France



Investigation of a creep of domain walls (DWs) in a combination of the in-plane and out-of-plane magnetic fields may give useful information about the Dzyaloshinskii-Moriya interaction in magnetic films. The first approach of the definition of the DMI field implied simple searching for the minimum in the dependence of the velocities of the DWs on the in-plane magnetic field, &lt;i&gt;v&lt;/i&gt;(&lt;i&gt;H&lt;/i&gt;&lt;sub&gt;x&lt;/sub&gt;) [1]. It was considered that when the in-plane magnetic field compensates the DMI field, the energy density of the DW is maximal and the velocity of the DW is minimal. However, papers were published in which the DMI fields measured by other methods like Brillouin light spectroscopy were very different from the magnetic fields of minima in the velocity curves &lt;i&gt;v&lt;/i&gt;(&lt;i&gt;H&lt;/i&gt;&lt;sub&gt;x&lt;/sub&gt;) [2]. To explain the shape of strongly asymmetric &lt;i&gt;v&lt;/i&gt;(&lt;i&gt;H&lt;/i&gt;&lt;sub&gt;x&lt;/sub&gt;) dependencies different extended models of the creep of DWs were inroduced [3, 4]. In this study, we investigate the DMI in the epitaxial Si(111)/Cu/Pd/Co(0.7 nm)/Pd system with the thickness of the bottom Pd layer varying from 0 to 3 nm. Since we measured non-zero DMI in the symmetric Pd/Co/Pd system in the previous paper [5], present study is aimed to understand the origin of the DMI in the symmetric Pd/Co/Pd system. The velocity curves &lt;i&gt;v&lt;/i&gt;(&lt;i&gt;H&lt;/i&gt;&lt;sub&gt;x&lt;/sub&gt;) were fitted by the extended dispersive stiffness model [4, 6] and the energy of DMI was derived from the fittings.&lt;br/&gt;We carried out the investigation of strains in the interfaces depending on the thickness of the Pd bottom layer, &lt;i&gt;d&lt;/i&gt;&lt;sub&gt;Pd&lt;/sub&gt;, by a reflection high energy electron diffraction during the process of fabrication of the samples. We found a large asymmetry of the strains between the bottom Pd/Co and top Co/Pd interfaces in the samples. The bottom Pd layer is initially largely strained when it grows on the Si/Cu(2 nm) surface. Stress is gradually relaxed by the introduction of misfit dislocations in the Pd layers. Co layers grown on the top of Pd underlayers are strained. Strains are partially relaxed during growth of the Co layers. Therefore, strains in the bottom and in the top of Co layers depend on the thickness of the Pd bottom layer.&lt;br/&gt;Magnetic hysteresis loops of the samples were measured by a vibrating sample magnetometer and the energy of a perpendicular magnetic anisotropy was defined in the series of samples. We investigated the creep of domain walls by means of a Kerr-microscope eqipped with in-plane electromagnet and home-made perpendicular coil. Measured velocity curves of the samples were strongly asymmetric. They may be fitted well only by using a variable velocity prefactor in the creep law, &lt;i&gt;v&lt;/i&gt; = &lt;i&gt;v&lt;/i&gt;&lt;sub&gt;0&lt;/sub&gt;exp(-Δ&lt;i&gt;E&lt;/i&gt;/kT), &lt;i&gt;v&lt;/i&gt;&lt;sub&gt;0&lt;/sub&gt; = &lt;i&gt;v&lt;/i&gt;&lt;sub&gt;0&lt;/sub&gt;*(1 + χ*cos(φ&lt;sub&gt;eq&lt;/sub&gt;)), where Δ&lt;i&gt;E &lt;/i&gt;is an energy barrier, χ* is the weight of a chiral damping effect, φ&lt;sub&gt;eq&lt;/sub&gt; is an azimuthal angle of the magnetization in the DW. Strictly speaking the nature of the dependence of the velocity prefactor on the in-plane magnetic field is still under debate. In some studies, this effect is named chirality-induced or additional asymmetry effect. The dependence of the chiral damping weight on the Pd bottom layer thickness is shown in Fig. 1. The absolute value of the chiral damping weight increases up to 0.95 at the thickness of the bottom Pd layer of 1 nm and oscillates near this value with further increasing of the thickness of Pd. This behavior of the chiral damping weight is correlated with the dependence of the asymmetry of the strains between the bottom and top interfaces on the thickness of the Pd bottom layer. Therefore, we assume that the possible origin of the strong asymmetry of the velocity curves may be related with strains gradient in the Co layers. The energy of the effective DMI, &lt;i&gt;D&lt;/i&gt;&lt;sub&gt;eff&lt;/sub&gt;, as a function of the thickness of the bottom Pd layer is outlined in Fig. 2. The negative energy of effective DMI slightly increases when &lt;i&gt;d&lt;/i&gt;&lt;sub&gt;Pd&lt;/sub&gt; = 0.2 nm. It may be explained by a structural inhomogeneity of the bottom interface and possible intermixing of Pd and Cu. With further increasing of &lt;i&gt;d&lt;/i&gt;&lt;sub&gt;Pd&lt;/sub&gt;, the magnitude of |&lt;i&gt;D&lt;/i&gt;&lt;sub&gt;eff&lt;/sub&gt;| decreases. &lt;i&gt;D&lt;/i&gt;&lt;sub&gt;eff&lt;/sub&gt; changes a sign at the thickness of Pd layer of 0.8 nm and increases with an increase in the thickness of the Pd bottom layer. The overall dependence of &lt;i&gt;D&lt;/i&gt;&lt;sub&gt;eff&lt;/sub&gt; on the &lt;i&gt;d&lt;/i&gt;&lt;sub&gt;Pd&lt;/sub&gt; may be explained by the dependence of strains in the interfaces on the thickness of Pd bottom layer. A kink in the &lt;i&gt;D&lt;/i&gt;&lt;sub&gt;eff&lt;/sub&gt;(&lt;i&gt;d&lt;/i&gt;&lt;sub&gt;Pd&lt;/sub&gt;) dependence correlates well with the kink in the dependence of the elastic strains on the &lt;i&gt;d&lt;/i&gt;&lt;sub&gt;Pd&lt;/sub&gt;. For the first sight the change in the sign of &lt;i&gt;D&lt;/i&gt;&lt;sub&gt;eff&lt;/sub&gt; may be easily explained by the growing influence of the contribution to the DMI from the bottom Pd/Co interface. However, we prepared asymmetric Pd(3 nm)/Co(0.7 nm)/Cu sample with the same strains distribution as in the Pd(3 nm)/Co(0.7 nm)/Pd sample but with near zero DMI. Therefore, the non-zero DMI in the symmetric Pd(3 nm)/Co/Pd(3 nm) sample is caused mainly by the contribution to the DMI from the top Co/Pd interface. We may conclude that unequal elastic strains in the bottom Pd/Co and top Co/Pd interfaces cause non-zero DMI in symmetric Pd/Co/Pd system.&lt;br/&gt;Authors acknowledge the financial support of the Russian Ministry of Science and Higher Education under the state task (0657-2020-0013), the Grant program of the Russian President (Grant No. MK-1384.2021.1.2) and Russian Foundation for Basic Research according to the research project No.19-02-00530.

**References:**

1. S.-G. Je, D.-H. Kim and S.-C. You, Phys. Rev. B., Vol. 88, p. 214401 (2013).&lt;br/&gt;2. R. Soucaille, M. Belmeguenai, J. Torrejon, Phys. Rev. B., Vol. 94, p. 104431 (2016).&lt;br/&gt;3. J. P. Pellegren, D. Lau and V. Sokalski, Phys. Rev. Lett., Vol. 119, p. 027203 (2017).&lt;br/&gt;4. D.M.F. Hartmann, R. A. Duine, M.J. Meijer, Phys. Rev. B., Vol. 100, p. 094417 (2019).&lt;br/&gt;5. A.V. Davydenko, A.G. Kozlov, A.G. Kolesnikov, Phys. Rev. B., Vol. 99, p. 014433 (2019).&lt;br/&gt;6. D. Lau, J. P. Pellegren, H. T. Nembach, Phys. Rev. B., Vol. 98, p. 184410 (2018).

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/616d83d450af5977728230e236ec3566.jpg)
Fig. 1. The Pd thickness dependence of the chiral damping weight in the Si(111)/Cu/Pd(0–3 nm)/Co(0.7 nm)/Pd(3 nm) samples.

![](https://assets.underline.io/uploads/markdown_image/1/image/5a613be85fa0118433cb6bb5ec4ab78d.jpg)
Fig. 2. The Pd thickness dependence of the energy of the effective DMI in the Pd/Co/Pd films.

