For a domain wall to become liberated to propagate through the sample, the reversed external field must be increased to the point where a characteristic pinning energy density, ε, is exceeded, i.e., Hc−(M2−M1)=ε. Here M1;M2 are the initial and final magnetization, and Hc is the switching field. If Hc is small compared to the cubic anisotropy field, coherent rotation of M1 and M2 from the zero-field equilibrium is negligible. For transitions from [100] to [010], a ˜90° domain wall with core magnetization along [110] is required to propagate across the sample, giving Hc·{circumflex over (x)}=−ε110/2M sin(45°−δ), in which ε110 is the corresponding domain wall pinning energy density. Considering all possible orientational trajectories, the loci of transitions can be described as Hc·{circumflex over (x)}=±ε110/2M sin(45°−δ) and H·ŷ=±ε110/2M× sin(45°−δ). At low field, these describe two parallel sets of lines that are in excellent correspondence with the switching pattern observed in the experiments as shown in FIG. 6(b). At high fields, two new pieces of physics become important. First, coherent rotation of M must be considered and, second, the energy density of a domain wall also becomes significantly reduced (see, A. Hubert and R. Scha{umlaut over ()}fer, Magnetic Domains: The Analysis of Magnetic Microstructures (Springer-Verlag, Berlin, 1998), p. 235.). As a result, high field transitions progressively evolve towards the (110) directions.

[0085] Several additional points should be noted. First, PHE measurements enable determination of crystallographic orientation with remarkable precision; the angular error in establishing the (110) crystalline axes is estimated to be less than 0:04°. Second, apart from the singularity along these (110) directions, neither single transitions nor three-transition processes are observed. This justifies an assumption that in-plane uniaxial anisotropy does not exist along the cubic easy axes (see, R. P. Cowburn, et al., Phys. Rev. Lett. 79, 4018 (1997).).

[0086] Both the cubic and uniaxial anisotropy fields can be deduced through PHE measurements. In one aspect, to achieve this, a large, constant magnetic field is applied in the plane while its orientation is rotated continuously. FIG. 6(c) shows data from such measurements for clockwise and counterclockwise sweeps of magnetic field orientation, for several different field magnitudes. When H<HcA, where HcA˜2500 Oe is the dominant cubic anisotropy field, the magnetoresistance reverses each time the magnetization switches across the cubic hard axis. The planar Hall resistance becomes reversible for fields greater than 2500 Oe, in which case the magnetization rotates coherently according to the Stoner-Wohlfarth model. Thus, for a given external field angle φH, the macroscopic in-plane magnetization orientation, φ, can be calculated by using the expression in Eq. (2). Fitting all the computed data sets (φH; φ) to Eq. (3), the anisotropy fields HcA=2K1/M=2400 Oe, HuA=2Ku/M=160 Oe are consistently and unambiguously extracted.

[0087] Progress has recently been made toward gaining a theoretical understanding of magnetic anisotropy in III-V magnetic semiconductors (see, e.g., T. Dietl et al., Phys. Rev. B 63, 195205 (2001); T. Dietl et al., Science 287, 1019 (2000).; M. Abolfath et al., Phys. Rev. B 63, 054418 (2001); J. Ko{umlaut over ()}nig, T. Jungwirth, and A. H. MacDonald, Phys. Rev. B 64, 184423 (2001).). It is generally agreed that, in addition to an intrinsic cubic anisotropy, (Ga,Mn)As possesses a substantial out-of-plane uniaxial component with sign that is dependent on whether tensile or compressive biaxial strain exists at the interface (see, H. Ohno et al., Science 281, 951 (1998).). While attention has focused almost exclusively on the out-of-plane magnetic anisotropy, recent work on (Ga,Mn)As magnetic tunnel junctions (see, M. Tanaka and Y. Higo, Phys. Rev. Lett. 87, 026602 (2001)) highlights the importance of the in-plane anisotropy. On the other hand, theoretical models of cubic anisotropy predict that the in-plane cubic easy axes can be oriented along either (100) or (110) depending upon the hole concentration and the degree of spin splitting. By contrast, all of the data taken on a variety of (Ga,Mn)As epilayers as described herein indicate that the cubic easy axes are aligned along (100) axes. Further PHE studies on additional epilayers may be needed to determine if the cubic anisotropy exhibits the variations predicted by theory. The experimental results do agree, however, with the predicted magnitude of the cubic anisotropy field, which is found to be about 2400 Oe.

