France



The exchange bias effect, which manifests itself as a shift of the hysteresis loop of a ferromagnet when it is adjacent to an antiferromagnet, is a key feature used in such technological spheres as magnetic recording and storage as well as magnetic sensorics [1,2]. The exchange bias field &lt;i&gt;H&lt;/i&gt;&lt;sub&gt;ex&lt;/sub&gt; – an amount by which the loop is shifted along the magnetic field axis – as well as the coercivity &lt;i&gt;H&lt;/i&gt;&lt;sub&gt;c&lt;/sub&gt; are very sensitive to the temperature. Changes in both of these characteristics are directly connected with changes of magnetic properties of the layers, mainly, the anisotropies of the layers. Directly determining this parameter for an antiferromagnet is often a demanding task, especially dealing with a nanostructure. This task, however, is not only of fundamental, but of applied interest too. By systemizing a collection of data on currently used antiferromagnets it is then possible to propose brand new ones, suitable for technological applications. The industrial standards currently are Pt- and Ir-based antiferromagnets, in which the presence of precious metals substantially affects the overall cost of manufacturing. Recently scientists have been studying nitrides, oxides and Heusler-type compounds as affordable alternatives, but those come with their own limitations, such as relatively low Néel temperatures or high interdiffusion tendency. In this regard use of machine learning for prediction of possible new antiferromagnets seems rather promising, although it requires a dataset on existing antiferromagnets. For example, one of the most crucial properties of any magnetic material, and especially of an antiferromagnet is its anisotropy constant [3].&lt;br/&gt;One way of obtaining this data is by using different measurement protocols. For instance, the York protocol [3,4] has proven itself as a useful tool that by subsequent measurements of the hysteresis loops (Fig. 1) of an ferromagnet/antiferromagnet bilayer estimate the anisotropy constant of the antiferromagnet. All the measurements are conducted at 100 K, but between them the sample is activated: it is maintained at subsequently raising temperatures &lt;i&gt;T&lt;/i&gt;&lt;sub&gt;act&lt;/sub&gt; during the activation time &lt;i&gt;t&lt;/i&gt; in a field of the sign opposite to that in which the sample was set initially. This approach is based on the fact that magnetron-sputtered films, such as those considered in this work, have a grain-size distribution of crystallites. Due to this distribution at any finite temperature the whole antiferromagnetic layer divides into three categories: 1) stable grains, which pin the ferromagnet’s magnetization thus contributing to &lt;i&gt;H&lt;/i&gt;&lt;sub&gt;ex&lt;/sub&gt;; 2) unstable grains, which irreversibly follow the ferromagnet’s magnetization thus contributing to &lt;i&gt;H&lt;/i&gt;&lt;sub&gt;c&lt;/sub&gt;; 3) paramagnetic and superparamagnetic grains, which do not contribute to the hysteresis. This segregation is due to the fact that the energy of a grain depends on its volume and when it equals the thermal energy, a grain becomes unstable. The equation is as follows: &lt;i&gt;K&lt;/i&gt;&lt;i&gt;&lt;sub&gt;AF&lt;/sub&gt;&lt;/i&gt;&lt;i&gt; V = k&lt;sub&gt;B&lt;/sub&gt; T&lt;/i&gt; (1),&lt;br/&gt;where &lt;i&gt;K&lt;/i&gt;&lt;sub&gt;A&lt;i&gt;F&lt;/i&gt;&lt;/sub&gt; is the anisotropy constant of an antiferromagnet, &lt;i&gt;V&lt;/i&gt; is the volume of a grain, &lt;i&gt;k&lt;/i&gt;&lt;sub&gt;B&lt;/sub&gt; is the Boltzmann’s constant and &lt;i&gt;T&lt;/i&gt; is temperature. Through a series of field-cooling procedures alternating with hysteresis loop measurements it is possible to estimate the median blocking temperature &amp;lt;&lt;i&gt;T&lt;/i&gt;&lt;sub&gt;B&lt;/sub&gt;&amp;gt; at which exactly one half of the type 1 grains pin the magnetization in one direction while the other half does so in the opposite direction so that the net shift of the loop is zero (Fig. 2). The value of the anisotropy constant &lt;i&gt;K&lt;/i&gt;&lt;sub&gt;AF&lt;/sub&gt; at this particular temperature can be calculated by using the following equation:&lt;br/&gt;&lt;i&gt;K&lt;/i&gt;&lt;sub&gt;AF&lt;/sub&gt; (&amp;lt;&lt;i&gt;T&lt;/i&gt;&lt;sub&gt;B&lt;/sub&gt;&amp;gt;) = (&lt;i&gt;K&lt;/i&gt;&lt;sub&gt;B &lt;/sub&gt;&amp;lt;&lt;i&gt;T&lt;/i&gt;&lt;sub&gt;B&lt;/sub&gt;&amp;gt; ln (&lt;i&gt;t f&lt;/i&gt;&lt;sub&gt;0&lt;/sub&gt;))/&lt;i&gt;V&lt;/i&gt;&lt;sub&gt;m&lt;/sub&gt; (2), where &lt;i&gt;t&lt;/i&gt; is the time for which the setting field was applied, set to be 30 min in our experiment, &lt;i&gt;f&lt;/i&gt;&lt;sub&gt;0&lt;/sub&gt; is switching attempt frequency from the Néel-Arrhenius law taken to be 2.1 10&lt;sup&gt;12&lt;/sup&gt; s&lt;sup&gt;-1&lt;/sup&gt;[5], &lt;i&gt;V&lt;/i&gt;&lt;sub&gt;m&lt;/sub&gt; is the median volume of the grains, which was estimated by using the X-ray diffractometry and electron microscopy.&lt;br/&gt;In this work by using the York protocol and employing a vibrating sample magnetometer with a temperature chamber we estimate the anisotropy constants of the antiferromagnetic Ni-Mn in a series of magnetron-sputtered Fe-Ni/Ni-Mn films of different thickness of the Ni-Mn layer. Ni-Mn is a metallic antiferromagnet, systems based on which are capable of sustaining exchange bias up to temperatures as high as 700 K [6]. Using this approach, we estimated that a 20-nm thick Ni-Mn has &lt;i&gt;K&lt;/i&gt;&lt;sub&gt;AF&lt;/sub&gt; (320 K) = 2.1 10&lt;sup&gt;6&lt;/sup&gt; erg/cc.&lt;br/&gt;Our results can help to form a dataset on the anisotropy constants of antiferromagnets and can also be used to study the finite-size scaling effect in Ni-Mn.&lt;br/&gt;&lt;br/&gt;This work was financially supported by the Russian Science Foundation (RSC), project No. 18-72-10044.