INTERMAG 2021

 Creep of domain walls in epitaxial Pd/Co/Pd(111) trilayers

The 15th Joint MMM-INTERMAG Conference (2022 Joint) is sponsored jointly by AIP Publishing and the IEEE Magnetics Society. Members of the international scientific and engineering communities interested in recent developments in fundamental and applied magnetism are invited to attend and contribute to the technical sessions. The technical program will include invited and contributed papers in oral and poster sessions, invited symposia, a plenary session, and an evening session, with about 1500 presentations overall. This Conference provides an outstanding opportunity for worldwide participants to meet their colleagues and collaborators and discuss developments in all areas of magnetism research.

The 2022 Joint Conference will offer both in-person and prerecorded on-demand content.

For tickets to the event, please register at the following link: https://magnetism.org/register

Please register for the Joint MMM-Intermag conference to join the event.

2022 Joint MMM-INTERMAG

This Conference provides an outstanding opportunity for worldwide participants to meet their colleagues and collaborators and discuss developments in all areas of magnetism research.

Investigation of a creep of domain walls (DWs) in a combination of the in-plane and out-of-plane magnetic fields may give useful information about the Dzyaloshinskii-Moriya interaction in magnetic films. The first approach of the definition of the DMI field implied simple searching for the minimum in the dependence of the velocities of the DWs on the in-plane magnetic field, v(Hx) [1]. It was considered that when the in-plane magnetic field compensates the DMI field, the energy density of the DW is maximal and the velocity of the DW is minimal. However, papers were published in which the DMI fields measured by other methods like Brillouin light spectroscopy were very different from the magnetic fields of minima in the velocity curves v(Hx) [2]. To explain the shape of strongly asymmetric v(Hx) dependencies different extended models of the creep of DWs were inroduced [3, 4]. In this study, we investigate the DMI in the epitaxial Si(111)/Cu/Pd/Co(0.7 nm)/Pd system with the thickness of the bottom Pd layer varying from 0 to 3 nm. Since we measured non-zero DMI in the symmetric Pd/Co/Pd system in the previous paper [5], present study is aimed to understand the origin of the DMI in the symmetric Pd/Co/Pd system. The velocity curves v(Hx) were fitted by the extended dispersive stiffness model [4, 6] and the energy of DMI was derived from the fittings. We carried out the investigation of strains in the interfaces depending on the thickness of the Pd bottom layer, dPd, by a reflection high energy electron diffraction during the process of fabrication of the samples. We found a large asymmetry of the strains between the bottom Pd/Co and top Co/Pd interfaces in the samples. The bottom Pd layer is initially largely strained when it grows on the Si/Cu(2 nm) surface. Stress is gradually relaxed by the introduction of misfit dislocations in the Pd layers. Co layers grown on the top of Pd underlayers are strained. Strains are partially relaxed during growth of the Co layers. Therefore, strains in the bottom and in the top of Co layers depend on the thickness of the Pd bottom layer. Magnetic hysteresis loops of the samples were measured by a vibrating sample magnetometer and the energy of a perpendicular magnetic anisotropy was defined in the series of samples. We investigated the creep of domain walls by means of a Kerr-microscope eqipped with in-plane electromagnet and home-made perpendicular coil. Measured velocity curves of the samples were strongly asymmetric. They may be fitted well only by using a variable velocity prefactor in the creep law, v = v0exp(-ΔE/kT), v0 = v0*(1 + χ*cos(φeq)), where ΔE is an energy barrier, χ* is the weight of a chiral damping effect, φeq is an azimuthal angle of the magnetization in the DW. Strictly speaking the nature of the dependence of the velocity prefactor on the in-plane magnetic field is still under debate. In some studies, this effect is named chirality-induced or additional asymmetry effect. The dependence of the chiral damping weight on the Pd bottom layer thickness is shown in Fig. 1. The absolute value of the chiral damping weight increases up to 0.95 at the thickness of the bottom Pd layer of 1 nm and oscillates near this value with further increasing of the thickness of Pd. This behavior of the chiral damping weight is correlated with the dependence of the asymmetry of the strains between the bottom and top interfaces on the thickness of the Pd bottom layer. Therefore, we assume that the possible origin of the strong asymmetry of the velocity curves may be related with strains gradient in the Co layers. The energy of the effective DMI, Deff, as a function of the thickness of the bottom Pd layer is outlined in Fig. 2. The negative energy of effective DMI slightly increases when dPd = 0.2 nm. It may be explained by a structural inhomogeneity of the bottom interface and possible intermixing of Pd and Cu. With further increasing of dPd, the magnitude of |Deff| decreases. Deff changes a sign at the thickness of Pd layer of 0.8 nm and increases with an increase in the thickness of the Pd bottom layer. The overall dependence of Deff on the dPd may be explained by the dependence of strains in the interfaces on the thickness of Pd bottom layer. A kink in the Deff(dPd) dependence correlates well with the kink in the dependence of the elastic strains on the dPd. For the first sight the change in the sign of Deff may be easily explained by the growing influence of the contribution to the DMI from the bottom Pd/Co interface. However, we prepared asymmetric Pd(3 nm)/Co(0.7 nm)/Cu sample with the same strains distribution as in the Pd(3 nm)/Co(0.7 nm)/Pd sample but with near zero DMI. Therefore, the non-zero DMI in the symmetric Pd(3 nm)/Co/Pd(3 nm) sample is caused mainly by the contribution to the DMI from the top Co/Pd interface. We may conclude that unequal elastic strains in the bottom Pd/Co and top Co/Pd interfaces cause non-zero DMI in symmetric Pd/Co/Pd system. Authors acknowledge the financial support of the Russian Ministry of Science and Higher Education under the state task (0657-2020-0013), the Grant program of the Russian President (Grant No. MK-1384.2021.1.2) and Russian Foundation for Basic Research according to the research project No.19-02-00530.

**References:**

1. S.-G. Je, D.-H. Kim and S.-C. You, Phys. Rev. B., Vol. 88, p. 214401 (2013). 2. R. Soucaille, M. Belmeguenai, J. Torrejon, Phys. Rev. B., Vol. 94, p. 104431 (2016). 3. J. P. Pellegren, D. Lau and V. Sokalski, Phys. Rev. Lett., Vol. 119, p. 027203 (2017). 4. D.M.F. Hartmann, R. A. Duine, M.J. Meijer, Phys. Rev. B., Vol. 100, p. 094417 (2019). 5. A.V. Davydenko, A.G. Kozlov, A.G. Kolesnikov, Phys. Rev. B., Vol. 99, p. 014433 (2019). 6. D. Lau, J. P. Pellegren, H. T. Nembach, Phys. Rev. B., Vol. 98, p. 184410 (2018).

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/616d83d450af5977728230e236ec3566.jpg)
Fig. 1. The Pd thickness dependence of the chiral damping weight in the Si(111)/Cu/Pd(0–3 nm)/Co(0.7 nm)/Pd(3 nm) samples.

![](https://assets.underline.io/uploads/markdown_image/1/image/5a613be85fa0118433cb6bb5ec4ab78d.jpg)
Fig. 2. The Pd thickness dependence of the energy of the effective DMI in the Pd/Co/Pd films.

technical paper

INTERMAG 2021 Welcome Video



	
<figure class="video_container">
 <iframe src="https://screencast-o-matic.com/player/crflqFVnUpZ?v=6&ff=1&title=0&controls=1" width=640 height=360 frameborder="0" allowfullscreen="true"> </iframe>
</figure>

Want to buy a ticket? 

Register for INTERMAG 2021 here

[![](https://assets.underline.io/uploads/markdown_image/1/image/17ce76e7960db633258024eef8d6da7f.png)](https://intermag.org/register)

Got your ticket, but still can’t access the event?

Contact us and our team will look into it.

