Magnetic Anisotropy of FeNi Multilayer Films with Different Orientations of the Magnetic Anisotropy Axes in Adjacent Layers

Abstract

FeNi films were prepared using the DC magnetron sputtering technique with an oblique deposition arrangement. Multilayers with different orientations of the magnetic anisotropy axes were obtained thanks to a rotary sample holder inside the vacuum chamber. Magnetic properties were studied using magneto–optical Kerr microscopy and a vibrating sample magnetometer. Single-layered FeNi films having thicknesses as high as 10 nm and 40 nm show in-plane uniaxial easy magnetization axes produced by the oblique incidence of incoming components of the beams. Magnetic anisotropy field for four-layered samples with orthogonal uniaxial magnetic anisotropy axes in the adjacent layers and the thickness of individual layers of 10 nm and 40 nm turned out to be less than in single-layered films. The magnetic properties peculiarities of the eight-layered sample FeNi (10 nm) × 8 obtained by rotation of the sample holder by 45° before deposition of each subsequent layer suggest the formation of a helix-like magnetic structure through the thickness of the multilayered sample similar to the magnetization arrangement in the Bloch-type magnetic domain wall.

1. Introduction

Magnetic anisotropy is one of the main parameters of functional magnetic materials. It is especially important in the case of thin magnetic films—a basic type of material of spintronics and spin-orbitronics devices that have a tendency to become smaller and smaller [1]. The reduction of the size makes the shape anisotropy contribution more and more essential, and therefore, overall control of all contributions to the effective anisotropy is a challenge. There are cases in which the competition between induced magnetic anisotropy and shape anisotropy is crucial for small devices’ functionality. For example, in magnetoimpedance (MI) magnetic field sensors, the highest sensitivity can be obtained in the case of transverse-induced magnetic anisotropy for a sensitive element in the shape of a rectangle. The direction of the long side of the rectangle corresponds to the orientation of the magnetization for which the magnetic shape anisotropy contribution is minimal [2]. During magnetoimpedance sensor functionality, the high-frequency alternating current flows through a magnetic conducting element and creates an alternating electromagnetic field. The highest sensitivity to the applied external magnetic field appears near the magnetic anisotropy field of the thin film-sensitive element, being the order of a few Oe [3]. Although different soft magnetic films and their combinations were proposed for MI applications, the most studied material was Fe₂₀Ni₈₀ magnetically soft permalloy with high magnetic permeability and close to zero magnetostriction [4]. Different spacers were used in order to design the best MI structures, and even multilayered structures without spacers were checked using the technological “stop deposition” condition when shatter closes the substrate, and the surface of the deposited film becomes passivated by the atoms left in the deposition atmosphere [3].

A number of practical applications require magnetically isotropic films [5,6]. Important examples of such applications are the mapping of elementary topological configurations of inhomogeneous magnetic fields or soft magnetic films for magnetoresistive sensing applications. Optimization of the magnetic properties of thin magnetic films requires an understanding of the physical origin of magnetic anisotropy. The most important point is understanding the relationships between microstructure and magnetic domain structure in soft magnetic thin films. Technological and biomedical applications require the development of magnetic materials in planar geometry with different types of effective magnetic anisotropy: isotropic, anisotropic, and helicoidal. As an example of the applications of FeNi thin films with well-defined uniaxial magnetic anisotropy, one can mention magnetoimpedance multilayered structures for which the highest sensitivity with respect to the external magnetic field can be achieved in the sensitive elements with transverse induced magnetic anisotropy [7]. Worth mentioning that different applications of magnetic thin films require different thicknesses. This means that magnetic anisotropy of a particular type may be requested for thin films of very different thicknesses.

