Study of magnetization reversal in Neel and Bloch regime of Ni and Py stripes using Kerr microscopy

TL;DR

This study investigates how magnetization reverses in Ni and Py micro-stripes with Neel and Bloch domain walls, revealing differences in domain wall density and propagation behavior using Kerr microscopy.

Contribution

It provides a systematic comparison of magnetization reversal in Neel and Bloch regimes for Ni and Py stripes, highlighting domain wall density differences.

Findings

Higher domain wall density in Bloch regime.

Domain propagation varies with stripe width and magnetic field direction.

Nucleation starts in lower width regions.

Abstract

We present a systematic study of the magnetization reversal of nickel and permalloy micro-stripes with Neel and Bloch domain walls using Kerr microscopy. Magnetic field driven domain propagation was observed from higher width to lower width stripes for magnetic fields applied along the length of micro-stripes. Stripe like domains were observed with nucleation starting in lower width region followed by their propagation to higher width regions for magnetic fields applied along the width of micro-stripes. The comparison of magnetization reversal in Bloch and N\'eel domain wall regime showed higher domain wall density in Bloch regime for both nickel and permalloy stripes.

Full text

magnetization reversal, Kerr microscopy, Shape anisotropy, Néel domain wall, Bloch domain wall,

I Introduction

Patterned magnetic thin films is a vast area of research from several perspectives55 ; 56 ; 57 . First, they are a good candidate to study the effect of lateral confinement on the spin re-arrangements and magnetization reversal processes1 ; 2 ; 3 ; 4 ; 5 ; 6 ; 7 ; Second, they provide a basis to understand the new physical phenomena which are of fundamental interest and key to future applications in hard disk drives, magnetic random access memory (MRAM) and other spintronics devices8 ; 9 . Using patterned magnetic films, we can control the local magnetic properties such as magnetization distribution and reversal by changing the shape and hence, magneto-static fields2 ; 10 . The magnetization reversal of one or more ferromagnetic components is the key functional aspect of spintronics devices. For example, lateral spin valves are one of the most widely utilized device structures in spintronics14 ; 15 . These devices involve the use of magnetic electrodes patterned in the lateral direction37 . Therefore, a systematic study of magnetization reversal for patterned ferromagnetic materials such as nickel and permalloy stripes with different aspect ratios is relevant for spintronics devices. In this context, various studies38 ; 39 ; 40 ; 41 ; 47 ; 48 ; 49 ; 50 ; 52 have been done. These studies have considerably improved the basic understanding of magnetization processes in patterned magnetic structures. In the same context, we present a comparison study of a controlled propagation of domains in connected micro-stripes for nickel and permalloy in Néel and Bloch regime of thickness. This study is also useful for domain wall superconductivity and domain wall induced spin-polarized supercurrents in superconducting-spintronics51 ; 54 .

In this report, we investigate magnetic field driven magnetization reversal and magnetic domain patterns in Néel and Bloch regime of nickel and permalloy stripes using Kerr microscopy. We observe a controlled propagation of domains due to shape anisotropy for magnetic fields applied along the length of stripes. We observe stripe like domains for magnetic fields applied along the width of stripes.

II EXPERIMENTAL DETAILS

Nickel and permalloy thin films with thickness of 15 nm and 90 nm were prepared at room temperature using DC magnetron sputtering in an Ar pressure of 1.5 Pa using high purity ($99.99\%$) nickel and permalloy targets on cleaned Si/SiO2 substrates. The base pressure of the deposition system was $2\times 10^{-8}$ mBar. Nickel and permalloy films were patterned using photo-lithography and Ar-ion milling techniques, as shown in Fig. 1(a). The magnetic domains were imaged by high resolution magneto-optical Kerr microscope supplied by M/s Evico Magnetics, Germany. All magneto-optical Kerr effect measurements were performed at room temperature in the longitudinal mode. The ramp rate of magnetic field was kept constant as 0.1 mT/s for all the measurements. M(H) loops were measured by selecting an area of interest in the stripe of a particular aspect ratio. The Kerr rotation and hence, the average intensity of the domain image in the region of interest is proportional to the magnetization of that region10 ; 11 . Therefore, the M(H) loop is a characteristic of an individual stripe with a specific aspect ratio. The coercive fields for the stripes of different aspect ratios were extracted from the M(H) loops of the respective stripes.