**References:**

F. Radu, H. Zabel, Magnetic heterostructures, Springer, 97 (2008)&lt;br/&gt;J. Nogués, J. Sort, V. Vanglais, et.al., Phys. Rep., Vol. 422, p.65 (2005)&lt;br/&gt;K. O’Grady, J. Sinclair, K. Elphick, et. al., J. Appl. Phys., Vol. 128, p.040901 (2020)&lt;br/&gt;K. O’Grady, L.E. Fernandez-Outon, G. Vallejo-Fernandez, JMMM, Vol. 322, p.883 (2010)&lt;br/&gt;S. Jenkins, R.W. Chantrell, T.J. Clemmer, et. al., Phys. Rev. B, Vol. 100, p.220405(R) (2019)&lt;br/&gt;V.O. Vas&#39;kovskiy et al., J. Alloys Compd., Vol. 777, p.264 (2019)

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/c0d8de8aac5aab99d9374a4497ab0d7a.jpg)
Fig. 1. Hysteresis loops of a Fe&lt;sub&gt;20&lt;/sub&gt;Ni&lt;sub&gt;80&lt;/sub&gt; (40 nm)/Ni&lt;sub&gt;30&lt;/sub&gt;Mn&lt;sub&gt;70&lt;/sub&gt; (20nm) film obtained at 100 K following the York protocol after being set at an activation temperature &lt;i&gt;T&lt;/i&gt;&lt;sub&gt;act&lt;/sub&gt;.

![](https://assets.underline.io/uploads/markdown_image/1/image/dc6843a0d73b5424ac6cbd49931ce501.jpg)
Fig. 2. Measurement of the blocking temperature for a Fe&lt;sub&gt;20&lt;/sub&gt;Ni&lt;sub&gt;80&lt;/sub&gt; (40 nm)/Ni&lt;sub&gt;30&lt;/sub&gt;Mn&lt;sub&gt;70&lt;/sub&gt; (20nm) film. The arrow denotes the median blocking temperature &lt;&lt;i&gt;T&lt;/i&gt;&lt;sub&gt;B&lt;/sub&gt;&gt;.

INTERMAG 2021

 Estimating the Anisotropy Constant in an Antiferromagnet Through Exchange Bias in Polycrystalline Ni-Mn/Fe-Ni Films

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.

The exchange bias effect, which manifests itself as a shift of the hysteresis loop of a ferromagnet when it is adjacent to an antiferromagnet, is a key feature used in such technological spheres as magnetic recording and storage as well as magnetic sensorics [1,2]. The exchange bias field Hex – an amount by which the loop is shifted along the magnetic field axis – as well as the coercivity Hc are very sensitive to the temperature. Changes in both of these characteristics are directly connected with changes of magnetic properties of the layers, mainly, the anisotropies of the layers. Directly determining this parameter for an antiferromagnet is often a demanding task, especially dealing with a nanostructure. This task, however, is not only of fundamental, but of applied interest too. By systemizing a collection of data on currently used antiferromagnets it is then possible to propose brand new ones, suitable for technological applications. The industrial standards currently are Pt- and Ir-based antiferromagnets, in which the presence of precious metals substantially affects the overall cost of manufacturing. Recently scientists have been studying nitrides, oxides and Heusler-type compounds as affordable alternatives, but those come with their own limitations, such as relatively low Néel temperatures or high interdiffusion tendency. In this regard use of machine learning for prediction of possible new antiferromagnets seems rather promising, although it requires a dataset on existing antiferromagnets. For example, one of the most crucial properties of any magnetic material, and especially of an antiferromagnet is its anisotropy constant [3]. One way of obtaining this data is by using different measurement protocols. For instance, the York protocol [3,4] has proven itself as a useful tool that by subsequent measurements of the hysteresis loops (Fig. 1) of an ferromagnet/antiferromagnet bilayer estimate the anisotropy constant of the antiferromagnet. All the measurements are conducted at 100 K, but between them the sample is activated: it is maintained at subsequently raising temperatures Tact during the activation time t in a field of the sign opposite to that in which the sample was set initially. This approach is based on the fact that magnetron-sputtered films, such as those considered in this work, have a grain-size distribution of crystallites. Due to this distribution at any finite temperature the whole antiferromagnetic layer divides into three categories: 1) stable grains, which pin the ferromagnet’s magnetization thus contributing to Hex; 2) unstable grains, which irreversibly follow the ferromagnet’s magnetization thus contributing to Hc; 3) paramagnetic and superparamagnetic grains, which do not contribute to the hysteresis. This segregation is due to the fact that the energy of a grain depends on its volume and when it equals the thermal energy, a grain becomes unstable. The equation is as follows: KAF V = kB T (1), where KAF is the anisotropy constant of an antiferromagnet, V is the volume of a grain, kB is the Boltzmann’s constant and T is temperature. Through a series of field-cooling procedures alternating with hysteresis loop measurements it is possible to estimate the median blocking temperature &lt;TB&gt; at which exactly one half of the type 1 grains pin the magnetization in one direction while the other half does so in the opposite direction so that the net shift of the loop is zero (Fig. 2). The value of the anisotropy constant KAF at this particular temperature can be calculated by using the following equation: KAF (&lt;TB&gt;) = (KB &lt;TB&gt; ln (t f0))/Vm (2), where t is the time for which the setting field was applied, set to be 30 min in our experiment, f0 is switching attempt frequency from the Néel-Arrhenius law taken to be 2.1 1012 s-1[5], Vm is the median volume of the grains, which was estimated by using the X-ray diffractometry and electron microscopy. In this work by using the York protocol and employing a vibrating sample magnetometer with a temperature chamber we estimate the anisotropy constants of the antiferromagnetic Ni-Mn in a series of magnetron-sputtered Fe-Ni/Ni-Mn films of different thickness of the Ni-Mn layer. Ni-Mn is a metallic antiferromagnet, systems based on which are capable of sustaining exchange bias up to temperatures as high as 700 K [6]. Using this approach, we estimated that a 20-nm thick Ni-Mn has KAF (320 K) = 2.1 106 erg/cc. Our results can help to form a dataset on the anisotropy constants of antiferromagnets and can also be used to study the finite-size scaling effect in Ni-Mn. This work was financially supported by the Russian Science Foundation (RSC), project No. 18-72-10044.