One of the effective ways to reduce magnetic anisotropy is the deposition of thin magnetic films in the presence of an external magnetic field rotated in the plane of the substrate [6,8]. Another possible method of anisotropy reduction is the creation of a multilayered film system in the adjacent magnetic layers of which the magnetic anisotropy axes are orthogonal to each other. From a formal point of view, such a two-layer system, consisting of two identical layers possessing uniaxial magnetic anisotropy, should be magnetically isotropic [9]. In the simplest case, it is assumed that there is no exchange and magnetostatic interaction between the adjacent layers. This condition can be ensured by introducing a thin, non-magnetic spacer between the magnetic layers [8]. In Ref. [10], it was shown that in the case of the strong exchange coupling, a simple model predicted nearly isotropic behavior for thin magnetic films much thinner than a Bloch magnetic domain wall width. Bloch wall is a transition region between neighboring magnetic domains, over which the magnetization changes from its value in one domain to that in the next, rotating through the plane of the wall. However, it was shown in practice that the magnetically isotropic state is not realized for multilayered systems with interlayer interaction. In particular, it was found [10,11] that for the case with the same thickness of layers, the second deposited film determined the magnetic anisotropy behavior of both films in the double layer. The reason for the predominant influence of the upper layer is considered to be a difference in the microstructure of the layers governed, for example, by the defects distribution [10] or a decrease in the magnetic anisotropy of the lower layer as a result of its possible heating in an orthogonal magnetic field during the deposition of the upper layer [11]. It was also found that for bi-layered films with orthogonal anisotropy axes in adjacent layers, the values of the anisotropy constants are smaller than the values corresponding to each individual magnetic layer [11,12].

The possible undesirable influence of the magnetic field on the magnetic anisotropy of the previous layer during the deposition of the subsequent layer can be avoided by the implementation of different deposition methods for obtaining layers with uniaxial magnetic anisotropy, namely, oblique deposition. For deposition in such geometry, the direction of the flow of particles and clusters deposited on the substrate deviates from the normal direction to the substrate plane [13]. In this case, the orientation of the magnetization is determined by the anisotropy of the shape of the components of the microstructure, arising due to the self-shading effect of the columnar microstructure [13]. The in-plane easy magnetization axis is oriented perpendicular to the incidence plane of the atoms or clusters of flux if incidence angles are less than 60–70° [13,14]. Nowadays, the method of oblique deposition is very often used for the deposition of films and multilayers and exchange bias structures with varied effective magnetic anisotropy in each one of the layers [15,16,17,18,19,20,21]. The reason is either to control the magnetotransport properties of granular films or to tune the magneto-optical properties of thin films of pitched columnar type. In addition, oblique deposition is a very promising technique for the synthesis of thin films and multilayered structures with high ferromagnetic resonance frequencies requested in many high-frequency applications [22].

The success of obtaining anisotropic FeNi films by oblique deposition has been repeatedly demonstrated [13,22,23,24,25]. But this method for thin film deposition does not exclude the occurrence of resulting magnetic anisotropy in the plane of bilayers with orthogonal uniaxial magnetic anisotropies, although in that case, the anisotropy constants are much smaller than the values of the magnetic anisotropy corresponding to each individual layer [23,26]. With the same layer thickness, the easy magnetization direction of the effective uniaxial magnetic anisotropy was tilted by 45° from the easy magnetization axis of each individual layer [23]. With an increase in the thickness of the magnetic layers, the formation of a Bloch-type magnetic domain wall extending across the thickness of the bilayer is possible [23]. Such behavior creates a perspective for the preparation of thin magnetic films with a helical magnetic anisotropy by the deposition of the multilayered structures with different orientations of the easy magnetization axis of each individual layer and possible manipulation of the magnetization of helimagnets through the spin-transfer torque effect [27]. The spin-transfer torque phenomenon is an effect in which the orientation of a magnetic layer in either a magnetic tunnel junction or spin valve appears to be modified using a spin-polarized current.