III RESULTS AND DISCUSSIONS

Figure 1(a) shows the FESEM (field emission scanning electron microscope) image of the nickel (15 nm) continuous stripes with widths varying from 5 $\mu$m to 50 $\mu$m but with a fixed length of 100 $\mu$m for each stripe. The consecutive stripes are marked as R1, R2, R3, R4, R5 and R6 with widths of 5 $\mu$m, 10 $\mu$m, 20 $\mu$m, 30 $\mu$m, 40 $\mu$m, and 50 $\mu$m, respectively. The aspect ratio (length/width) of the respective stripes are 20, 10, 5, 3.33, 2.5, and 2, as shown in Fig. 1(a). Since the background motivation was to study the continuous propagation of domain walls, as mentioned in the abstract, experiment had to be performed in connected stripes. Generally, the domain wall energy per unit area (the sum of anisotropy, exchange and stray field energy densities) gradually increases with increasing film thickness for Néel walls, whereas for Bloch walls, the domain wall energy density decreases with increasing film thickness. Therefore, Néel walls are energetically favorable at lower thickness while Bloch walls are favorable at thickness beyond a certain threshold value21 ; 22 ; 23 . It has been predicted theoretically that the crossover thickness in nickel films is $\sim$50 nm24 ; 25 whereas it is $\sim$30 nm in case of permalloy thin films26 . Therefore, throughout the text, we will refer to nickel and permalloy films of 15 nm thickness as the Néel regime and the 90 nm thickness as the Bloch domain wall regime.

In order to compare the magnetization reversal of stripes with different aspect ratios, M(H) loops along with simultaneous domain images were recorded for different stripes for an in-plane applied magnetic field along the length (x-axis) and the width (y-axis) of the stripes. Henceforth, keeping in mind that the magnetic field is always applied in the plane of the substrate, throughout the text, we will refer to these two field configurations as parallel and perpendicular field, respectively. Due to the stepped pattern of the entire stripe, the corners in each segment of the stripe (corresponding to different aspect ratio) may act as a nucleation site for domains. The present experiment does not have a real control over the domain nucleation site. However, even if we assume that nucleation in a certain segment (aspect ratio) starts as a result of propagation of a domain from an adjacent segment, the full magnetization reversal in the said segment should depend on the geometry of that segment. Therefore, even though the numerical value of the obtained coercive field may not be the exact coercive field of that segment if it was isolated, the dependence of coercive field on aspect ratio, in this case, is still a relevant quantity. Fig. 1 shows the coercive fields ($H_{c}$) of nickel and permalloy stripes as a function of their width for Néel and Bloch domain wall regimes in parallel and perpendicular configuration, respectively. In parallel configuration, we observe a decrease in ($H_{c||}$) with increase in the width of the stripes. This shows that the sample geometry has a dramatic effect on the reversal of the magnetization. This is because demagnetizing fields come into picture on reducing the dimensions in patterned structures. The demagnetization field for a single stripe is proportional to $M_{s}d/w$, where $M_{s}$ is the saturation magnetization, d is the thickness and w is the width of the stripe27 ; 28 . The demagnetization field, $H_{d}$, reduces the internal field of a stripe to $H_{i}=H_{a}-H_{d}$, where $H_{a}$ is the applied field. Therefore, the coercive field30 ; 31 ; 32 of stripe of width w becomes:

$H_{c}=H_{0}+A/w$

where $H_{0}$ is the coercive field of a stripe with infinite width (thin film) and A is a constant parameter which depends on the thickness, finite length shape anisotropy factor and saturation magnetization of the stripe. The fitting model assumes that a small part of the stripe reverses magnetization coherently and propagates along the stripe. Here, $H_{0}$ and $A$ are free fitting parameters. The fitting results gave the $H_{0}$ values of 65 Oe and 20 Oe for nickel (15 nm) and nickel (90 nm) films while it was 25 Oe and 2 Oe for permalloy (15 nm) and permalloy (90 nm) films. The values of $H_{0}$ show the soft magnetic nature of nickel and permalloy thin films and are consistent with the literature33 ; 34 . The stripe with lower width (higher aspect ratio) has a larger shape anisotropy and therefore, requires a larger energy to switch, resulting into higher Hc. On the other hand, in perpendicular configuration, we observe an increase in $H_{c}$ with increase in the width of stripes. Due to the shape anisotropy, the short axis of the stripes acts like a hard axis in these patterned structures. The stripe with lower width (higher aspect ratio) has a higher tendency to act like a hard axis, resulting into lower Hc. Similar results were reported in previous literature for magnetic fields applied along the easy axis and hard axis of Co nanowires35 . We emphasize here that for a pure study of the aspect ratio dependence of the coercive field one must do experiments on isolated segments, as reported earlier35 . However, one of the background motives of this work is to study the propagation of domain walls in magnetic stripe in a controllable manner so that one can design experiments where a magnetic domain wall (due to its built-in natural magnetic inhomogeneity) may be used to trigger a functional response. Since propagation is the key issue, the experiment had to be done on a single stripe where the segments with different aspect ratio are connected to each other. The important point we would like to study here is that by successively changing the aspect ratios of the segments whether we are able to get some control over the propagation of the domain walls.

Figure 2(a) shows the M(H) loop for a rectangular region marked in yellow dashed line in Fig. 2(b) for a permalloy (15 nm) stripe of aspect ratio 3.33 (marked as R4 in Fig. 1(a)) in parallel configuration. Fig. 2(b), (c), (d) show the Kerr microscope images of magnetic domains corresponding to points A, B, C marked in Fig. 2(a). We observe a propagation of domains from higher width stripe to the adjacent lower width stripe as shown in Fig. 2(c).

In Fig. 2(e), we show the M(H) loop for a rectangular region marked in yellow dashed line in Fig. 2(f) for a permalloy (15 nm) stripe of aspect ratio 3.33 in perpendicular configuration. We observe that domain nucleation initiates in the lower width region, at some nucleation points as shown in Fig. 2(f). Then, domains started to grow around the nucleation points and stripe like domains were formed giving rise to multiple domain walls throughout the structure, as shown in Fig. 2(g). The formation of multiple domain walls may be due to the minimization of demagnetization energy along the width of stripe. We notice that the sequence in which the domain nucleation starts in different stripes, follows the variation of coercive fields as a function of width shown in Fig. 1(h).

Figure 3(a) shows the M(H) loop for a rectangular region marked in yellow dashed line in Fig. 3(b) for a permalloy (15 nm) stripe of aspect ratio 10 (marked as R2 in Fig. 1(a)) in parallel configuration. We observe that domain nucleation initiates in the higher width (lower aspect ratio) region, at some nucleation points, as shown in Fig. 3(c). A monotonic decrease in the coercivity of alternate stripes results in a propagation of domains from one stripe to other as shown in Fig. 3(c), (d). The comparison of domain propagation in Fig. 2 and Fig. 3 shows that the propagation is more directional and hence, controlled for stripes of higher aspect ratio in parallel configuration.

In Fig. 3(f), we show the M(H) loop for a rectangular region marked in yellow dashed line in Fig. 3(g) for a permalloy (15 nm) stripe of aspect ratio 10 in perpendicular configuration. We observe multiple domain walls throughout the structure, as shown in Fig. 3(g), 3(h). The comparison of Kerr images in Fig. 2 and Fig. 3 shows similar stripe domain patterns for different aspect ratios in perpendicular configuration.