[0088] The temperature dependence of PHE has also been studied. The temperature dependence should be of significant importance in elucidating the underlying physical mechanisms. FIG. 7(a) shows the results for a 10 μm Hall bar device, measured under conditions of careful temperature regulation, stepped downward from 50 to 0.32 K, for fixed-orientation, swept-magnitude, applied magnetic fields. The magnitude of both PHE and the coercive fields increases rapidly with decreasing temperature as shown in FIG. 7(b). For T<10 K, both the PHE and sheet resistivity diverge logarithmically down to the lowest measured temperatures, while the ratio ΔRH/R□ remains nearly constant. Here, ΔRH is the PHE resistance jump and R□ is the zero-field sheet resistance. This ratio, ΔRH/R□, provides valuable information about the hole spin polarization, which has been found to decrease monotonically with increasing temperature, qualitatively tracking the magnetization of a (3×3) mm2 sample measured by SQUID magnetometry as shown in FIG. 7(c).

[0089] The giant planar Hall effect in ferromagnetic semiconductor devices, and in particular in (Ga,Mn)As devices, as described herein enables systematic investigation of in-plane magnetic anisotropy and magnetization reversal via electrical transport measurements. In semiconducting materials, GPHE measurements provide unique advantages over magneto-optical techniques. For example, carrier concentration changes arising from sample illumination can be circumvented. It is also notable that the high signal-to-noise attainable in GPHE measurements permits observation of behavior that emerges only in structures of reduced dimensions (e.g., Barkhausen jumps in 6 μm devices). Given the minimal excitation power required, the present invention is compatible with very low temperature measurements (˜mK), thus offering new possibilities for investigations in micro- and nanoscale spintronic devices.

[0090] Sensing Systems and Methods

[0091] In one aspect, the Giant Planar Hall Effect (GPHE) in diluted magnetic semiconductors (DMS), combined with its super magnetostrictive effect, offers, inter alia, high-resolution pressure sensing and enhanced magnetic domain switching detection when this novel class of materials and devices are incorporated into a sensor system.

[0092] In certain aspects, as discussed above the Giant Planar Hall effect reflects the large Hall voltage change due to the instantaneous switching of magnetization when a magnetic field is swept in the film plane of a high-quality epitaxial magnetic semiconductor. The Hall resistance jump is extremely sharp with a magnitude on the order of about 1000 times larger than observed in normal ferro-magnetic metals such as Fe and Permalloys.

[0093] As discussed above, the GPHE in a DMS such as (Ga,Mn)As, or other III-V magnetic semiconductors, originates from the strong spin-orbit inter-action in the valence band of Zinc-blend structure. The same coupling also provides a fairly large magnetoelastic coupling constant in (Ga,Mn)As and hence yields a high magnetostrictive effect (or its inverse effect). The correlation between mechanical stress and magnetic anisotropy in thin film (Ga,Mn)As can be preliminarily estimated by the measured lattice-mismatch induced out-of-plane anisotropy field as follows:

H u =1072 Oe for ε=−0.15%.  (6)

[0094] Compared to normal ferro-magnetic materials such as Fe, to generate the same amount of coercivity, a strain variation of about 8% is required. (See, e.g., D. Sander, R. Skomski, C. Schmidthals, A. Enders, J. Kirschner, Phys. Rev. Lett. 77, 2566 (1996).) The magneto-elastic coupling is more than about 50 times larger for diluted semiconductor (Ga,Mn)As. It should be appreciated that other III-V semiconductor materials as described herein may be used, with similar calculations as above, e.g., eq. 6.

[0095] Pressure Sensor Design

[0096] FIG. 12 illustrates a micro-machined pressure sensor 100 according to one embodiment of the present invention. A pressure difference across a diaphragm 110 causes the diaphragm to deflect and become mechanically strained. In one aspect, the strain is detected by measuring the change of amplitude or switching field of a Giant planar Hall signal from a Hall bar, e.g., (Ga,Mn)As Hall bar, gage 120 patterned proximal the edge of the membrane 110. FIG. 12 shows a circular diaphragm, and cavity, however, other geometries and configurations may be used, such as, for example, rectangular, oval, etc. FIG. 15 and the associated description below discusses aspects of fabricating a pressure sensor device 100 and others.