Artificial spin ice (ASI) consists of a lithographically manufactured two-dimensional array of ferromagnetic nanoislands with a strong shape anisotropy resulting in single-domains that behave like giant Ising spins. In natural spin ices, such as, for example, the rare earth pyrochlore Ho2 Ti2 O7, the local ordering of magnetic moments is particularly difficult to measure [1]. Despite the extraordinary advances in imaging techniques in the last decade, which allows us now to access the three-dimensional magnetic nanostructure of micrometric bulk samples [2], we still need to work together with simplified models to understand more about the underlying physics that governs these highly complex systems. In that direction, artificial frustrated magnetic systems allow us to test and reproduce theoretical models as well as to understand more about their three-dimensional analogs. But even more importantly, it allows us to design a system rather than discover it [3]. Here we study an squared artificial spin ice, with dipolar interactions, where there is a height separation h between the magnetic islands which are oriented perpendicularly [3, 4] (see Fig. 1). In a tetrahedrical lattice, such as the one found in natural spin ices, the distance between any pair of spins (belonging to the same tetrahedron) is the same, and therefore, the dipolar energy has the same value for any pair. On the contrary, in the square ASI, the distance between collinear spins is different from the distance between perpendicular spins. Hence, the need to add this height parameter. In this way, this system is halfway between three-dimensional (natural) and two-dimensional (artificial) spin ices. The effect of disorder is still an open question and a field to study in ASI [1,4]. Budrikis et al (2012) study from the point of view of theory, simulations and experiments the influence of disorder on the response of the system to an external field [5]. To do this, in the simulations, they propose that each spin had an internal (coercive) field given by a Gaussian distribution. With this model, they manage to estimate the strength of disorder in a sample. There are several works that quantify the disorder, what is new in Budrikis’ work is that they also study how disorder affects the dynamics of ASI. Chern et al (2014) take this same idea to model disorder and study the avalanches and critical behavior in square and kagome ASI when a magnetic field is applied in-plane [6]. This system shows a phase transition out-of-equilibrium induced by disorder strength. Reichhardt et al (2015) show that the avalanche distributions of this process follow a power law [7]. Another way to include disorder in the system is to disconnect the islands by eliminating a given percentage of them at random. Greenberg et al (2018) show through simulations that this system changes its thermal behavior as the percentage of holes increases [8]. These results coincide with what was observed experimentally in diluted spin ice, once again showing that the study of ASI allows modeling and better understanding of the behavior of spin ice. Our aim is to study, through simulations and an energetic analysis, the effect of a geometrical quenched disorder by design on an square ASI. To do this, we propose to include disorder in the length of the magnetic islands. In this way, the strength of disorder could be easily controlled experimentally. As the system behavior depends on its geometrical parameters (a, h and d in Fig. 1), we focus on studying it in the proximity of the ice regime, where Type I and II vertices are equally likely (see Fig. 2). This regime is quite difficult to thermalize both in experiments and simulations. In this work, we show how length disorder affect the antiferromagnetic (all Type I vertices) and locally ferromagnetic (all Type II vertices) ordering, by inducing the system, in the case of weak disorder, to intermediate or mix states. An important result is that this means that disorder strength can be chosen and tuned to thermalize the ice regime. Moreover, in the case of strong disorder, ferromagnetic vertices prevail regardless of whether the mean length of the islands corresponds to an antiferromagnetic ordering. The energetic analysis let us understand the intermediate regimes that we found and also why disorder does not permit defects (Type III and IV vertices) at low temperatures in this model, contrary to expectations. Also, with this analysis, we show in detail how the ferromagnetic–antiferromagnetic transition with h is rounded by quenched disorder. Finally, we found that disorder in the geometry causes a slowdown in the dynamics of the system which increases with disorder.

**References:**

[1] C. Nisoli, R. Moessner and P. Schiffer, Rev. Mod. Phys., Vol. 85, p. 1473 (2013) [2] C. Donnelly, M. Guizar-Sicairos, V. Scagnoli et al, Nature, Vol. 547, p. 328–331 (2017) [3] R. F. Wang, C. Nisoli, R. S. Freitas et al, Nature, Vol. 439, p. 303 (2006) [4] G. Möller and R. Moessner, Phys. Rev. Lett., Vol. 96, p. 237202 (2006) [5] Z. Budrikis, J. P. Morgan, J. Akerman et al, Phys. Rev. Lett., Vol. 109 p. 037203 (2012) [6] G. W. Chern, C. Reichhardt and C. Nisoli, Appl. Phys. Lett., Vol. 104, p. 013101 (2014) [7] C. J. O. Reichhardt, G. W. Chern, A. Libál and C. Reichhardt, J. Appl. Phys., Vol. 117, p. 172612 (2015) [8] N. Greenberg and A. Kunz, AIP Adv., Vol. 8, p. 055711 (2018)

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/40c1929d56323d1f53e49ae6ec1ca273.jpg)
Fig. 1: Square artificial spin ice array with dipolar interactions: a is the lattice spacing and d is the island length; in this model, the width is considered negligible. There is a height separation h between the position of islands which are oriented in perpendicular directions; this means that dark-gray shaded islands are located at height z = h while light-gray shaded islands are located at height z = 0.

![](https://assets.underline.io/uploads/markdown_image/1/image/403dd6ed545da510cecae93a68da98cd.jpg)
Fig. 2: The point where four islands concur is called a vertex. There are 16 possible configurations for each vertex of an squared ASI, according to the orientation of its spins. Furthermore, these can be classified into four different topological groups. Type I and II satisfy the so called ice rule (two spins point inwards the vertex and two point outwards), while Type III and IV do not. Type II and III vertices have a non-trivial net magnetization.

 Disorder by Design in an Artificial Spin Ice with Dipolar Interactions: An Energetic Analysis

There have been reports for different systems that there is an asymmetry in the saturation magnetisation in exchange bias systems [1, 2]. This report contains the first comprehensive study of these effects in the CoFe/IrMn system which is used in the read head of almost all HDDs [3]. The magnetic behaviour of exchange bias systems has two quasi-independent contributions [3]. The first contribution derives from the ferromagnetic (F) layer which in device applications is in the range 2 to 5 nm. This thickness produces the maximum loop shift when the structure is field cooled. However, it has been shown that there can be a further significant contribution to the magnetisation from spins lying at the interface which can have their origin either from the F layer or the antiferromagnetic (AF) layer. It has been shown that ordering of interface spins increases the loop shift by increasing the coupling between the F and AF layers and that doping the interface in the CoFe/IrMn system with atomic layers of the high-moment Mn can increase the spin ordering significantly [3, 4]. In this work the asymmetry in the saturation moment (ms) and its temperature (T) dependence to &lt;4.2 K have been measured for sputtered polycrystalline multilayer Si/Ta (2)/Ru (2)/IrMn (10)/Mn (dMn)/CoFe (5)/Ru (5) (thicknesses in nm) thin films where 0≤ dMn ≤0.2 nm in a LakeShore 8600 Series VSM. The York Model of Exchange Bias was used to separate and interpret the effect of the F/AF interface [3]. Low-temperature measurements were undertaken to ensure that the bulk of the AF grains remained magnetically frozen. Thus there was no reorientation of the magnetic moments in the bulk of the AF grains in subsequent measurements. The magnetisation of an exchange bias system undergoes an athermal training effect whereby the first and second magnetisation loops exhibit different loop shifts after the initial field cooling, even when the temperature of measurements is such that the moment of the bulk of the AF grain does not vary. This effect on the coercivity, which is defined as half loop width, has been shown to originate in the interfacial spins [5]. Figure 1 shows the measured hysteresis loop, the structure of which is shown in the insert, at 5.5 and 300 K. For the system measured we established by separate measurements (not shown) that the loop shift for the system shown was thermally stable at T = 300 K by holding the magnetisation at reverse saturation for a period of 30 mins after which time the hysteresis loop show at 300 K reproduced. In figure 1 the difference between the loops measured was due to changes in the orientation and ordering of interfacial spin clusters. The increase in the measured loop shift comes from ordering of spins at the interface. Furthermore, the first loop training effect is very small at T = 300 K but becomes significant at T = 5.5 K. This shows that the training effect is caused by interfacial spins, as is the coercivity. There is an asymmetry in the magnetisation. This data was obtained for the same sample without it being removed from the magnetometer. A significant increase in ms occurs at 5.5 K, which is greater at negative saturation than is the case in a positive field. We designate this difference in ms in positive (ms+) and negative (ms–) fields as Δms = |ms+|–|ms–|. (1) Figure 2 shows the temperature dependence of Δms for samples where there is a variation in the thickness of the Mn interface doping layer (dMn). It has been found that the optimum loop shift for this structure is affected by dMn and has been found to be a layer of subatomic thickness of the order of 1/6 of an atom. The use of layers of an atomic thickness or greater were found to result a reduction in Δms [4]. Figure 2 shows Δms for the first (i) and second (ii) loops measured immediately after the field annealing process. There is a change in the interfacial spin order which gives rise to the interface training effect but also effects the change in Δms. For all films figure 2 shows that the increase in Δms is small at T &gt; 100 K. We attribute this to the interface spins not being ferromagnetically ordered and hence not responding to the modest fields used to measure the loops. Below 100 K the change in Δms with T becomes more rapid as it appears these spins become F ordered, probably in a spin cluster structure similar to a spin glass [6]. It appears that at &lt;6 K Δms goes through a peak for all samples which is consistent with a spin cluster structure. It would be expected that dMn would not be optimum for a layer of subatomic thickness. It is most likely that the Mn layer is making a compositional adjustment at the interface. In studies of a similar system by Baltz et al. using 3-D atom probe tomography on an exchange bias system containing Co and Mn significant diffusion was found to occur [7]. CoMn alloys have complex properties existing in both F, AF and superparamagnetic phases dependent on composition and temperature but no high resolution phase diagram has been determined [8]. In the production of an exchange bias system the field annealing requires up to 600 K which promotes interfacial diffusion [3, 7]. We conclude that the role of the Mn interface layer is to replace Mn deficiencies in the bulk of the AF.