Figure 4(a) shows the M(H) loop for a rectangular region marked in yellow dashed line in Fig. 4(b) for a nickel (15 nm) stripe of aspect ratio 3.33 in parallel configuration. Note that the Kerr images shown in this manuscript have been taken from stripes of aspect ratio 3.33 for representation and comparison purpose. We observe a propagation of domains in the parallel configuration as shown in Fig. 4(c) and (d). Since the growth of domains happens in a particular direction, we propose that a 180*∘* domain wall propagates from wider stripes to narrower stripes of the patterned geometry. R.H.Wade29 in 1962 showed that the polycrystalline nickel thin film of 16 nm thickness comprises of 180*∘* domain walls with a domain wall thickness of 55$\pm$0.5 nm. The comparison of domain propagation in permalloy (15 nm) in Fig. 2 and nickel (15 nm) in Fig. 4 shows that the propagation is more gradual in nickel compared to permalloy.

In Fig. 4(f), we show the M(H) loop for a rectangular region marked in yellow dashed line in Fig. 4(g) for a stripe of aspect ratio 3.33 in perpendicular configuration. The magnetization reversal resulted in a growth of stripe like domains, as shown in Fig. 4(h), 4(i). The comparison of Kerr images in Fig. 2(g) and 4(h) shows that the domain patterns are similar irrespective of the material in perpendicular configuration. This might be due to the dominance of shape anisotropy in the process of domain formation.

Figure 5(a) shows the M(H) loop for a rectangular region marked in yellow dashed line in Fig. 5(b) for a permalloy (90 nm) stripe of aspect ratio 3.33 in parallel configuration. Similar to the case of permalloy (15 nm), domain wall propagates from the stripe with higher width (lower aspect ratio) to the stripe with lower width (higher aspect ratio), as shown in Fig. 5(c) and (d). In Fig. 5(f), we show the M(H) loop for a rectangular region marked in yellow dashed line in Fig. 5(g) for a stripe of aspect ratio 3.33 in perpendicular configuration. Stripe like domains were observed in perpendicular configuration, as shown in Fig. 5(g), (h). The comparison of Kerr images shown in Fig. 2(g) and 5(g) suggests that the domain width decreases in Bloch regime compared to Néel regime and hence, the domain wall density increases in Bloch regime. This may be because of an increase of demagnetization energy with increasing thickness, as evident from the experssion $H_{d}=M_{s}d/w$.

Figure 6(a) shows the M(H) loop for a rectangular region marked in yellow dashed line in Fig. 6(b) for a stripe of aspect ratio 3.33 in parallel configuration. We observe a propagation of domains from higher width region to lower width region as shown in Fig. 6(c), (d). In Fig. 6(f), we show the M(H) loop for a rectangular region marked in yellow dashed line in Fig. 6(g) for a stripe of aspect ratio 3.33 in perpendicular configuration. Stripe like domains were observed in perpendicular configuration, as shown in Fig. 6(g). The comparison of Kerr images shown in Fig. 4(i) and 6(g) suggests that the domain width decreases and hence, the domain wall density increases in Bloch regime compared to Néel regime. This result is similar to the one obtained from comparison of Néel and Bloch regime in case of permalloy. The comparison of domain images in parallel and perpendicular configuration shows a higher domain wall density in perpendicular configuration in all the cases discussed above.

IV Conclusion

We have investigated magnetic field driven magnetization reversal and imaged magnetic domains in nickel and permalloy stripes with Néel and Bloch domain walls in different field orientations by Kerr microscopy. We observe propagation of domain walls in parallel field configuration and growth of multiple stripe like domains in perpendicular field configuration. We observed more controlled propagation of domains in stripes with higher aspect ratio. The comparison of stripes with Néel and Bloch domain walls showed higher domain wall density in Bloch regime. From these results, we conclude that it is advantageous to use Bloch domain walls and perpendicular configuration in studies requiring higher domain wall density such as domain wall magneto-resistance, domain wall switching etc. in spintronics43 ; 44 . This study may find an importance in the field of domain wall superconductivity and domain wall induced triplet superconductivity51 ; 13 ; 42 .

Acknowledgements.

We acknowledge the funding from National Institute of Science Education and Research(NISER), DST-Nanomission (SR/NM/NS-1183/2013) and DST SERB (EMR/2016/005518) of Govt. of India.