It was recently shown that the transfer of spin momentum in a helimagnetic system leads to rotation of the spiral magnetization structure of the helimagnet around its axis under the influence of a spin current flowing in the helimagnet. An estimate for the frequency of such rotation of the magnetization spiral was also given. It was proposed to use such magneto-chiral structures for the creation of a generator of electromagnetic radiation with tunable frequency [27,28]. In addition, chirality-based all-optical logic gates have recently been experimentally demonstrated. This direction holds great potential for optical computing logic [29]. Perhaps, by analogy with chirality-based all-optical logic, magneto-chiral logic gates will be created if one develops suitable magneto-chiral structures. An example of “natural” carriers of magnetic structures with helical anisotropy are single crystals of dysprosium and holmium, in which a helical magnetic order is realized at low temperatures [30,31]. It is also possible to achieve the formation of helical anisotropy in a FeNi film by deposition in a rotating external magnetic field [32]. However, in this case, the film thickness should be approximately 1 µm. However, it is too large for the implementation in the devices based on the spin-orbit torque effect using chiral helimagnet [27,33].

In this work, thin permalloy films obtained by DC magnetron sputtering using the oblique deposition method were prepared and studied. The main features of the induced magnetic anisotropy were also studied in multilayered film structures without additional separating non-magnetic layers with different orientations of the magnetic anisotropy axes in adjacent magnetic layers. The features of magnetic anisotropy for four-layered samples having orthogonal uniaxial magnetic anisotropy axes in the adjacent layers were compared with the existing literature data for two-layer films. The possibility of creating a helix-like magnetic structure through the thickness of the multilayered sample similar to the magnetization arrangement in the Bloch-type magnetic domain wall by increasing the number of layers and decreasing the angle between the anisotropy axes of adjacent layers was also analyzed.

2. Materials and Methods

FeNi thin films and FeNi multilayered samples were deposited onto commercial Corning glass substrates using the DC magnetron sputtering technique without additional surface modifications of the substrates. All depositions were done at room temperature using a Fe₂₀Ni₈₀ alloyed target. The composition of the deposited films was checked to be close to the composition of the target. Background pressure inside the vacuum chamber was as high as 6 × 10⁻⁷ mbar, and the working argon gas pressure during deposition was kept constant close to the value 2.6 × 10⁻³ mbar. The oblique deposition angle defined with respect to the normal surface of the substrate was 35°. The deposition rate was about 0.13 nm/s. It was defined prior to the deposition of the investigated samples. A film about 100 nm thick with a sharp step was deposited for calculation of the deposition rate. The exact thickness of the film (for known deposition time) was measured by the atomic force microscope (NT-MDT Ltd., Moscow, Russia) and Dektak 150 Surface Profiler (Veeco Instruments Inc., Plainview, NY, USA). The thickness of the layers in the experimental studies was selected to be 10 nm or 40 nm. Multilayered structures with different orientations of the magnetic anisotropy axes were obtained using a rotary sample holder inside the sputtering vacuum chamber. For different measurements, the whole glass substrates were cut into smaller pieces of appropriate size.

The microstructure was studied by transmission electron microscopy (TEM) using a JEM-2100 (JEOL Ltd., Tokyo, Japan) microscope operating at the accelerating voltage of 200 kV. The selected area electron diffraction (SAED) was employed to determine the structure of the crystallites. For TEM measurements, FeNi films were deposited onto NaCl single crystal substrates, on which a 5 nm thick layer of SiO₂ was previously deposited. The magnetic properties of the samples, i.e., surface hysteresis loops and magnetic domain structure for different values of the applied magnetic field, were measured by a magneto-optical Kerr effect (MOKE) using the Evico Magnetics microscope equipped with a digital camera. Both wide and narrow field microscope configurations were used. A lakeshore vibrating sample magnetometer (VSM) was employed for the measurements of in-plane magnetic hysteresis loops at room temperature.