**References:**

[1] U. Nowak et al., Phys. Rev. B, Vol. 66(1), p. 014430 (2002) [2] J. Keller et al., Phys. Rev. B, Vol. 66(1), p. 014431 (2002) [3] K. O'Grady et al., J. Magn. Magn. Mater., Vol. 322(8), p. 883-899 (2010) [4] R. Carpenter et al., IEEE Trans. Magn., Vol. 48(11), p. 4351–4354 (2014) [5] R. Carpenter et al., J. Appl. Phys., Vol. 115(17), p. 17D715 (2014) [6] J. L. Tholence and R. Tournier, Phys. B+C, Vol. 86-88, p. 873-874 (1977) [7] F. Letellier, et al., J. Appl. Phys., Vol. 116, p. 203906 (2014) [8] A. Z. Men’shikov, et al., Sov. Phys. JETP, Vol. 62, p. 734 (1985)

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/c4880f4f4f141c9f2e76b5ec7dfed9c9.jpg)
Figure 1: The first (i) and second (ii) hysteresis loops at 5.5 and 300 K for a film with structure shown inset.

![](https://assets.underline.io/uploads/markdown_image/1/image/b484480c5ac00e6406cf86b341a6f8a8.jpg)
Figure 2: The temperature dependence of the magnetisation offset (Δms) measured at ± 5 kOe shown for the first (i) and second (ii) magnetisation reversal for the thin films shown inset where 0≤ dMn ≤0.2 nm.

 Magnetisation Asymmetry in Exchange Bias Systems

Here we present an X-ray, magnetometry, and electron microscopy study of novel 40 nm FeCr spinel oxide nanoparticles (NPs) containing ~4 nm FeO clusters occupying up to 20% of the volume. The unique structure provides a large effective interface area between the antiferromagnetic (AFM) FeO clusters and the ferrimagnetic (FiM) spinel structure, leading to one of the largest exchange bias (EB) observed in FeO-based NP systems. Nanoparticles composed of Wüstite FeO and Magnetite Fe3O4 have been studied extensively and exhibit a wide range of magnetic properties [1-3]. Both oxides share the same fcc oxygen lattice and differ only in the occupancy and valence of the Fe cations in the tetrahedral (A) and octahedral (B) interstitial sites. Wüstite has a rock salt (RS) structure with Fe2+ occupying B-sites in the fcc O-lattice. In bulk, FeO is not stoichiometric and thermodynamically stable under ambient conditions, which results in a defect FexO RS structure. The Fe2+ ions in the RS structure can diffuse and oxidize into Fe3+ at the surface, forming the inverse spinel structure of Magnetite, where Fe3+ occupies the A-lattice while the B-lattice is equally shared by the Fe2+ and Fe3+ cations. At the nanoscale, the transformation of Wüstite into Magnetite leads to FexO/Fe3O4 core/shell (CS) NPs where exchange coupling between the AFM core and the FiM shell leads to the emergence of EB [1-6], which manifests as a horizontal shift of the hysteresis and coercivity increase. Here, NPs with an average diameter of 40 nm are generated using an aerosol technique based on spark ablation [7-9] where the material is evaporated from stainless steel grade 430 electrodes containing 97.4% FeCr with a 4:1 ratio of Fe:Cr [9]. The evaporated material is transported where adiabatic expansion and mixing with the carrier gas cool down the vapor, which condensates into sub 10 nm primary particles that form larger agglomerates as they collide. The agglomerates are transported through a furnace where they are compacted at 1200 oC before being size-selected and deposited onto a substrate using an electric field. The structural characterization was performed by high-resolution electron microscopy (HRTEM) and the compositional analysis was carried out with energy dispersive X-ray spectroscopy (EDXS) in scanning transmission electron microscopy (STEM) mode, see Figure 1 (a and b), along with X-ray photoelectron spectroscopy (XPS) performed at the Species beamline, MAX IV [10]. The NPs appear to have a single crystalline phase with a spinel oxide structure, and homogeneously distributed Fe and Cr within the NP with the expected Fe:Cr ratio of 4:1. However, X-ray diffraction (XRD) acquired at the 11-ID-B beamline of the Advanced Photon Source [11] reveals the co-existence of up to 20% RS crystallites with a size (~4 nm) about one order of magnitude smaller than the spinel phase (~40 nm). The significant volume fraction and the small size of the RS phase rule out a CS structure and point to FexO RS clusters embedded in the spinel structure as schematically illustrated in Fig. 1 c). In stainless steel, Cr prevents the corrosion of metallic Fe by forming an oxide that blocks further diffusion and oxidation of Fe at the surface. We propose that Cr plays a similar role here, but now preventing the full transformation from FexO → Fe3O4, both in the primary particles and during compaction. X-ray magnetic circular dichroism (XMCD) data, acquired at the X-treme beamline of the Swiss Light Source [12] shows the FM interaction between the Fe2+/Fe3+ and Cr3+ ions in the B-lattice, and an AFM coupling to the Fe3+ ions in the A-lattice. Differences observed in XMCD spectra recorded at 2 K and 300 K at the Fe L3-edge indicate an AFM to paramagnetic transition, further supporting the presence of FexO in the NPs. Element and site specific m(H) curves were recorded and the differences between Fe and Cr will be discussed, along with a more precise determination of the individual ions’ magnetic contribution by means of multiplet simulations. Temperature-dependent magnetization measurements also indicate the presence of FexO, where a rapid increase in magnetization is observed around the Neél temperature of FeO (TN ~ 200 K - Fig.2 left-panel). The magnetic moment as a function of applied magnetic field at a temperature of 2 K after cooling in a field of 1 T is shown in the right-panel of Figure 2. As can be seen, the hysteresis is significantly shifted with respect to the field axis. The observed shift of µ0H = 460 mT is among the highest values ever reported for Fe-oxide NPs, and significantly larger than the typical values for FexO/Fe3O4 NPs [13] (a record high of 860 mT was reported by Lottini et al [14] after Co doping). Furthermore, a 38% increase in the coercivity at 2 K (580 mT) after cooling in a magnetic field is also observed. That the EB is the result of interactions with the AFM FexO clusters is also corroborated by the temperature dependence of the HEB which vanishes above the FeO TN, immediately followed by a rapid coercivity decrease. The large EB is attributed to the unique structure in which the embedded AFM clusters have an increased effective AFM-FiM interface area compared to AFM/FiM core/shell structures.

**References:**

[1] Sun, Xiaolian, et al. "Tuning exchange bias in core/shell FeO/Fe3O4 nanoparticles." Nano letters 12.1 (2012): 246-251. [2] Estrader, Marta, et al. "Origin of the large dispersion of magnetic properties in nanostructured oxides: Fe x O/Fe 3 O 4 nanoparticles as a case study." Nanoscale 7.7 (2015): 3002-3015. [3] Wetterskog, Erik, et al. "Anomalous magnetic properties of nanoparticles arising from defect structures: topotaxial oxidation of Fe1–x O| Fe3− δO4 core| shell nanocubes to single-phase particles." ACS nano 7.8 (2013): 7132-7144. [4] Iglesias, Oscar, Amilcar Labarta, and Xavier Batlle. "Exchange bias phenomenology and models of core/shell nanoparticles." Journal of nanoscience and nanotechnology 8.6 (2008): 2761-2780. [5] Meiklejohn, William H., and Charles P. Bean. "New magnetic anisotropy." Physical review 102.5 (1956): 1413. [6] Nogués, Josep, et al. "Exchange bias in nanostructures." Physics reports 422.3 (2005): 65-117. [7] R. Hallberg, L. Ludvigsson, C. Preger, B. Meuller, K. Dick, and M. Messing, "Hydrogen-assisted spark discharge generated metal nanoparticles to prevent oxide formation," Aerosol Science and Technology, vol. 52, no. 3, pp. 347-358, 2018. [8] B.O. Meuller, M.E. Messing, D.L.J. Engberg, et al. ”Review of spark discharge generators for production of nanoparticle aerosols” Aerosol Sci. Technol., 46, 1256, 2012. [9] C. Preger et al., "Controlled Oxidation and Self-Passivation of Bimetallic Magnetic FeCr and FeMn Aerosol Nanoparticles", The Journal of Physical Chemistry C, vol. 123, no. 26, pp. 16083-16090, 2019. [10] S. Urpelainen, C. Såthe, W. Grizolli, M. Agåker, A. R. Head, M. Andersson, S.-W. Huang, B. N. Jensen, E. Wallén, H. Tarawneh, et al., "The species beamline at the max iv laboratory: a facility for soft x-ray rixs and apxps", Journal of synchrotron radiation, vol. 24, no. 1, pp. 344-353, 2017. [11] U. Rütt, M. Beno, J. Strempfer, G. Jennings, C. Kurtz, and P. Montano, "Diffractometer for high energy x-rays at the APS", Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, vol. 467, pp. 1026-1029, 2001. [12] C. Piamonteze, U. Flechsig, S. Rusponi, J. Dreiser, J. Heidler, M. Schmidt, R. Wetter, M. Calvi, T. Schmidt, H. Pruchova, et al., “X-treme beamline at sls: X-ray magnetic circular and linear dichroism at high field and low temperature", Journal of synchrotron radiation, vol. 19, no. 5, pp. 661-674, 2012. [13] Phan, Manh-Huong, et al. "Exchange bias effects in iron oxide-based nanoparticle systems." Nanomaterials 6.11 (2016): 221. [14] Lottini, E., et al. "Strongly exchange coupled core| shell nanoparticles with high magnetic anisotropy: a strategy toward rare-earth-free permanent magnets." Chemistry of Materials 28.12 (2016): 4214-4222.