[0097] The strain distribution of a uniformly loaded diaphragm has been previously calculated in standard text books. (See, e.g., Timoshenko and Woinowski-Krieger, Theory of Plates and Shells 2nd Ed.) For small deflection, the strain introduced at the edge of a circular diaphragm is given by: 5$\begin{array}{cc}\epsilon =\frac{\Delta P}{E}\frac{3a^2}{4t^2}.& \left(7\right)\end{array}$

[0098] Here, ΔP is the pressure difference across the diaphragm, E is Young's modulus of the diaphragm, t is the thickness and a is the radius. For a rectangular diaphragm, a numerical factor of order 1 should be expected. For larger deflections, analytical solutions may be hard to determine and finite element simulations may be necessary to calculate the exact strain distribution.

[0099] Combining equations (6,7) the gauge factor represented by coercive field change in response to applied pressure can be deduced:

ΔH/ΔP=7(a/t)2 Oe/MPa  (8)

[0100] For example, for a device with a=60 μm, t=150 nm, this gauge factor is approximately 1.2 Oe/Pa. (A Young's modulus for GaAs of approximately 75 GPa is used.) If the radius of the membrane is increased, for example to 300 μm, the sensitivity can be as high as about 30 Oe/Pa (at 4K). Compared to conventional pressure sensors, sensors according to the present invention provide an extremely sensitive solid state pressure gauge. In GPHE measurement, the resolution in field can be as high as about 0.2 Oe. Therefore, the minimum detectable pressure change is about 7 mPa.

[0101] FIG. 13 illustrates a rectangular diaphragm type pressure sensor 200 according to an embodiment of the present invention. In addition to a strained Hall cross 210 at the edge of the membrane 202, a similar Hall cross 220 is patterned over an unsuspended area of the chip 205 as a reference. The resistance response to a sweeping in-plane magnetic field from both the sensor Hall cross 210 and reference Hall cross 220 are shown in FIGS. 14(a) and (b), respectively. In one aspect, signals from the Hall crosses 210 and 220 are provided to differential amplifiers 222 and 224, respectively, the outputs of which are provided to final stage amplifier 240. The differential amplifiers 222, 224 and 240 may be provided on chip 205 or they may be off-chip and electrically coupled with sensor 200. The signal 230 provided by final stage amplifier 240 is also shown in FIG. 14(c). The pressure induced strain shifts the magnetic switching and yields a pulse-like signal 230 as the output of the final differential amplifier 240. The width of the pulse reflects the strength of the applied pressure. In certain aspects, the signal 230 (and/or the signals output by each of the differential amplifiers 222 and 224 in certain aspects) is processed, for example, using a microprocessor or other intelligence module, to determine various parameters associated with the applied pressure, such as for example, the strength, or magnitude, as well as timing, frequency and other parameters.

[0102] Pressure Sensor: Fabrication

[0103] In preferred aspects, a sensor is fabricated on a ferromagnetic semiconductor-based epitaxial heterostructure, e.g., a (Ga, Mn)As epitaxial heterostructure. For example, in one embodiment, the epitaxial heterostructure includes (Ga,Mn)As (150 nm)/GaAs (150 nm)/Al0.7Ga0.3As (400 nm)/GaAs (100 nm) layers on a <100> GaAs substrate (300 μm thick) as shown in FIG. 15(a). The various layers may be formed by various techniques as are well known in the art. Such techniques include low temperature molecular beam epitaxy (MBE), chemical vapor deposition (CVD), diffusion, and other similar techniques. It should be appreciated, however, that other materials and thicknesses, as well as different concentrations of materials, may be used. In particular, type III-V semiconductor materials are preferred.

[0104] FIGS. 15(a)-(d) illustrate a process of fabricating a pressure sensor device according to an embodiment of the present invention. First, with reference to FIG. 15(a), an epitaxial heterostructure (e.g., (Ga,Mn)As/GaAs/Al0.7Ga0.3As/GaAs) is formed on a substrate 300 (e.g., wafer) and the bottom side of the substrate is polished. Next, a rectangular window(s) 310 are photolithographically patterned in thick photoresist. A cavity 320 is formed, e.g., by wet chemical etching through a photoresist mask (FIG. 15(b)). In one aspect, the etching is anisotropic, and the AlGaAs layer (e.g., 400 nm thick) serves as the etch stop, which is in-turn removed in diluted HF. The top (Ga,Mn)As layers are patterned (e.g., to provide gage 120 (FIG. 12) or Hall gages 210 and 220 (FIG. 13)). The cavity 320 is then sealed, e.g., by low temperature bonding, to a Si or other substrate (FIG. 15(d)). (see, e.g., Z. Ma, G. L. Zhou, H. Morkoc, L. H. Allen, K. C. Hsieh, Appl. Phys. Lett. 64, 772 (1994).)