3. Results and Discussion

Figure 1 shows the TEM images for the FeNi (10 nm) and FeNi (40 nm) films. It confirms the polycrystalline structure of both films and indicates that the crystallites are randomly oriented. The SAED pattern (insets in Figure 1) confirms the fcc (face cubic centered) crystalline structure of the films of both thicknesses. The mean crystallite size of the grains was about 10 nm for FeNi (40 nm) film and 3–4 nm for FeNi (10 nm) samples. Despite the difference in the structural state, both films had well-defined uniaxial magnetic anisotropy in each sample plane.

Figure 2 shows the magnetic hysteresis loops measured on each sample along the easy magnetization axis (EA) and perpendicular to it, i.e., along the hard magnetization axis (HA). For each sample, the easy magnetization axis was expectedly formed during the deposition process in the direction perpendicular to the plane of incidence of the deposited material.

In this case, the anisotropy field H_a = 8 ± 0.5 Oe for the FeNi (10 nm) film and for the FeNi (40 nm) film H_a = 5 ± 0.5 Oe. The presence of uniaxial magnetic anisotropy is also confirmed by the shape of the angular dependences of M_r/M_s curves, constructed on the basis of hysteresis loops measured in different directions in the plane of the studied films (Figure 2c,d). On the graph, the φ angle is measured from the induced easy magnetic anisotropy axis. In addition to the above-mentioned, one can observe that all deposited samples were of very good quality with low deviations of the anisotropy from the expected value.

Let us now analyze the FeNi multilayered structures. In the case of deposition of four-layered films, it was assumed that individual uniaxial magnetic anisotropy would be formed in each layer due to the deposition conditions, oblique deposition. By rotating the sample holder, the easy magnetization axis of each subsequent layer should be orthogonal to the easy magnetization axis of the previous layer. Based on a set of magnetic hysteresis loops measured by means of a vibrating sample magnetometer in the plane of four-layered samples, it was found that the films have a resulting uniaxial magnetic anisotropy. Figure 3 shows hysteresis loops measured along the resulting easy magnetization axis and hard magnetization axis for the samples with individual layer thicknesses of 10 nm or 40 nm (FeNi (10 nm) × 4 and FeNi (40 nm) × 4). For each sample, the easy magnetization axis turned out to be oriented at an angle of 45° with respect to the easy magnetization axis orientation of the first magnetic layer (Figure 3c,d). In this case, the value of the magnetic anisotropy field for both types of samples coincides with the measurement error corresponding to the employed technique (H_a = 3.5 ± 0.5 Oe), and it is smaller than the anisotropy field for single-layered films.

For single-layer and four-layer FeNi films, the angular dependencies of the coercive force H_c(φ) were also plotted (Figure 4a,b). The found values of the anisotropy field for single-layer and four-layer films were used to describe the angular dependence of the coercive force H_c(φ) within the Stoner–Wohlfarth model using the equation H_c(φ) = H_a × |cosφ|, where φ is the angle between the measurement field and the easy magnetization axes [34]. It can be seen that for both samples, the best agreement is observed in the range of angles ±15° near the hard axis, where magnetization reversal of the film occurs through coherent rotation of magnetic moments. In other angular ranges, the difference between the experimental and model dependences H_c(φ) is due to the fact that magnetization reversal of the film occurs both through the coherent rotation of magnetic moments and the movement of domain walls. Thus, the magnetization reversal process is the same for single-layered and four-layered FeNi samples. These films differ from each other only by the value of the magnetic anisotropy field. These results are consistent with the observations reported for two-layered FeNi systems with orthogonal magnetic anisotropy axes [26]. Note that in the above-mentioned work [26], the magnetic layers of different (Fe₁₀Ni₉₀) compositions were used, and the magnetic anisotropy field of individual Fe₁₀Ni₉₀ magnetic layers was as high as 100 Oe.