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/d2fd8d2c49ec21852eb227573753590a.jpg)
Fig 1. a) left: HRTEM image of a NP; right: FFT indicating spinel; b) EDXS showing Fe, Cr and O distribution; c) Schematic model of a spherical spinel NP containing cubic FeO clusters (brown).

![](https://assets.underline.io/uploads/markdown_image/1/image/04541974969216210d0c73ac8ae0bc2a.jpg)
Fig 2. Left: temperature susceptibility in a 10 mT field in ZFW and FCC cases; right: 1 T FC m(H) showing the horizontal hysteresis shift.

 Large Exchange Bias in FeCr Spinel Oxide Nanoparticles with Embedded FexO Clusters

The research on hybrid systems combining dissimilar materials with artificial interfaces is among the major challenges of the nanotechnology during the last decades. Different functionalities may be engineered by combining dissimilar materials. To this regard, bilayers composed of a ferromagnetic (FM) layer interfaced with a transition-metal-oxide (TMO) layer can be considered as model hybrid systems V2O3 is an archetypical TMO system exhibiting a first-order metal insulator transition (MIT) when the temperature is decreased below TMIT= 170 K [1], from a paramagnetic metallic state to an antiferromagnetic (AFM) insulating phase, accompanied simultaneously by a structural phase transition [2]. We are provided with an additional degree of freedom in order to control the FM properties at the FM/TMO interface. Below TMIT, several interfacial exchange-coupling effects have already been found in similar FM/V2O3 bilayers [3], ranging from moderate to negligible exchange bias fields, enhanced giant coercive fields and unique dynamical effects. To understand these interesting effects it is important to investigate in detail the temperature dependence of the magnetic anisotropies of this system. Here we present a detailed angular and temperature dependent angle-resolved vector magnetometry study of the model FM/TMO system mentioned across the MIT. This provides fundamental insights on how to tailor and control the interfacial exchange effects. More specifically, we investigated the magnetic anisotropy landscape of Co/V2O3 bilayers, where the Co layer has a well-defined uniaxial magnetic anisotropy at room temperature (see Fig.1a1 y Fig.1b1). We have studied the effects using different magnetic anisotropy configurations by setting a magnetic orientation of the FM layer across the MIT transition. Different field cooling (FC) procedures strongly affect the magnetic configuration of the hybrid bilayer. Just after the FC, the low temperature (50 K) loops display strong interfacial exchange coupling effects (see Fig.1a2 y Fig.1b2), i.e., enhanced coercive field, clear exchange bias, and asymmetric magnetization reversal [4]. In general, depending on the FC procedure, different angle-dependent and temperature evolutions are found, which are ascribed to the modifications of the anisotropy configuration. Our findings demonstrate that we have an effective control of the interfacial exchange phenomena in the FM/TMO heterostructures, especially across the MIT, which enables the development of novel magnetic devices with specific functionalities. This research was partially supported by the Spanish Ministry of Economy and Competitiveness (MINECO) (grant numbers FIS2016-78591-C3-1-R, RTI2018-097895-B-C42, and SEV-2016-0686) and by the Regional Government of Madrid (through project NANOMAGCOST-CM). JMD ackowledges support from MINECO through FPI program (BES-2017-080617). Work at UCSD supported by the Office of Basic Energy Science, U.S. Department of Energy, BES-DMS funded by the Department of Energy’s Office of Basic Energy Science, DMR under grant DE FG02 87ER-45332. We thank C. Urban, and J. Trastoy for initial work and discussions on this problem.

**References:**

[1] Imada, M., Fujimori, A. and Tokura, Y. Rev. Mod. Phys. 70, 1039-1263 (1998). [2]De La Venta, J., Wang, S., Ramirez, J. G., and Schuller, I. K. Applied Physics Letters, 102(12), 122404 (2013). [3] E. Jiménez, J. Camarero, J. Sort, J. Nogués, A. Hoffmann, N. Mikuszeit, J. M. Garcia-Martín, B. Dieny, and R. Miranda. Phys. Rev. B. 80,014415 (2009). [4] J. Camarero, J. Sort, A. Hoffmann, J. M. Garciaa-Martín, B. Dieny,R. Miranda, and J. Nogués. Phys. Rev. Lett. 95, 057204 (2005)

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/84121644807d7ea9c1108a6365c3b517.jpg)
Representative v-MOKE loops of a V2O3 (100 nm)/Co (15 nm) heterostructure above (a1,b1) and below (a2,b2) the MIT transition using different field-cooling (FC) conditions: a) Negative-FC along the easy axis and b) Positive-FC around 50° off the easy axis. Note that after FC, the hysteresis field-shift (indicated with arrows) is opposite to the FC direction, while the reversal pathways are very different [notoriously in M⊥(H)].

 Tailoring interfacial phenomena in hybrid V2O3/Co bilayers

Introduction: Exchange Bias(EB) originates from the interfacial interaction between a ferromagnetic (FM) and an Antiferromagnetic (AFM) material, after cooling the sample in a magnetic field below the Neel temperature (TN) of the AFM [1]. The main feature of EB is training effect which is a gradual reduction of EB field when hysteresis loop (n) measured many times at constant temperature. The training effect is inversely dependence on FM layer thickness and interface structure, etc [2]. We have carried out the Kerr microscopy measurements to get direct information about magnetization irreversible of FM layers along with hysteresis loops. Experimental details: Thin films of Al(10nm)/Fe(30nm)/Fe57(10nm)/Fe(30nm)/Si substrate were deposited using Ion-beam sputtering technique. The films were implanted with 35 keV Fluorine ions. With this ion energy, the maximum energy has been deposited only at the centre of the Fe film to obtain FeF2 at the center, which is an antiferromagnet and its Neel temperature is 78 K. For magnetic characterization, the measurements were done using in-plane Magneto-optic Kerr effect microscope (MOKE) at 50 K after field cooling from room temperature. Results and discussions Fig. 1(a) shows the hysteresis loop as function of number of loops. It clearly shows the exchange bias and it becomes quite constant. Similar observation has also been showed for the sample irradiated with 5×1016 F-ions/cm2. Magnetic domain evolution during the magnetization irreversible and the training effect have been studied after F-ion implantation at 50 K. Kerr microscopy measurements were carried out at different temperatures after the field-cooling across the Neel temperature of the antiferromagnetic (AFM) FeF2 layer. It is observed that with the increasing anti-ferromagnetic (AFM) FeF2 layer thickness, the exchange bias increase and the training effect becomes weaker. Analysis of the temperature variation of the exchange bias field indicates the disorder decreases at the FM-AFM interface with the increasing F-Ion fluences. Higher fluence specimen indicates a behavior that is changing from non-equilibrium state to equilibrium state and behavior of training effect is suggesting step-like. For 1×1017 ions/cm2 sample, change in the exchange bias field after measuring the 6th cycle is 4.42% while in case of 5×1016ions/cm2 27.32%. Remarkably, for the largest fluences, the relative training becomes smaller than the one for the lower fluence. It indicates higher ion fluence EB system is suitable for device application.

**References:**

 [1] W. H. Meiklejohn and C. P. Bean, Phys. Rev. 102, 1413(1956). [2] M. Gruyters and D. Riegel, Phys. Rev. B 63, 052401 (2000).

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/a97c3e88eac11173ccbf237d54b463aa.jpg)
Figure 1: (a) The hysteresis loops as a function of number of loops (n) for the specimen 1×1017 ions/cm2 (b) Value of Exchange bias for both the films implanted with ion fluences 5×1016 ions/cm2 and 1×1017 ions/cm2 at 50 K.