[0105] Pressure Sensor: Packaging

[0106] Once the sensor devices are fabricated, the individual dies are diced and packaged in a non-magnetic plastic carrier in order to make the electrical contacts, as well as form the inlet port to apply the pressure. In one aspect, a wire-wound magnetic coil is included, e.g., to sweep an in-plane magnetic field.

[0107] Read Head Sensors

[0108] In certain aspects, the materials and devices of the present invention are particularly useful for reading magnetic domains as are typical in HDD media and other magnetic storage media. Given the increased sensitivity to magnetic fields as discussed above, materials of the present invention advantageously provide enhanced magnetic field and domain change detection capabilities as are desirable in read head technology to enable significant scaling and size reduction.

[0109] According to one embodiment, a read head (or merged read/write head) for reading a magnetic medium, such as an HDD read head, is configured with a ferromagnetic semiconductor epitaxial heterostructure configured to provide GPHE signals as discussed herein. In preferred aspects, the epitaxial heterostructure includes Mn doped GaAs epilayer(s) on a GaAs substrate layer as disclosed herein above. Alternatively, any type III-V semiconductor material may be used as the substrate and epilayer(s), such as for example, one or more epilayers of Mn doped GaN on a GaN substrate. Voltage probes (such as provided on device 10) and other electrical contacts may be provided as desired to optimize field/domain detection and signal analysis as well as device construction.

[0110] When such GPHE materials and structures are incorporated in a read head as shown in FIG. 16, the GPHE-based read head 400 provides enhanced magnetic signal detection capabilities relative to known read heads, such as for example, MR- and GMR-based read heads. As discussed above, the sensitivity is typically 4 orders of magnitude greater than conventional technologies, thereby enabling enhanced signal detection and concomitantly scaling and dimension reduction in HDD and other magnetic media. FIG. 17 shows a merged read/write head configuration including a GPHE sensor. The magnetic recording medium is typically circular, including a plurality of tracks arranged in circular order on the medium. Reading (and writing) typically includes rotating the magnetic medium with the read (write) head proximal the rotating medium. In certain aspects, a read head according to the present invention includes one or more integrated electric coils for generating a saturation magnetic field within the read-head materials. U.S. Pat. Nos. 6,519,119, 6,542,341, 6,549,382, 6,563,678, 6,563,679, and 6,580,270, the contents of which are hereby incorporated by reference in their entirety, discuss various aspects of known read heads and non-GPHE read head materials.

[0111] While the invention has been described by way of example and in terms of the specific embodiments, it is to be understood that the invention is not limited to the disclosed embodiments. To the contrary, it is intended to cover various modifications and similar arrangements as would be apparent to those skilled in the art. Therefore, the scope of the appended claims should be accorded the broadest interpretation so as to encompass all such modifications and similar arrangements.

Claims

1. A pressure sensing device, comprising: a semiconductor housing structure having an opening defined therein, said opening having a perimeter; a thin semiconductor membrane covering the opening so as to define an enclosed cavity within the housing structure, said membrane defining a pressure sensing region within the perimeter; and a ferromagnetic semiconductor Hall bar gage structure positioned proximal at least a portion of the perimeter of the pressure sensing region; wherein the Hall bar gage structure produces a signal responsive to a deflection of the membrane in said pressure sensing region due to a pressure difference between the interior of the cavity and the exterior of the cavity, said signal being proportional to the pressure difference.

2. The device of claim 1, wherein the membrane includes one of GaAs and GaN.

3. The device of claim 1, wherein the Hall bar gage structure includes one of Mn doped GaAs and Mn doped GaN.

4. The device of claim 1, wherein the housing structure includes one or more of GaAs, GaN, and Si.

5. The device of claim 1, wherein the pressure sensing region of the membrane is substantially circular.

6. The device of claim 1, wherein the pressure sensing region of the membrane is substantially rectangular.

7. The device of claim 6, further including a second ferromagnetic semiconductor Hall bar gage structure positioned on the membrane away from the pressure sensing region, wherein said second Hall bar gage provides a reference signal.