Figure 5 shows Kerr images of the magnetic structure of the samples obtained by reversing the magnetization of a FeNi (40 nm) × 4 multilayers along the resulting easy magnetization axis and perpendicular to it using a wide-field Kerr microscope mode for the longitudinal Kerr effect configuration. Generally speaking, magneto-optical Kerr microscopy allows the collection of magnetic information from the metallic sample surface at a depth of about 20 nm. It is defined as “the information depth” [35]. Therefore, Kerr image data characterize the behavior of the top layer of the sample. Considering the deposition conditions, its individual easy magnetization axis should have been formed at an angle of 45° defined with respect to the position of the effective, easy magnetization axis. However, Kerr images of the magnetic domain structure indicate that the effective, easy magnetization axis of this layer coincides with the effective, easy magnetization axis of the whole sample: magnetization reversal along the easy magnetization axis is carried out due to the displacement of the magnetic domain boundaries (Figure 5a–c), and along the hard magnetization axis due to the process of the rotation of the magnetization vectors (Figure 5d,e). In the present study, the signs of the formation of a Bloch-type magnetic wall extending across the whole thickness of the sample were not detected either using measurements with a magnetometer or using a MOKE Kerr microscope, in contrast to the results of the following work [23].

A multilayered sample containing eight layers (FeNi (10 nm) × 8), each one having a thickness of 10 nm, was prepared aiming to form a thin film magnetic structure through a sample thickness similar to the Bloch-type magnetic wall. Before deposition of each subsequent layer, the substrate holder was rotated by an angle of 45°. VSM analysis of magnetic hysteresis loops measured in the plane of the sample indicates that the orientation of the effective, easy magnetization axis of the sample differs from the orientation of the expected easy magnetization axis of the first layer of the multilayered structure by an angle of 60° (Figure 6b).

Hysteresis loops measured in this direction and perpendicular to it are shown in Figure 6a. Unlike all previous samples, the easy-axis hysteresis loops become less squared, and the hard-axis loops exhibit a higher coercivity. Despite the observed effective uniaxial anisotropy, the type of hysteresis loops and the shape of the curve describing the angular dependence of the remanent magnetization indicates a large degree of dispersion of the effective magnetic anisotropy axes (compare Figure 3 and Figure 6). This is also confirmed by the fact that the angular dependence of the coercive force H_c(φ) could not be described using the H_c(φ) = H_a × |cosφ| (Figure 4c).

As was noted previously, in two-layered systems with orthogonal anisotropy axes in adjacent identical layers deposited in the presence of an external magnetic field, the upper layer is dominant in the formation of effective magnetic anisotropy of the two-layered sample [10,11]. We do not discuss here all possible reasons for such an asymmetry (difference in thermal conditions, different conditions of crystallization or re-sputtering, etc.). However, in the scientific literature, there are examples of similar behavior, which must be taken into account for the successful development of micro- and nanoelectronics devices having complex fabrication technologies and complex thin film-based patterned structures [7,36].

In the FeNi (10 nm) × 8 multilayered sample described above, the easy magnetization axis of the top layer also coincided with the effective, easy magnetization axis of the whole sample. This is evidenced by the MOKE hysteresis loop measured along the effective, easy magnetization axis of the sample (Figure 7a). The corresponding Kerr images are shown in Figure 7c. On the one hand, they indicate that magnetization reversal occurs by domain wall displacements, which is typical for thin films with uniaxial magnetic anisotropy. On the other hand, a more complex structure inside the large magnetic domains is visible, in contrast, for example, to the FeNi (40 nm) × 4 multilayered sample (Figure 5). Most likely, this is a consequence of the complex magnetic structure through the thickness of the whole sample [37]. Additional confirmation of this assumption is provided by Kerr images obtained by reversing the magnetization of the sample along the hard magnetization axis (Figure 7b,d). Taking into account the fact that, in this case, the Kerr images are the averaged signal of approximately two upper layers (due to the limitation of the “information depth”), it seems that the averaged magnetic anisotropy axis corresponding to these Kerr images changes its orientation when the external magnetic field changes. This situation is apparently possible if the magnetic structure of the sample is similar to the Bloch-type magnetic wall magnetic structure. A magnetization helix magnetic structure across the sample thickness was previously observed in bilayer films having orthogonal, in-plane easy magnetization axes [37].