 Observation of training effect in Fe thin film implanted with F+ ions

Magnetic tunnel junctions (MTJs) with perpendicular magnetic anisotropy (PMA) have advantages of high packing density due to their high thermal stability, efficient current-induced magnetic switching, and unrestricted cell aspect ratio.[1] Therefore, PMA-based MTJs (p-MTJs) have broad application prospects on high-density and non-volatile spin-transfer torque magnetic random access memories (STT-MRAMs) and other spintronics devices.[2-4] In particular, p-MTJs with CoFeB/MgO structures show high tunnel magnetoresistance ratio.[4] In the W/CoFeB/MgO stacks, the interfacial magnetic anisotropy (Ki) of 1.98 mJ/m2 is acquired after annealing at 350 °C.[5] To further improve Ki, it is an effective way to use MgO/CoFeB/X/CoFeB/MgO double-interface free layer structures, where the X represents the spacer material. [6] However, higher Ki is needed to scale down the MTJs devices and maintain high thermal stability.[7] In this work [8], we study the magnetic anisotropy of the double-interface free layer stacks with different annealing temperatures and different spacer materials. The samples of Ta(5)/MgO(2)/Co20Fe60B20(t)/X(0.4) /Co20Fe60B20(1.2)/MgO(2)/Ta(3) are deposited by sputtering system, where the numbers in parentheses indicate the thickness in nanometers, X including Ta, W, and Mo. The thickness of the bottom CoFeB layer t ranges from 0.4~3.2 nm. The films with Mo spacer layer exhibit the highest Ki of 4.06 mJ/m2 compared with the Ta- and W-based films. Furthermore, the Ki of Mo-based films remains at 2.93 mJ/m2 even after annealing at 500 °C for 1 h. Fig.1a and 1b show the structures with bottom CoFeB thickness of 0.8 nm and with spacer materials of Mo and W. As shown in the M-H (Magnetization M versus magnetic field H) curves, all the films show good PMA. Fig.1c and 1d show the corresponding high-resolution transmission electron microscopy (HR-TEM) and energy-dispersive X-ray spectroscopy (EDX) images. Saturation magnetization Ms and the total thickness of magnetic dead layer tDL can be extracted from the curves ( M/A versus tCoFeB curves), where A is the area of the sample and tCoFeB is the total thickness of CoFeB layers. The Ki can be extracted from the intercept of the Keff*teff versus teff curves, where Keff is effective magnetic anisotropy energy density and teff = tCoFeB- tDL. As shown in Fig.2a and 2b, Ki of 2.15±0.11 mJ/m2 and 2.90±0.34 mJ/m2 can be obtained for Ta- and W-based stacks, respectively. More importantly, the maximum of 4.06±0.14 mJ/m2 is acquired for the Mo-based structure after annealing at 350 °C as shown in Fig.2c. According to previous study, when the required thermal stability factor is 70, if the Ki reached 4 mJ/m2 , the device size can be reduced to 10 nm.[8] It can be shown in Fig.2d- 2f that the Ki decreases with the increase of annealing temperature. Note that the Ki reaches 2.93 mJ/m2 even after annealing at 500 °C for 1 h, which demonstrates the high thermal stability of this structure. Revealed by the HR-TEM and EDX images, the sharper interface, weaker interdiffusion, better interfacial uniformity and higher crystallinity can be seen in the Mo-based stack than the W-based stack. These results show that Mo plays an important role in improving the film quality. When the annealing temperature increases, the interfacial mixing intensifies, which leads to the decrease of Ki in Mo-based films. In summary, we study the dependence of magnetic anisotropy and film quality of the MgO/CoFeB/X/CoFeB/MgO double-interface free layer structures on the spacer material and the annealing temperature. We find that, the Mo-based structure exhibits the strongest PMA (4.06 mJ/m2). Furthermore, Ki in the Mo-based stack is still as strong as 2.93 mJ/m2 even after annealing at 500 °C. The HR-TEM and EDX images of W- and Mo-based stacks show that the high thermal stability can be attributed to the high film quality in the Mo-based films. Our work provides a practical solution to scale down MRAM device while maintaining a high thermal stability.

**References:**

[1] Ikeda, S., Miura, K., Ohno, H. Nature materials, 9(9), 721-724 (2010). [2] Wang, M., Cai, W., Zhao, W. Nature electronics, 1(11), 582-588 (2018). [3] Peng, S., Zhu, D., Zhao, W. Nature Electronics, 1-8 (2020). [4] Wang, M., Cai, W., Zhao, W. Nature communications, 9(1), 1-7 (2018). [5] An, G. G., Lee, J. B., Hong, J. P. Acta Materialia, 87, 259-265 (2015). [6] Hwang, C. S. Advanced Electronic Materials, 1(6), 1400056 (2015). [7] Peng, S., Kang, W., Zhao, W. IEEE Magnetics Letters, 8, 1-5 (2017). [8] Cheng, H., Chen, J., Zhao, W. Advanced Electronic Materials, 6(8), 2000271 (2020).

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/c74384cddd3ce3d115a89232da765fba.jpg)
Fig.1. (a,b) The schematic of Ta(5)/MgO(2)/Co20Fe60B20(0.8)/X(0.4) /Co20Fe60B20(1.2)/MgO(2)/Ta(3) with spacer material of (a) X=Mo and (b) X=W. (c,d) The corresponding HR-TEM images and EDX line profiles with spacer material of (c) X=Mo and (d) X=W. (e,f) The corresponding hysteresis loops with the in-plane (IP) field (red line) and out-of-plane (OP) field (black line) for different spacer materials of (e) X=Mo and (f) X=W. In the inset of (e) and (f), a M-H image under the small OP field with spacer material of (e) X=Mo and (f) X=W.

![](https://assets.underline.io/uploads/markdown_image/1/image/75ebcd2b300cee56aef34e4a13c0417a.jpg)
Fig.2. (a-c) The Ki fittings for MgO (2)/Co20Fe60B20(t )/X(0.4)/ Co20Fe60B20(1.2)/MgO(2) stacks under annealing at 350 °C with spacer layer X of (a) Ta, (b) W, and (c) Mo. (d-f) The Ki fittings for MgO(2)/Co20Fe60B20(t ) /Mo(0.4)/Co20Fe60B20(1.2)/MgO(2) stacks with the annealing temperature of (d) 400 °C, (e) 450 °C, and (f) 500 °C.

 Giant perpendicular magnetic anisotropy in Mo-based double-interface free layer structure for advanced magnetic tunnel junctions

Soft magnetic films consisting of CoFeB alloys are widely applied in magnetic sensor [1] and MRAM devices [2], but also used as cores for microfabricated transformers used to power specific subunits on microprocessors [3]. These applications typically require a well-defined in-plane uniaxial anisotropy axis which aligns the magnetic moments and permits operation modes where a field is applied either along the anisotropy axis to obtain a domain wall motion easy axis or a magnetization rotation governed magnetization process. Several implementations have been discussed in literature to obtain a well-defined uniaxial in-plane anisotropy axis and to tune the anisotropy. Apart from the geometry of a sensor element which can be chosen to obtain a shape anisotropy , oblique sputtering of seed layers for the successively deposited magnetic layer(s) [4], sputter deposition or post-deposition annealing in a magnetic field [5] have been employed. Here, we deposit CoFeB layers on Ta seed layers grown by oblique sputtering and compare the resulting anisotropies to those obtained by deposition of the magnetic layer in an applied field. Moreover, for the CoFeB layers on the oblique-sputtered seeds we compare the anisotropies obtained for different CoFeB-thicknesses to distinguish between a surface or interface induced anisotropy or a bulk anisotropy term. Specifically, we deposited Ta(2nm)/CoFeB(6nm)/Ta(Xnm)-seed films onto a Si-wafer covered by its native oxide using angles of 40°, 50° and 60° between the sample normal and the Ta source, and Ta-seed layer thicknesses from 2nm to 10nm. For an angle of 60° the Ta-seed layer was additionally deposited with a thickness of 15nm and 20nm. In addition, we deposited 4nm, 8nm, 16nm and 32nm CoFeB films on 4nm-thick Ta-seed layer deposited at an angle of 50°. Selected systems were deposited in fields of 4mT and 40mT, applied parallel to the substrate surface. VSM was then used to obtain the magnetization of the CoFeB films. Hard- and easy axis magnetization loops were measured with a Kerr Microscope to determine the in-plane anisotropy (see figure 1 & figure 2). We find that the in-plane anisotropy increases with the deposition angle and also the seed layer thickness. For a deposition angle of 60° the anisotropy increases from 4kJ/m3 to 25kJ/m3 with thicknesses from 2nm to 20nm. For smaller deposition angles, we measure a much smaller increase of in-plane anisotropy. Concerning the scientific question whether the anisotropy arises from a surface or a bulk effect, our analysis of CoFeB films with thicknesses from 4nm to 32nm revealed that the anisotropies stemming from the oblique-sputtered Ta-seed layers is well modeled by a surface anisotropy term. While with oblique sputtering, a wide range of in-plane anisotropies can be covered, the anisotropies remain much smaller when films are deposited in fields. However, it has already been demonstrated that with CoSm-alloys much larger anisotropies have been obtained [5]. To correlate the seed-layer morphology with the in-plane anisotropy, high resolution AFM imaging on the Ta-seed layer was performed. Our work shows that in-plane anisotropies in magnetically soft systems can be tuned over a wide range by oblique seed-layer depositions. However, the anisotropy arises from an interface effect and thus decays as 1/t with CoFeB film layers thickness "t". The anisotropy of CoFeB films deposited in an applied field does not depend on film thickness and remains rather small. We propose that alloying Sm into the CoFeB films may increase the field-induced anisotropy, while keeping the soft magnetic properties of the CoFeB films.