**References:**

F. Radu, H. Zabel, Magnetic heterostructures, Springer, 97 (2008) J. Nogués, J. Sort, V. Vanglais, et.al., Phys. Rep., Vol. 422, p.65 (2005) K. O’Grady, J. Sinclair, K. Elphick, et. al., J. Appl. Phys., Vol. 128, p.040901 (2020) K. O’Grady, L.E. Fernandez-Outon, G. Vallejo-Fernandez, JMMM, Vol. 322, p.883 (2010) S. Jenkins, R.W. Chantrell, T.J. Clemmer, et. al., Phys. Rev. B, Vol. 100, p.220405(R) (2019) V.O. Vas'kovskiy et al., J. Alloys Compd., Vol. 777, p.264 (2019)

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/c0d8de8aac5aab99d9374a4497ab0d7a.jpg)
Fig. 1. Hysteresis loops of a Fe20Ni80 (40 nm)/Ni30Mn70 (20nm) film obtained at 100 K following the York protocol after being set at an activation temperature Tact.

![](https://assets.underline.io/uploads/markdown_image/1/image/dc6843a0d73b5424ac6cbd49931ce501.jpg)
Fig. 2. Measurement of the blocking temperature for a Fe20Ni80 (40 nm)/Ni30Mn70 (20nm) film. The arrow denotes the median blocking temperature <TB>.

poster

Magnetic tunnel junctions (MTJs) are known to be one of the main building blocks of spintronics. In these devices, the integration of molecular layers as tunnel barrier is envisioned as an opportunity to allow the engineering of spintronics at the molecular scale. Actually, thanks to the spin-dependent hybridization at ferromagnet/molecule interfaces, spin polarization and thus tunnel magnetoresistance can now be tailored by molecules. For example, it was shown that spin polarization could be inverted or enhanced depending on the hybridization coupling strength [1]. In this direction, among molecular systems, self-assembled monolayers (SAMs) appear as one of the most promising tool to tailor MTJs properties. Indeed, the anchoring group, body and head of molecules can be changed to tune the coupling strength of both ferromagnetic electrodes independently and at will. However, up to now, only basic “passive” molecules such as alkane chain have been integrated into MTJs [2-4]. These molecules could be defined as “passive” since their electronic properties cannot be switched by external stimulus. In this talk we will present molecular MTJs integrating “active” molecules from the diarylethene molecule family. These molecules can be switched between two stable states corresponding to an open and a closed form of its central part. As the electron delocalization, energy gap and coupling strength to the electrodes depend on the molecule state, the tunnel resistance and tunnel magnetoresistance (TMR) is expected to be tuned by switching the molecule. We will then discuss the magnetotransport properties of NiFe/diarylethene/Co MTJs (cf. Figure 2(a)). After presenting the fabrication process of the molecular MTJs, we will first show that the diarylethene molecules grafting over the NiFe ferromagnetic bottom electrode do not alter its surface properties. We will in particular show X-Ray Photoelectron Spectroscopy (XPS) measurements with spectra recorded at Ni and Fe 2p edges (figure 1) ruling out any trace of oxidization and validating our surface recovery protocol [5]. We will then focus on the transport results. We will show that we have been able to obtain non-linear I(V) and parabolic conductance curves in both open and closed forms. In addition, we have performed inelastic electron tunneling spectroscopy (IETS) in order to detect molecular vibrations inside the tunnel barrier. Those measurements confirm that these active molecules act as an effective tunnel barrier in our MTJs. We will also show magnetoresistance curves that have been recorded at different bias voltages, temperatures and magnetic field directions. We have obtained magnetoresistance effects in both open and closed formed of diarylethene molecules (result shown for the open form in figure 2(b)). The angular dependence of the magnetoresistance reveals that both TMR and tunneling anisotropic magnetoresistance effects coexist and allows us to disentangle them. All these studies demonstrate for the first time that switchable active molecules can be successfully integrated into MTJs to build multifunctional spintronic devices electrically and/or optically controllable.

**References:**

[1] C. Barraud et al., Nat. Phys. 6, 615 (2010) [2] J. R. Petta et al., Phys. Rev. Lett., 93, 136601 (2004) [3] W. Wang et al.,, Appl. Phys. Lett. 89, 153105 (2006) [4] M. Galbiati et al., Adv. Mater. 24, 6429 (2012) [5] M. Galbiati et al., AIP Advances. 5, 057131 (2015)]

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/8c88e8460872455a93231591ad25f5da.jpg)
Figure 1: XPS spectra at nickel and iron 2p edges of diarylethenes SAMs grafted on a NiFe electrode.

![](https://assets.underline.io/uploads/markdown_image/1/image/d3b94195d50c63cf8755c10a625fbf11.jpg)
Figure 2: (a) Schematic representation of the molecular MTJ (b) Magnetoresistance curve of a NiFe/diarylethene/Co MTJ at 50 K obtained for the open form of the diarylethene molecules

 Magnetic Tunnel Junctions Based on Photoswitchable Self Assembled Monolayers

The areal density of hard disk drives (HDD) must be increased to meet the needs of today’s society and its ever-increasing data creation. This requires both the writability of smaller bits and the ability to read these bits. This can only be achieved by reducing the size of conventional tunnelling magnetoresistance (TMR) sensors. A possible solution to this is tunnel anisotropic magnetoresistance (TAMR) which arises due to a change of density of states at the Fermi level when a magnetic sublattice rotates with respect to the crystal field. It is particularly present in materials which possess large spin-orbit (SO) interactions such as CoPt [1]. Very large TAMR signals have been observed at low temperatures consisting of an IrMn based magnetic tunnel junction (MTJ) [2] which if optimised at room temperature would remove the requirement for a second ferromagnetic electrode and thus reducing the size of the sensor. VAMPIRE [3] brings together a simple user interface through the use of plain text files to produce an open-source atomistic spin simulation software. The software is used in conjunction with vibrating sample magnetometry (VSM) and ferromagnetic resonance (FMR) to investigate the bilayer system of Ni80Fe20/Co70Fe30 deposited via magnetron sputtering, in order to optimise the ferromagnetic layer of the MTJ. Simulation parameters are achieved directly from experiment and supplied to VAMPIRE. Integration of simulation and experiment in this manner can yield insight into the magnetisation dynamics on the atomic scale and guide experimental development. The inclusion of the thin layer of Co70Fe30 within simulations of a permalloy thin film has been found to enhance the coercive field of ferromagnetic layer of the MTJ which can be seen in figure 1. This has been experimentally confirmed in addition to an increase in magnetisation saturation. Building on this, we found that there is an increase in exchange with 1nm of CoFe on top of NiFe through FMR analysis as shown in figure 2. Further research has been carried out on Ta, Ru and Pd seed layers to assess whether or not they would make viable seed layers for the FM layer of the MTJ. VSM measurements, FMR analysis and X-Ray diffraction (XRD) scans have been carried out to further analyse the stacks. As each layer of the TAMR structure is simulated, experimentally confirmed and finally optimised; it will pave the way for the future of data storage.