8. The device of claim 7, wherein the signal and the reference signal are processed to determine one or more parameters associated with the pressure difference on the membrane in the sensing region.

9. A method of producing a ferromagnetic semiconductor-based pressure sensor, comprising: providing a substrate; forming an epitaxial heterostructure comprising two or more layers on the substrate; forming a cavity in the substrate such that the cavity is exposed to a portion of a first one of said two or more layers of the epitaxial heterostructure, the exposed portion of the first layer defining a sensing region having a perimeter; sealing the cavity; patterning the layer adjacent the first layer so as to form a Hall bar gage structure proximal the perimeter of the sensing region and so as to expose the sensing region of the first layer to the atmosphere; wherein the Hall bar gage structure produces a signal responsive to a deflection of the first layer in said pressure sensing region due to a pressure difference between the interior of the cavity and the exterior of the cavity, said signal being proportional to the pressure difference.

10. The method of claim 9, wherein the substrate is one of a GaAs substrate and a GaN substrate.

11. The method of claim 9, wherein the heterostructure includes GaAs in the first layer and Mn doped GaAs in the adjacent layer.

12. The method of claim 11, wherein the heterostructure further includes AlGaAs in a second layer between the first layer and the substrate, said second layer serving as an etch stop during the step of forming the cavity.

13. The method of claim 9, wherein the heterostructure includes GaN in the first layer and Mn doped GaN in the adjacent layer.

14. The method of claim 9, wherein the sensing region is substantially circular.

15. The method of claim 9, wherein the sensing region is substantially rectangular.

16. The method of claim 9, wherein sealing includes bonding the substrate to a second substrate so as to seal the cavity from the atmosphere.

17. The method of claim 9, wherein the epitaxial heterostructure is formed using molecular beam epitaxy.

18. A ferromagnetic semiconductor-based read head sensor configured to detect magnetic domain orientations in a magnetic recording medium having a plurality of domains, each domain having a magnetization, the sensor comprising: a substrate defining a plane; a ferromagnetic semiconductor epilayer formed on said substrate, said epilayer having a cubic hard axis; and first and second read current contacts, each contact coupled proximal an end of the epilayer, said contacts being configured to provide an electrical current flow along the hard axis; and one or more read probes, in electrical contact with the epilayer, configured to detect transverse magnetic resistance in the epilayer; wherein application of an in-plane magnetic field, non-aligned with the cubic hard axis, produces a transition in the transverse magnetic resistance of the epilayer, and wherein the magnetization of each domain produces a magnetic field having a component non-aligned with the cubic hard axis when the read head is positioned proximal thereto.

19. The sensor of claim 18, wherein the epilayer is substantially elongated and oriented along the cubic hard axis.

20. The sensor of claim 18, wherein the substrate is one of a GaAs substrate and a GaN substrate, and wherein the epilayer includes one of a Mn doped GaAs layer and a Mn doped GaN layer.

21. The sensor of claim 18, wherein the epilayer includes a type III-V semiconductor material.

22. The sensor of claim 18, further including at least one electric coil proximal the substrate and epilayer for generating a saturation magnetic field of desired orientation and magnitude within the epilayer.

23. A method of detecting changes in magnetic domain orientations in a magnetic recording medium using a ferromagnetic semiconductor-based read head sensor, the method comprising: positioning a read head sensor proximal a magnetic recording medium having a plurality of domains, each domain having a magnetization, wherein the read head sensor includes a ferromagnetic semiconductor epilayer structure defining a plane and having a cubic hard axis; moving the read head position relative to the domains in a sequential order; and detecting changes in the transverse magnetic resistance of the epilayer structure; wherein application of an in-plane magnetic field, non-aligned with the cubic hard axis, produces a transition in the transverse magnetic resistance of the epilayer, and wherein the magnetization of each domain produces a magnetic field having a component non-aligned with the cubic hard axis when the read head is positioned proximal thereto.

24. The method of claim 23, wherein the substrate is one of a GaAs substrate and a GaN substrate, and wherein the epilayer includes one of a Mn doped GaAs layer and a Mn doped GaN layer.

25. The method of claim 23, wherein the magnetic recording medium is substantially circular, and wherein moving includes rotating the magnetic recording medium.

26. The method of claim 23, wherein the epilayer includes a type III-V semiconductor material.

27. The method of claim 23, further including generating a saturation magnetic field of desired orientation and magnitude within the epilayer using at least one electric coil positioned proximal the substrate and epilayer.