Some confirmation of the helix-like magnetic structure through the thickness of the whole sample comes from the magnetic hysteresis loop measured from the side of the glass substrate at an angle of 45° with respect to the expected easy magnetization axis of the first magnetic layer (Figure 8a). The shape of the hysteresis loop and the corresponding Kerr effect images show that there is an asymmetrical magnetization reversal of this part of the film in two opposite directions (Figure 8a,c). When the applied magnetic field changes from +H_max to –H_max, the magnetization reversal occurs due to the rotation of the magnetization vectors. In the opposite direction, magnetization reversal is carried out mainly through magnetic domain wall displacement. The MOKE hysteresis loop measured along this axis from the film side has quite the usual symmetric shape, and the magnetization reversal process occurs through a combination of magnetization rotation and domain wall displacement (Figure 8b,d).

The possibility of obtaining a thin magnetic film with helical magnetic anisotropy was demonstrated more than 70 years ago. For this purpose, the FeNi film was deposited in a rotating magnetic field [32]. In this case, the film thickness should be equal to a critical thickness l_cr = 2π(A/K)^1/2, where A is the exchange constant, and K is the anisotropy constant. In this case, the magnetic anisotropy field overpowers the exchange interaction field. For FeNi films, assuming K = 2 × 10³ erg/cm³, and the exchange constant A = 5.5 × 10⁻⁷ erg/cm value, one predicts l_cr = 1.0 µm [32]. However, this thickness is too large for the implementation of the devices based on the spin-orbit torque effect using chiral helimagnet [27,33]. Creating multilayered structures with sequential changes in the orientation of the magnetic anisotropy axis in adjacent layers seems to be a promising way to reduce the value of the critical thickness l_cr. This strategy will be facilitated by a decrease in the exchange interaction between the layers in comparison with the exchange interaction level in a continuous film.

In turn, the value of the exchange interaction between the layers will depend on the microstructure (orientation, maybe even size due to the grain boundary contributions of the columns) of adjacent layers during inclined deposition, as well as the duration of the possible pause between the deposition of layers. The last technological method was shown to be quite efficient for the deposition of the multilayers used for sensor applications. It seems interesting to obtain multilayered structures with different layer thicknesses, as well as with different values of the induced magnetic anisotropy constant. Oblique deposition allows success and reaching the goal in a quite simple way by changing the incidence angles of a flux of depositing material [38,39]. In addition to the other advantages of the proposed technique for the fabrication of thin magnetic films with helical anisotropy, one can add that it is simple, cheap, and basic properties of FeNi films can be well adjusted to the desired level using existing technological steps.

4. Conclusions

Single-layered and multilayered FeNi films prepared by DC magnetron sputtering using the oblique deposition method have been investigated using Kerr microscopy and vibrating sample magnetometer. The thickness of the magnetic layer was either 10 or 40 nm. Single-layered FeNi films had uniaxial magnetic anisotropy in the sample plane. Exchange-coupled four-layer samples with orthogonal uniaxial axes were also characterized by uniaxial magnetic anisotropy but with a smaller value of the magnetic anisotropy field in comparison with the single-layered films. In addition, the magnetic anisotropy axis was oriented at an angle of 45° to the axes of the individual layers. The shape of the hysteresis loops and images of the magnetic domain structure of the eight-layered sample with different orientations of the magnetic anisotropy axes in adjacent layers suggest the formation of a magnetic structure through a sample thickness similar to the Bloch-type magnetic domain wall. This motivates us to continue research in the described direction, promising the possibility of constructing thin magnetic films with helical anisotropy, which can be very useful for the construction of spin-orbitronics devices and their applications.