**References:**

[1] - Phys. Rev. Letts 100, 087204 (2008) [2] - Nature Mat. 10, 347 (2011) [3] - R. F. L. Evans et al., Appl. Phys. Lett. 104, 082410 (2014)

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/333fae02046ce4def0303705c39e4ebd.jpg)
Fig.1 VAMPIRE hysteresis loop Ni80Fe20(Y)/Co70Fe30(X) of normalised magnetisation against applied field. Inset is the visual representation of the generated system with a total z height of x + y

![](https://assets.underline.io/uploads/markdown_image/1/image/03e12db268de7ee7a1ade20819f602a3.jpg)
Fig.2 Spin wave stiffness, D against CoFe thickness, x for Ta(5)/NiFe(70)/CoFe(x)/Ta(5) sample set.

 Atomistic Simulations for TAMR Applications

CoFeB-based magnetic tunnel junctions (MTJ) is one of the most studied spintronic devices[1] due to its high tunneling magnetoresistance (TMR)[2] and low switching current density[3]. Here, using a direct characterization technique of spin-resolved photoemission, the spin polarization of an amorphous Co40Fe40B20 thin film prepared by magnetron sputtering is obtained. The spin-resolved photoemission measurement consists of a UVS 300 unpolarized helium lamp () and a SPECS PHOIBOS 150 hemisphere energy analyzer equipped with 3D Micro-Mott detector operated at a scattering energy of 25 kV. As shown in Fig. 1, the measured spin polarization at Fermi level is as high as 60%, which is comparable to the values obtained by other techniques. It is shown that our experimental value of spin polarization at is positive while the calculated spin polarization at ground state using first-principle calculation has a negative spin polarization at . Fig. 2 displays the density of states after considering the photoionization cross section, thermal broadening and experimental resolution broadening. The modified calculated spin polarization indicates a positive spin polarization at which is consistent to the experimental results. Moreover, we find that the maximum spin polarization is located at ~0.2 eV above , supported by the modified calculated spin polarization spectrum.

**References:**

1. C. Fu, X. Zhang, and Y. Duan. Magn. Magn. Mater. 492, 165657 (2019). 2. S. Ikeda, J. Hayakawa, and Y. Ashizawa, Appl. Phys. Lett. 93 (8), 082508 (2008). 3. B. Liu, X. Ruan, Z. and Y. Xu, Appl. Phys. Lett. 109 (4), 042401 (2016).

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/dd331884eb1494b055cdc054f6e0d160.jpg)
Fig. 1 (a) Spin-resolved photoemission spectra of the CoFeB thin film. (b) Corresponding spin polarization curve, with the black line is a guide to eyes.

![](https://assets.underline.io/uploads/markdown_image/1/image/961e789d99472be07ad7f2f0828f2dbc.jpg)
FIG. 2 Calculated density of states after considering (a) the photoionization cross sections, (b) temperature and (c) experimental resolution. (d) Spin polarization curve corresponds to (c).

 Direct Observation of Spin Polarization in Amorphous CoFeB Thin Film

At the interface of a heavy metal (HM) with a strong spin-orbit coupling (SOC) and a ferromagnetic material (FM), a break in inversion symmetry arises leading to the appearance of several phenomena. One of them is the antisymmetric exchange interaction called the Dzyaloshinskii-Moriya Interaction (iDMI). It is responsible for the stabilization of chiral magnetic structures such as skyrmions and for the non-reciprocity of spin wave propagation. Another interface phenomenon, subject of extensive research in magnetic heterostructures (e.g. for HM /FM / metal oxide (MOx) type stacks) is interfacial Perpendicular Magnetic Anisotropy (iPMA) [2]. This originates from the level of oxidation, from the interface roughness and from the SOC [3]. These two quantities are of considerable interest because of their crucial importance for applications in spintronic devices, such as non-volatile magnetic memories and hard disk recording [4]. As an example, an MgO layer is often used as a tunnel barrier in magnetic tunnel junctions (MTJs), providing high spin transfer torque (SST), or in tunnel magnetoresistance (TMR), which are key properties for such devices. In order to optimize the magnetic properties at the interfaces of thin films, a thermal treatment is often used because it yields structural changes and atomic rearrangement and thus induces phase transition or alloys formation. However, annealing at high temperatures can also induce oxidation thus reducing the volume magnetic moment or increasing the magnetic anisotropy. In this work, we investigate the effect of the annealing temperature on iPMA and iDMI in Pt/Co/MgO stacks, with variable Co thickness (1≤tCo≤6 nm). These stacks were deposited by magnetron sputtering on a Si substrate and annealed at different temperatures (Ta), up to 400°C. For all the studied films, the saturation magnetic moment per unit area was measured by Vibrating Sample magnetometry (VSM) for each Co thickness. This allowed us to accurately determine the saturation magnetization Ms and the dead layer thickness td which are necessary to evaluate iDMI and iPMA. These quantities have been measured using the Brillouin light scattering (BLS) in the Damon-Eshbach (DE) configuration. Effective magnetizations (Meff) were deduced from measurements with fixed wave vector (ksw = 8.08 µm-1) and variable magnetic fields. The presence of the DMI in the studied structures induces a frequency shift between the two Stokes and anti-Stokes modes, this allowed us to quantify the iDMI effective constant (Deff) which is proportional to this shift. The variation of Meff as a function of the inverse of the Co thickness, shows the presence of two different linear regimes separated by a critical thickness (tc), which depends on Ta. This is generally observed in the case of epitaxial growth with a lattice misfit between the constituents, the strain can contribute to Kv⊥ (regime I: t &lt; tc), or to Ks⊥ (regime II: tc &lt; t). In order to separate the different surface/interface anisotropies (magnetocrystalline, magnetoelastic and Neel type) from the volume anisotropies, we have used an adapted model [1]. In figure (1), we can see the evolution of these different anisotropies as a function of the annealing temperature. For Ta above 200°C, a variation in the different contributions of the surface and volume anisotropies is observed. Similar to the iPMA, the evolution of the iDMI interaction was investigated as a function of the annealing temperature and the thickness of the effective magnetic layer. The measurements show that the effective iDMI constant (Deff) is proportional to the inverse effective magnetic thickness. Therefore, for low thicknesses we find that the values of Deff deviate from the linear regime probably due to the reduction of magnetic thickness (effect of dead layer). The iDMI surface constant Ds is then deduced form Deff wich decreases linearly (with absolute value) with the annealing temperature Ta (see figure(2)). It varies from (-1.9pJ/m) at room temperature to (-0.7pJ/m) at 400°C annealing temperature. We then conclude the possibility to modulate the strength of the iDMI by changing the annealing temperature.