**References:**

[1] T. Ogasawara, M. Oogane and M. Al-Mahdawi; Sci. Rep. 9, 17018 (2019) [2] S. Zhanga, Y. Sua and X. Li; Appl. Phys. Lett. 114, 042401 (2019) [3] J. Michel et al. IEEE Transactions on Magnetics 55, 7, pp. 1-7, (2019) [4] R. D. McMichael, C. G. Lee and J. E. Bonevich; Journal of Applied Physics 88, 5296 (2000) [5] F. Magnus, R. Moubah and A. H. Roos; Appl. Phys. Lett. 102, 162402 (2013)

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/6f342ce0ad4cb0acb37eff2b59d57a63.jpg)
Hard- and easy axis magnetization loops measured with Kerr Microscope of deposited Ta(2nm)/CoFeB(6nm)/Ta(Xnm)-seed layer at an angle of 60°

![](https://assets.underline.io/uploads/markdown_image/1/image/4484fc484b9b980d8019b0d9f4b39d69.jpg)
In-plane anisotropy of deposited Ta(2nm)/CoFeB(6nm)/Ta(Xnm)-seed layer at an angle of 60°

 Tuning the inplane anisotropy of CoFeB films

In this work, we present the results of polar magneto-optical Kerr effect (PMOKE) magnetometry and microscopy study of Ta/CoFeB/MgO stacks with a prototypical perpendicular easy axis system in spintronics applications [1]. Multilayer films with a structure of Ta (5 nm)/(Co0.25Fe0.75)75B25 (d=1.24÷1.60 nm)/MgO (2 nm)/Ta (2 nm) were deposited by dc/rf magnetron sputtering on thermally-oxidized Si substrates and after were annealed at 300 oC for 1 h under a perpendicular magnetic field 0.4 T in vacuum [2,3]. The magnetization processes statics, magnetization reversal dynamics and domain structures evolution driven by magnetic field were studied as a function of CoFeB thickness d. Magnetization curves are measured for samples with various d applying PMOKE magnetometry. For d=1.24 to 1.36 nm square-shaped hysteresis loops with full remanence, indicating an out-of-plane magnetization easy axis, are observed. For d=1.40 nm the shape of magnetization curve corresponds to the transitional state (at the edge of out-of-plane and in-plane magnetization states). For d =1.44 nm the magnetization curves demonstrate no hysteresis, which indicates on the reversible magnetization rotation and in-plane magnetization. The values of magnetic anisotropy parameters were deduced from the fitting of hard axis hysteresis curves using macrospin approximation model. In the case of the samples without in-plane magnetization additional 0.05 T magnetic field was applied to prevent domain structure formation. In the case of square hysteresis loop (d=1.24÷1.36 nm) magnetization reversal dynamics was studied exploiting PMOKE microscopy imaging technique driven by sequential magnetic field pulses (duration 200 ms) with a constant amplitude. Magneto-optical images were acquired in zero field after each pulse. The derived normalized magnetization m can be obtained as a function of number of field pulses expressed as time t using image processing techniques [4]. The dependence of magnetization m for direct process (starting from full saturated state) on time t for the sample with d= 1.24 nm is shown in Fig. 1. The logarithm of time t1/2, time when the sample has magnetization m=0, depends linearly on amplitude of pulses, which indicates on thermal activation of jump-like domain wall motion [5] and allows to calculate Barkhausen length lB » 120 nm [6]. Magnetization reversal in the stacks with out-of-plane anisotropy (d = 1.24÷1.36 nm) goes through domain nucleation and domain wall propagation. In the case of the samples with d = 1.24, 1.28 nm, while increasing H few nucleation point appears initially in the field of view and later quickly expands covering all the area excluding narrow domain between neighboring expanding domains named as narrow stripe domains (NSDs). NSDs are stabilized due to magnetostatic repulsion between there opposite domain walls and finite value of coercivity [7]. The increase of thickness d lead to significant reduction of straights cuts of NSDs, and for d = 1.32, 1.36 nm (close to dSRT) NSDs network is converted to the more complex dendrite like structures. The evolution of NSDs network in external out-of-plane magnetic field was studied exploiting PMOKE microscopy imaging technique driven by sequential magnetic field pulses (duration 500 ms) with a growing amplitude. The digital image processing of acquired images yielded normalized magnetization and NSDs length in the field of view. The obtained dependences of normalized magnetization and NSDs length for the stack with d=1.24 nm are presented in Fig. 2 as the functions of the amplitude of applied field pulses. NSDs appear after beginning of nucleation process at 1.25 mT, quickly reach maximum at 1.45 mT (when normalized magnetization mL » 0.92), then gradually annihilate and disappear after 5.5 mT. So saturation field is few times bigger than coercivity due to the NSDs. In conclusion, we have preformed an experimental study of the static and dynamic magnetization processes in ultrathin FeCoB films depending on their thickness. We find that magnetization reversal is occurred via thermal activated domain wall motion and is influenced by NSDs. The dependence of NSDs network length on applied out-of-plane magnetic field was studied.

**References:**

[1] S. Ikeda, K. Miura, H. Yamamoto, K. Mizunuma, H. D. Gan, M. Endo, S. Kanai, J. Hayakawa, F. Matsukura and H. Ohno, A perpendicular-anisotropy CoFeB–MgO magnetic tunnel junction,, Nature Mater. 9, 721–724 (2010), 10.1038/nmat2804 [2] S. Fukami, T. Suzuki, Y. Nakatani, N. Ishiwata, M. Yamanouchi, S. Ikeda, S. Kasai, and H. Ohno, Current-induced domain wall motion in perpendicularly magnetized CoFeB nanowire, Appl. Phys. Lett. 98, 082504 (2011), 10.1063/1.3558917 [3] C. Burrowes, N. Vernier, J.-P. Adam, L. Herrera Diez, K. Garcia, I. Barisic, G. Agnus, S. Eimer, Joo-Von Kim, T. Devolder, A. Lamperti, R. Mantovan, B. Ockert, E. E Fullerton, and D. Ravelosona, Low depinning fields in Ta-CoFeB-MgO ultrathin films with perpendicular magnetic anisotropy, Appl. Phys. Lett. 103, 182401 (2013), 10.1063/1.4826439 [4] M.Kisielewski, Z.Kurant, M.Tekielak, W.Dobrogowski, A.Maziewski, A.Wawro, L.T.Baczewski, Magnetooptical micromagnetometry of ultrathin Co wedges, Physica Status Solidi (A), vol.196, No.1, 129-132 (2003), 10.1002/pssa.200306370. [5] J. Ferré, V. Grolier, P. Meyer, S. Lemerle, A. Maziewski, E. Stefanowicz, S. V. Tarasenko, V. V. Tarasenko, M. Kisielewski, and D. Renard, Magnetization-reversal processes in an ultrathin Co/Au film, Phys Rev B 55, 15092–102 (1997), 10.1103/PhysRevB.55.15092. [6] A. K. Dhiman, T. Dohi, W. Dobrogowski, Z. Kurant, I. Sveklo, S. Fukami, H. Ohno, A. Maziewski, Magnetization processes and magnetic domain structures in Ta/CoFeB/MgO stacks, accepted in JMMM. [7] N. Vernier, J.-P. Adam, S. Eimer, G. Agnus, T. Devolder, T. Hauet, B. Ocker, F. Garcia, and D. Ravelosona, Measurement of magnetization using domain compressibility in CoFeB films with perpendicular anisotropy, Appl. Phys. Lett. 104, 122404 (2014), 10.1063/1.4869482

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/b5d1c9989052d51f4a6fef7bbdf051ff.jpg)
Fig. 1 The dependence of normalized magnetization m (derived from domains area) on total duration of magnetic field pulses t for the stack with d= 1.24 nm. The amplitudes of magnetic field pulses in series are 0.75, 0.85 and 0.95 mT.

![](https://assets.underline.io/uploads/markdown_image/1/image/7b3b405252fcac9a2760f9498043027a.jpg)
Fig. 2 The dependence of the normalized magnetization m (left Y-axis) and total length of NSDs, extracted from the field of view 1.4 mm2 (right Y-axis) on amplitude of magnetic field pulses m0H (pulse duration 500 ms) for the stacks with d=1.24nm.