**References:**

[1] M. T. Johnson, P. J. H. Bloemenz, F. J. A. den Broeder, and J. J. de Vries, Rep. Prog. Phys. 59, 1409 (1996). [2] B. Dieny and M. Chshiev, Rev. Mod. Phys. 89, 025008 (2017). [3] N.-H. Kim, et al , AIP Adv. 7, 035213 (2017). [4] Xueying Zhang, et al ACS Applied Materials & Interfaces 2018 10 (19), 16887-16892 [5] J. H. Moon, et al, Phys. Rev. B 88, 184404 (2013)

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/dd67d9db22ce764185c90945ff722646.jpg)
Variation of the different anisotropy constants magnetocristalline (Kmc⊥), volume (KIme,v⊥) magnetoelastic , Neel-type interface (KN⊥) and the surface (KIIme,s⊥) magnetoelastic as a function of annealing temperature.

![](https://assets.underline.io/uploads/markdown_image/1/image/a29103b80c55a2aa457283b58a111134.jpg)
iDMI surface constant Ds as a function of annealing temperature.

 Annealing temperature and thickness dependencies of perpendicular magnetic Anisotropy and Dzyaloshinskii-Moriya Interaction of Pt/Co/MgO thin film.

Over the past few years, spintronic devices have attracted considerable attention due to their potential application compared to the classical electronics. Recently, a new concept of racetrack memory (RM) is developed, where domain walls (DWS) are replaced by magnetic chiral spin textures (skyrmions) [1] stabilized in the presence of the interfacial Dzyaloshinskii-Moriya interaction (iDMI). Their small size, high speed, and low energy consumption make skyrmions good candidates for future spintronic data storage. Skyrmion-RM will would allow to considerably increasing storage densities and information transfer speed [1, 2]. Therefore such devices performance depends significantly on the proper modulation of the Nonmagnetic/ Ferromagnetic (FM) interface and interfacial associated phenomena. In addition to the Heavy-metal (HM)/ (FM) interfaces and related spin-orbit coupling (SOC) interaction, it has recently been demonstrated the presence of the perpendicular magnetic anisotropy (PMA) and the iDMI in the FM/ Metal Oxide (Mox) interfaces. The PMA in such interface is due to the hybridization between the ferromagnetic metal and oxygen ions [3]. Recently, density functional theory has predicted the controlling of the iDMI sign and magnitude by changing the oxygen coverage of the FM layer [4]. However, knowing that these effects are dependent on the interface configuration, the study of the iDMI in the FM/MOx is still lacking. An efficient way to determine the role and the contribution of such interface is to investigate a FM/MOx system with or without a HM underlayer. Thus, in this work we propose the investigation of the PMA and, the iDMI as a function of the annealing temperature, the PtOx thickness and the oxidation state in CoFeB/PtOx interfaces. We expect to obtain overoxidation state for the thinner Pt films, where oxygen penetrates the underlying magnetic layer, whereas underoxidazied and optimally oxidized states are obtained for thicker Pt layers. Two categories of samples were grown by DC magnetron sputtering on thermally oxidized silicon, with or without Ta (3 nm) buffer layer. In the first category, the Co8Fe72B20 (CFB) thickness was varied up to 10 nm (2nm≦ tCFB≦10 nm) and capped by TaOx (0.8 nm). In the second category, CFB thickness was kept at 1.5 nm and was capped by variable thicknesses of PtOx (0.7nm≦ tPt≦1.6 nm). Both as grown and 225 °C annealed are studied for this second category. For this experimental investigation, Vibrating sample magnetometer (VSM), Brillouin light scattering (BLS) in Damon-Esbach (DE) configuration and Microstrip line ferromagnetic resonance (MS-FMR) were combined. The VSM was used to investigate the static magnetic properties (magnetization at saturation (Ms) and magnetic dead layer thickness (td)). The FMR was used for the determination of the gyromagnetic ratio, and for the investigation of the PMA in the thicker films. The BLS was employed for the study of the PMA (thinner films) and the iDMI. For the PMA, the spectra were measured at spin wave (SW) vector ( ksw= 8.08 µm−1) under variable in-plane applied magnetic fields, while for the iDMI measurements, BLS spectra were measured as a function of the ksw under in-plane saturating magnetic field. VSM measurement revealed that td is approximately the same for the CFB/TaOx (td≈ 0.7 nm) and CFB/PtOx (td≈ 1 nm) systems. For the CFB/PtOx stackcs we notice slightly higher value for the samples with Ta buffer layer due to the supplement of intermixing at the bottom interface. The larger value of the td (≈ 1 nm) obtained for CFB/PtOx compared to CFB/TaOx samples may be due to the fact that the Pt does not oxidize, the oxygen penetrates more into the CFB films thus increasing the thickness of the magnetic dead layer. Ms is also approximately the same for all the samples (around 1510 emu/cm3) which is in good agreement for systems of CFB rich on Fe [5]. The PMA investigation in the CFB/TaOx samples has shown the degradation of the interface anisotropy by inserting a Ta buffer layer. Indeed, the surface anisotropy constant (Ks) in the Ta/CFB/TaOx is twice weaker for the samples without Ta buffer layer confirming the degradation of the interface anisotropy by the Ta buffer layer. The PMA and iDMI measurements as a function of the Pt thickness have shown that the effective magnetization (Meff) and the effective iDMI constant (Deff) have the same trend for all the systems. Deff and Meff increase with increasing tPt thus allowing characterizing the strength of the iDMI and PMA for the CFB/PtOx interface. In the as deposited CFB/PtOx films, the thinnest Pt (tPt = 0.7 nm) sample is perpendicularly magnetized. The increasing of Meff as Pt thickness increases suggested that the oxidation of CoFeB decreases. The experimental data of Meff and Deff were then fitted using exponential grow function, to extract a decay thickness (λ). It is found that, λ for the Meff is higher compared to that of Deff for as grown CFB/PtOx stacks, confirming that the iDMI is slightly more localized at the first monolayers at the interface compared to the PMA. Measurements of spin pumping induced damping have also been performed by FMR and BLS will also be presented.

**References:**

[1] R. Tomasello, E. Martine., R. Zivieri et al. . Sci Rep 4, 6784, (2014) [2] W. Kang, X. Chen, D. Zhu et al., IEEE 7th Non-Volatile Memory Systems and Applications Symposium (NVMSA), Hakodate, pp. 7-12, (2018) [3] S. Monso, B. Rodmacq, S. Auffret, et al. Applied Physics Letters, 80, no 22, pp 4157, (2002) [4] A. Belabbes, G. Bihlmayer, S. Blügel, et al. Scientific Reports 6, 24634, (2016). [5] T. Ogasawara, M. Oogane, M. Al-Mahdawi et al, AIP Adv. 9, 125053, (2019).

**Images:**

 Interfacial Dzyaloshinskii Moriya interaction and perpendicular magnetic anisotropy in CoFeB/PtOx based systems

In this work, we investigate the influence of interface parameters on the magnetic energies such as the perpendicular magnetic anisotropy (PMA) and the interfacial Dzyaloshinskii-Moriya interaction (DMI)[1, 2] of Pd(111)/Cox/CoO1-x/Pd ultrathin epitaxial films with different oxidation depths. Epitaxial Pd(3 nm)/Co(1 nm) films were grown on a Si(111)/Cu(2nm) surface by molecular beam epitaxy in ultrahigh vacuum chamber, and after that, films were exposed the procedure of artificial oxidation. We varied the oxide thickness and this way we changed ferromagnetic thickness. The lattice parameters of bulk Pd and Co materials are mismatched by about 9,7%, which leads to the appearance of significant elastic strains on the bottom interface, which induced strong perpendicular magnetic anisotropy [3]. Oxidation in heavy metal(HM)/ferromagnetic metal(FM)/Oxide polycrystalline trilayers leads to significant changes of the magnetic properties [4]. The lattice period of the metal layers during growth and their structure were analyzed by means of RHEED. Magnetic hysteresis loops of the samples were measured using a vibrating sample magnetometer (VSM). Magnetic structure was investigated by a magneto-optical Kerr effect (MOKE) microscope. The MOKE microscope was equipped with a hand-made coil applying out-of-plane magnetic fields and an in-plane electromagnet. The out-of-plane coil was used in pulse mode and produced magnetic fields with an amplitude up to 50 mT and a width down to 2 msec. Artificial oxidation of magnetic layer leads to symmetry breaking on the bottom and top magnetic interfaces, which reduces compensation of interfacial DMI. Determination of the thickness of the magnetic layers carried out by measuring the saturation magnetization with vibrating samples magnetometry. Increasing of oxidation time leads to decreasing of volume of magnetic material and to increasing of effective PMA energy. The reported study was partially funded by Council for Grants of the President of the Russian Federation (project - MK-1384.2021.1.2) and the Russian Ministry of Science and Higher Education (the state task 0657-2020-0013).

**References:**

[1] Dzyaloshinsky I., Journal of Physics and Chemistry of Solids, Vol. 4.4, p.241-255 (1958) [2] Moriya T., Physical review, Vol. 120, p. 91 (1960) [3] Davydenko A. V., et al., Phys. Rev. B, Vol. 95, 064430 (2017) [4] Feng, Junxiao, et al., Physical Review Applied, Vol. 13.4, 044029 (2020)

**Images:**

 Effect of Oxidation of Top Interface on Magnetic Parameters of Epitaxial Films with Dzyaloshinskii-Moriya Interaction

Cobalt ferrite (CoFe2O4, CFO) is a material used for technological applications such as permanent magnets, spintronic devices and recording media. This is due to its high magnetocrystalline anisotropy, large saturation magnetization in comparison to other ferrites, elevated Curie temperature and large magnetostrictive constant. Structurally, the ideal CFO corresponds to an inverse spinel (AB2O4), i.e with all the cobalt cations (Co2+) and half of the iron ions (Fe3+) located in octahedral sites (B), and the other half of the Fe3+ occupying tetrahedral sites (A). However, in the case of thin films, depending on the growth conditions and subsequent annealing treatments, a fraction of the cobalt cations can be forced to move to tetrahedral sites. The subject of cationic distribution in cobalt ferrite thin films has been already addressed in the past, with Mössbauer spectroscopy playing a fundamental role. In most cases, however, the films have been grown on oxide substrates whilst films grown on metallic substrates have been scarcely studied. In this work, we have studied the properties of ultrathin CoFe2O4 films grown by oxygen assisted molecular beam epitaxy on a Pt(111) single crystal. The deposition was performed at 523 K temperature, and the resulting film was subsequently annealed at intermediate temperatures (up to 773 K) in order to prevent the formation of multiphasic films, island growth or dewetting phenomena. Samples were grown from a few nanometers thickness to 20 nm. We characterized their composition and structural properties after each step of the growth process by Conversion Electron Mössbauer and Auger electron spectroscopies (CEMS and AES, respectively), low energy electron diffraction (LEED), and scanning tunneling microscopy (STM). An important particularity of this work was that both the growth and the characterization of the films were performed in-situ, i.e. in the same ultra high vacuum chamber, without exposure of the samples to the laboratory atmosphere in order to avoid any chemical transformations resulting from such exposure.

**References:**

V.A.M. Brabers, Handbook of Magnetic Materials, Vol. 8, pp.189–324 (1995) J. Smit and H. Wijn, Philips’ technical library, Vol. 278, pp. 136-172 (1959) M. De Santis et al, Acta Cryst B, Vol. 75, pp. 8-17 (2019)

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/b6c2363e9d22f051bc8da9b2820f1ee6.jpg)
Fig 1. Mössbauer spectra obtained for the 5 nm cobalt ferrite thin film for the different stages: a) as grown measured at room temperature, b) and c) annealed to 673 K and 773 K measured at room temperature and 125 K, respectively.

 Ultrathin CoFe2O4 films grown by molecular beam epitaxy on Pt(111)