 Magnetization Processes and Magnetic Domain Structures in Ta/CoFeB/MgO Stacks

Magnetic properties of an ultra-thin CoFeB film, particularly at the CoFeB/MgO interface, are of critical importance for the development of magnetic tunnel junction based magnetoresistive random access memory (MRAM) such as spin transfer torque MRAM (STT-MRAM) and voltage-controlled MRAM (VC-MRAM), because these MRAM cells have a CoFeB free layer with a thickness of around 1 nm in direct contact with a MgO tunnel barrier layer. In STT-MRAM cells, the CoFeB free layer is magnetized perpendicularly to the film plane by taking advantage of interfacial perpendicular magnetic anisotropy (PMA) induced at the interface with MgO [1,2]. Larger interfacial PMA is advantageous for shrinking the cell size while maintaining a sufficiently large thermal stability factor of the free layer magnetization. In VC-MRAM cells, the interfacial PMA is not only used to perpendicularly magnetize the free layer, but also actively modulated by the application of electric fields to trigger the so-called precessional toggle switching [3]. Therefore, thorough understanding of the physical mechanism of interfacial PMA is indispensable for material/process engineering of these MRAM technologies. The source of the interfacial PMA is considered to be preferential occupation of the dxy electron orbit that lies in-plane due to broken symmetry at the interface and spin-orbit interaction [4,5], which revives the orbital angular moment of the d-electrons that is almost completely quenched in the bulk. The magnitude of the orbital angular moment can be evaluated by the deviation of spectroscopic g factor from 2, i.e., g=2+L/S, where S and L are the spin and orbital angular moments, respectively. It has been observed that the magnitude of interfacial PMA is strongly dependent on the annealing temperature, suggesting that the crystalline state affects the electronic structure at the CoFeB/MgO interface, thus changing the interfacial PMA. According to these considerations, the spectroscopic g factor is expected to increase at the CoFeB/MgO interface with an annealing temperature dependence similar to that of the interfacial PMA. To examine this notion, we studied static and dynamic magnetic properties of CoFeB/MgO film stacks as a function of CoFeB thickness and annealing temperature. The samples are Ta(5) / CoFeB (t) / MgO(2) / Ta(5) / Ru(5) film stacks sputter deposited on a thermally oxidized Si substrate (from bottom to top, number in parentheses is the thickness of each layer in nm), where t is the nominal CoFeB thickness ranging from 1.1 nm up to either 1.5 or 5 nm for different annealing temperatures. These films were left as deposited or annealed between 150°C and 350°C for 1 hour in vacuum. We performed vibrating sample magnetometry (VSM) measurement for static characterization, and broadband ferromagnetic resonance (FMR) measurement for dynamic characterization, respectively. The CoFeB films have a magnetically dead layer at the Ta/CoFeB interface due to Ta diffusion with thicknesses depending on the annealing temperature, and the saturation magnetization Ms also varies on annealing temperature due to different degrees of crystallization. From the static M-H loop obtained by VSM, Ms as a function of t was first plotted, from which the effective thickness teff (t minus dead layer thickness) and effective Ms of the CoFeB layer were estimated for each annealing temperature. Next, the total PMA energy was calculated by integrating the static M-H loop. The interfacial PMA energy was estimated from the linear regression on the plot of total PMA energy as a function of teff for each annealing temperature. Figure 1 shows the interfacial PMA energy as a function of annealing temperature, which becomes maximum at 250°C. In the dynamic magnetic characterization, the broadband FMR measurement performed under out-of-plane bias magnetic field HB observed a Kittel mode FMR spectrum. The FMR frequency showed a liner relation on HB, from the slope of which the spectroscopic g factor was determined. The value of g factor showed a significant increase as teff decreased. Figure 2 shows the increase of g factor at the CoFeB/MgO interface from the bluk value as a function of annealing temperature, which becomes maximum at 200°C, and exhibits a very similar dependence on the annealing temperature to the interfacial PMA. These results support the notion that the increase of g factor and interfacial PMA have the same physical origin, i.e., revival of the orbital angular moment due to broken symmetry at the CoFeB/MgO interface. In the talk, we will present the detailed experimental and analytical procedures for studying the static and dynamic magnetic properties of the ultra-thin CoFeB film at MgO interface performed in this work, and possible physical origin to explain the observation results. This presentation was partly based on results obtained from a project, JPNP16007, commissioned by the New Energy and Industrial Technology Development Organization (NEDO), Japan, and JSPS KAKENHI 18K14118.

**References:**

[1] S. Yakata et al., J. Appl. Phys. 105, 07D131 (2009). [2] S. Ikeda et al., Nat. Mater. 9.9, 721 (2010). [3] Y. Shiota et al., Nat. Mater. 11.1, 39 (2012). [4] S. Miwa et al., J. Phys. D 52.6, 063001 (2018). [5] B. Dieny and M. Chshiev, Rev. Mod. Phys. 89.2, 025008 (2017).

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/448829d743d9b248d164789bdfd98766.jpg)
Fig. 1, Interfacial PMA at CoFeB/MgO interface as a function of annealing temperature.

![](https://assets.underline.io/uploads/markdown_image/1/image/4b735ce897f2db9be15f9a3abeeb48b4.jpg)
Fig. 2, Increase of g factor at CoFeB/MgO interface as a function of annealing temperature.

 Effect of Annealing Temperature on Spectroscopic g Factor at CoFeB/MgO Interface

Bulk perpendicular magnetic anisotropy materials are proposed to be a promising candidate for next-generation ultra-high density and ultralow energy-consumption spintronic devices (1~5). In this presentation, we will report the experimental investigation on the structure, thermal stability, and magnetic properties of FePd thin films seeded by a Ru layer. An fcc-phase Ru layer induces the highly-ordered L10-phase FePd thin films with perpendicular magnetic anisotropy (Ku ~ 10.1 Merg/cm3). The thermal stability of FePd samples is then studied through the annealing process. It is found that a Ku ~ 6.8 Merg/cm3 can be obtained with the annealing temperature of 500 oC. In addition, the damping constant α, an important parameter for switching current density, is determined as a function of the testing temperature. We observe that α increases from 0.006 to 0.009 for the as-deposited FePd sample and from 0.006 to 0.012 for the FePd sample annealed at 400 oC as the testing temperature changes from 25 oC to 150 oC (6). These results suggest that Ru-seeded FePd provides great potential in scaling perpendicular magnetic tunnel junctions below 10 nm for applications in ultralow energy-consumption spintronic devices.

**References:**

 1. M. Weisheit, S. Fähler, A. Marty, Y. Souche, C. Poinsignon, and D. Givord, Science 315, 349-351 (2007). 2. S. Iihama, A. Sakuma, H. Naganuma, M. Oogane, T. Miyazaki, S. Mizukami, and Y. Ando, Appl. Phys. Lett. 105, 142403 (2014). 3. S. Iihama, A.Sakuma, H. Naganuma, M. Oogane, S. Mizukami, and Y. Ando, Phys. Rev. B 94, 174425 (2016). 4. D.-L. Zhang, K. B. Schliep, R. J. Wu, P. Quarterman, D. R. Hickey, Y. Lv, X. Chao, H. Li, J.-Y. Chen, Z. Zhao, M. Jamali, K. A. Mkhoyan, and J.-P. Wang, Appl. Phys. Lett. 112, 152401 (2018). 5. D.-L. Zhang, C. Sun, Y. Lv, K. B. Schliep, Z. Zhao, J.-Y. Chen, P. M. Voyles, and J.-P. Wang, Phys. Rev. Applied 9, 044028 (2018). 6. D.-L. Zhang, D. Huang, R. J. Wu, D. Lattery, J. Liu, X. Wang, D. B. Gopman, K. A. Mkhoyan, J.-P. Wang, and X. Wang, Appl. Phys. Lett. 117, 082405 (2020)

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/759ad544c5fd3f1e39164b751e71e26b.jpg)
FIG. 1. (a) Atomic-resolution HAADF-STEM image of the as-deposited Ru-seeded FePd thin film. (b) The out-of-plane (θ-2θ scans) XRD patterns as a function of the annealing temperatures for Ru-seeded FePd thin films. The very clear (001) and (002) peaks of L10-phase are observed, indicating the high-ordered FePd thin films. (c), (d) The temperature-dependent effective anisotropic field (Hk,eff) and damping constant extracted from TR-MOKE measurements for both as-deposited and 400 oC-annealed samples.

Downloads

Next from INTERMAG 2021

Disorder by Design in an Artificial Spin Ice with Dipolar Interactions: An Energetic Analysis

Similar lecture

Effect of high Tc interface layer in high Tc/ low Tc hybrid memory layer on thermally assisted spin-transfer-torque switching

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from INTERMAG 2021

Disorder by Design in an Artificial Spin Ice with Dipolar Interactions: An Energetic Analysis

Similar lecture

Effect of high Tc interface layer in high Tc/ low Tc hybrid memory layer on thermally assisted spin-transfer-torque switching

Downloads