The properties of superconductor/ferromagnet multilayers and junctions have been intensively studied from the standpoint of the interplay between superconductivity and ferromagnetism. Though in a special case superconductivity and ferromagnetism can coexist and exhibit novel phenomena such as π state in superconductor/ferromagnet/superconductor junctions [1], in general, the superconducting order parameter is expected to suppress in the superconducting layer but to penetrate into the ferromagnetic layer. A number of the studies have been carried out with ferromagnetic component consisting of 3d transition metals like Fe, Co and Ni [2]. In contrast, only few reports exist for epitaxial NbN/Heusler alloy layered films and their devices so far; for example, giant coercivity enhancement in NbN/Co2FeSi bilayers [3], efficient injection of pure spin currents in NbN/Co2MnSi devices [4], superconductivity in NbN/Co2(Fe,Mn)Si/Au trilayers [5,6], and spin polarization measurements of Co2FeSi by the epitaxial contact Andreev reflection spectroscopy [7]. Therefore, we focus on superconducting properties of epitaxial films of a superconducting NbN layer attached to a highly spin polarized Heusler alloy Co2FeSi layer, where the superconductivity of the NbN layer penetrates into the ferromagnetic Co2FeSi layer. We investigate how the Co2FeSi layer with the high spin polarization affects the superconductivity of the NbN layer in epitaxial NbN/Co2FeSi/Au films through their interface. We have fabricated NbN/Co2FeSi/Au epitaxial films on an MgO-substrate by an ultrahigh-vacuum (UHV)-compatible magnetron sputtering system. The thickness of the CFMS film was 5 nm while the thickness of the NbN film was fixed to 100 nm. Considering the NbN film was thicker enough than its coherence length ξNbN(0) at 0 K, the superconductivity of the layered film should be in the three-dimensional (3D) regime, due to satisfying dNbN &gt;&gt; ξNbN(0). Superconducting properties of the epitaxial films were evaluated by resistivity measurements with the standard four-terminal technique under both parallel and perpendicular magnetic fields up to 17 T. The superconducting transition temperature Tc of NbN/Co2FeSi/Au trilayer films was determined to be 16.1 K in the absence of magnetic field. Temperature dependence of the resistivity ρ(T) was measured in both magnetic fields parallel and perpendicular to the surface of NbN/Co2FeSi/Au trilayer films. Figure 1 shows the field dependence of the pinning potential U0(H) for both applied magnetic field directions deduced from the Arrhenius plots for the ρ(T,H) curves. As shown in Fig. 1, U0(H) was well described by H−α(1−H/H*)β above 5 T by the similar model with the exponents in the field dependence of the pinning force density, where H* is a characteristic field representing the irreversibility field at 0 K. Temperature dependence of the upper critical field μ0Hc2(T) was evaluated from the ρ(T) measurements. Figure 2 represents the upper critical fields μ0Hc2||(T) and μ0Hc2⊥(T) of a NbN/Co2FeSi/Au trilayer film in two magnetic field directions applied parallel and perpendicular to the surface of the trilayer From Fig. 2, the upper critical field Hc2(0) at 0 K was also deduced to be μ0Hc2||(0) = 23.2 T for the parallel magnetic field and μ0Hc2⊥(0) = 16.2 T for the perpendicular magnetic field, respectively. The experimental results under magnetic fields revealed the superconductivity of the NbN layer was affected by the interplay between the superconducting NbN layer and the highly spin polarized Co2FeSi layer.

**References:**

[1] T. Kontos, M. Aprili, J. Lesueur, F. Genêt, B. Stephanidis, R. Boursier, Phys. Rev. Lett. 89 (2002) 137007. [2] L. Lazar, K. Westerholt, H. Zabel, L. R. Tagirov, Y. V. Goryunov, N. N. Garif'yanov, I. A. Garifullin, Phys. Rev. B 61 (2000) 3711. [3] A. Gupta, G. Singh, D. Kumar, H. Kishan, and R. C. Budhani, Appl. Phys. Lett. 103 (2013) 182602. [4] D. Kumar, P. C. Joshi, Z. Hossain, and R. C. Budhani, Appl. Phys. Lett. 102 (2013) 112409. [5] I. Shigeta, T. Kubota, K. Makise, S. Kimura, S. Awaji, B. Shinozaki, K. Koyama, K. Takanashi, and M. Hiroi, IEEE Magn. Lett. 8 (2017) 3305605. [6] I. Shigeta, T. Kubota, Y. Sakuraba, S. Kimura, S. Awaji, K. Takanashi, and M. Hiroi, Physica B 536 (2018) 310. [7] I. Shigeta, T. Kubota, Y. Sakuraba, C. G. Molenaar, J. N. Beukers, S. Kimura, A. A. Golubov, A. Brinkman, S. Awaji, K. Takanashi, and Masahiko Hiroi, Appl. Phys. Lett. 112 (2018) 072402.

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/33cb1ef7173a0adf01a17dd4170bd8e5.jpg)
Fig. 1. Pinning potential U0(H) as a function of magnetic field H. The blue circles are U0(H) parallel magnetic fields H|| to the surface for the NbN/Co2FeSi/Au trilayer film and the red squares are perpendicular magnetic fields H⊥. Inset: Relationship between ln ρ0f and U0 for both magnetic fields H|| and H⊥.

![](https://assets.underline.io/uploads/markdown_image/1/image/508049c8022712d9aedc6040944a89ff.jpg)
Fig. 2. Temperature dependence of the upper critical field Hc2(T) for both directions deduced from ρ(T) measurements at various magnetic fields.

 Superconducting properties of epitaxial films of superconducting NbN and highly spin polarized Co2FeSi under high magnetic fields

Antiferromagnetic coupling between magnetic films has long been studied for the design and tuning of spin structures, and is a core issue in spintronics research. Such a coupling has been widely reported in metallic multilayers, but the effect is seldom seen in the more complex multilayer film interfaces [1]. Here we studied the antiferromagnetic coupling at the interface between terbium iron garnet (Tb3Fe5O12, TbIG) and yttrium iron garnet (Y3Fe5O12, YIG) thin films, being manifested in form of an inverted hysteresis loop at low temperatures. TbIG (7 nm)/YIG (23 nm) was deposited on Y3Al5O12 (YAG) substrates with (110) orientation by pulsed laser deposition at 437 K under 100 mTorr oxygen pressure. As a control, a sample with 5 nm Gd3Ga5O12 (GGG) spacer between the two iron garnet layers was also prepared. The crystal structure of the films was characterized by a high-resolution X-ray diffractometer (HRXRD). The surface morphology of samples was characterized by atomic force microscopy (AFM). The microstructures and thickness of the samples were examined by high-resolution transmission electron microscopy (HRTEM). For transport measurements, Pt (5 nm) Hall bar patterns were deposited on the garnet samples by rf sputtering. Anomalous-Hall effect (AHE) of samples was measured by a physical property measurement system (PPMS). The magnetic properties of the samples were measured by the vibrating sample magnetometer (VSM) option of the PPMS. Fig. 1(a) shows the HRXRD pattern of TbIG (7 nm)/YIG (23 nm), demonstrating epitaxial growth of the films on YAG with strong (110) orientation (rocking curve with a full width at half maximum (FWHM) of 0.064°, Fig. 1(b)). AFM scan indicates a small root mean square roughness of 0.38 nm, with clear atomic terraces. Fig. 1(c) and (d) show the in-plane magnetic hysteresis loops of single layer YIG (30 nm) and TbIG (30 nm) measured at 10 K, respectively; both films were deposited on YAG (110) substrates under the aforementioned conditions. Due to the strong contribution from the Tb at low temperatures, the TbIG monolayer film exhibited much stronger magnetization than YIG of the same thickness did. A strong magnetocrystalline anisotropy was observed in the TbIG monolayer as evidenced by the high coercivity (Hc-TbIG = 5.5 kOe; Hc-YIG = 0.19 kOe) [2]. As shown in Fig. 2(a), the hysteresis loop of the TbIG (7 nm)/YIG (23 nm) sample measured at 10 K shows an obvious soft/hard magnetic coupling phenomenon, supported by the appearance of double coercivity behavior. The switching at the low field was mainly attributed to YIG, while the reversal at the high field was attributed to TbIG. The most characteristic feature of the hysteresis curve was the inversion under the low magnetic field (Inset of Fig. 2(a)): when the external field decreased from the positive saturation, the YIG layer first reversed its magnetization at a small positive field, while the magnetization of TbIG remained undisturbed. This clearly indicated the antiferromagnetic coupling between the TbIG/YIG at the interface [3]. In the case of the control sample, when a 5 nm GGG spacer is present at the TbIG/YIG interface, one could only observe the normal magnetic hysteresis loop as shown in Fig. 2(b). The inverted loop at the low field disappeared but the double coercivity could still be clearly observed, indicating the inverted loop at the low field was caused by the antiferromagnetic coupling at the interface of two different ferrimagnets. In summary, the antiferromagnetic coupling effect has been observed between two neighboring ferrimagnets, manifested in the form of an inverted hysteresis loop at the low magnetic field. Temperature-dependence of such an antiferromagnetic coupling effect and the anomalous-Hall effect of samples would be conducted to further elucidate the nature of the coupling effect. Financial support by the RGC, HKSAR (PolyU 15302320) is acknowledged.

**References:**

[1] T. Maity, D. Kepaptsoglou and S. Roy, "Observation of complete inversion of the hysteresis loop in a bimodal magnetic thin film," Physical Review B., 95(10):100401, 2017. [2] R. F. Pearson, "Magnetocrystalline Anisotropy of Rare-Earth Iron Garnets," Journal of Applied Physics., 33(3):1236-1242, 1962. [3] H. Zhang, J. Zhang and J. R. Sun, "Antiferromagnetic interlayer coupling of (111)-oriented La0.67Sr0.33MnO3/SrRuO3 superlattices," Chinese Physics B., 28(3):037501, 2019.

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/1d530bd49dfedcea853b1f515b39583c.jpg)
Fig.1. (a) θ/2θ scan and (b) rocking curve of TbIG (7 nm)/YIG (23 nm)/YAG (110) sample. Inset of (a) shows the AFM image (scan height = 10 nm). In-plane magnetic hysteresis loops of YIG (30 nm) and TbIG (30 nm) single layer films at 10 K are shown in (c) and (d), respectively.

![](https://assets.underline.io/uploads/markdown_image/1/image/931b651b61d4958badb65d7240caf8fe.jpg)
Fig 2. In-plane magnetic hysteresis loops of (a) TbIG (7 nm)/YIG (23 nm) and (b) TbIG (7 nm)/GGG (5 nm)/YIG (23 nm) samples at 10 K. The insets in each figure show the corresponding enlarged images at low field. The arrows indicate the field sweep directions.

 Observation of Antiferromagnetic Coupling between Ferrimagnetic Garnet Thin Films

As one of the methods to induce uniaxial magnetic anisotropy in magnetic films, the method of “periodically patterning of films” has attracted more and more research interest because of its promising application in magnetic electronic devices. It has been confirmed that in-plane uniaxial anisotropy of magnetic films can be induced by substrate with ripple patterns，and the direction of easy axis of induced anisotropy is parallel to the ripple direction [1]. Ki in Ref. 2 produced strong in-plane uniaxial anisotropy in NiFe films by annealing the sapphire substrate to form substrate of periodical ripple patterns; the anisotropy in samples is from the dipole effect by surface magnetic charge and shape anisotropy by rippled surface morphology [2]. However, the relationship between annealing-time dependent faceting and coarsening mechanisms of substrate on permeability spectrum of films has not been clarified. In our work, we fixed the annealing temperature at 1400 degrees centigrade and changed the annealing time of substrate to prepare rippled sapphire substrates with different periodicity. Then, we deposited Ta (5 nm)/NiFe (50 nm)/Ta (2 nm) three-layer films on these substrates to investigate the annealing-time dependent faceting and coarsening mechanism of substrate on permeability spectrum of film. The surface morphology of substrates was characterized by atomic-force microscopy, see Fig. 1. The effective modulation of substrate ripple is confirmed. Fig. 2 (a)-(f) show the permeability spectra of real parts μ' and imaginary parts μ'' at no external field for films deposited on substrates annealed for 0-10 h. The results show both μ' and μ'' of NiFe films that were deposited on substrates with periodical ripples are smaller than the film on substrate not annealed. The magnitude of permeability can be modulated by annealing time of substrate and it is related to the number of ripples per unit area and the geometric size of the ripples. The resonance frequency of films on substrates with periodical ripples is higher than that of film on substrate not annealed because of the in-plane uniaxial magnetic anisotropy induced by rippled patterns. This work provides an optimized annealing time to obtain high natural resonance frequency and a method to design permeability of magnetic films. This work would be helpful for film-based magnetic devices on ripple substrate.

**References:**

[1] Sarathlal K V, Kumar D, Gupta A, Appl. Phys. Lett. 98, 123111 (2011). [2] Ki S and Dho J, Appl. Phys. Lett. 106, 212404 (2015).

**Images:**

![](https://assets.underline.io/uploads/markdown_image/1/image/10f0e07017bd484924898841ec835ef7.jpg)
Fig. 1 Dependence of wavelength and height of substrate ripples on different annealing time. The inset maps are surface morphologies of these annealed substrates characterized by atomic-force microscopy.

![](https://assets.underline.io/uploads/markdown_image/1/image/96353c611ffb4f505e21b8cb9c14c705.jpg)
Fig. 2 Permeability spectra of real parts μ' and imaginary parts μ'' at no external field for films deposited on substrates annealed for different time: (a) 0 h, (b) 2 h, (c) 4 h, (d) 6 h, (e) 8 h, (f) 10 h.

Next from INTERMAG 2021

Magnetic Tunnel Junctions Based on Photoswitchable Self Assembled Monolayers

Similar lecture

Highly nonlinear magnetoelectric effect in buckled honeycomb antiferromagnetic Co4Ta2O9