Abstract

In our information society, the importance of energy conservation is increasing year by year. To tackle this issue, the transition from volatile memory to non-volatile memory can be a solution. One of the candidates for non-volatile memories is a racetrack memory[1], driven by current-induced domain wall motion (CIDWM). For practical use, faster CIDWM and lower threshold current density are key factors. In order to reach this goal, we have been focusing on and investigating Mn4N as a promising material. Mn4N film is an antiperovskite ferrimagnet without rare-earth elements, which is advantageous in fast magnetization reversal due to its perpendicular magnetic anisotropy (Ku ~ 1.1×105 J/m3) and a small saturation magnetization (MS ~ 80 kA/m)[2]. Previous study on 1-2μm-wide Mn4N strips showed the fastest spin-transfer-torque-driven domain wall motion (vDW ~ 900 m/s at 1.3×1012 A/m2) at room temperature (RT)[2], comparable to those reported in the system including rare-earth 4f magnets or heavy metals. To achieve faster vDW, Mn4N based mixed crystals have been studied in pursuit of the use of magnetic and/or angular momentum compensation for more efficient CIDWM thanks to diverged damping constants. Recently, Mn4-xNixN films have been suggested to have a magnetic compensation (MC) point between x = 0.1 and 0.25 at RT[3]. We found that Mn4-xCoxN films have a compensation point between x = 0 and 0.8 from x-ray magnetic circular dichroism (XMCD) measurements[4]. However, there is a lack of information about the magnetic behavior of Mn4-xCoxN at values x much smaller or larger than 0.8. Considering that the compensation in Mn4-xNixN takes place in a small range of composition x, further compensation points can be found when x is far from 0.8. In this work, we performed XMCD measurements on Mn4-xCoxN epitaxial films at x = 0.2 and 1.3 and investigated the change in magnetic structures by composition ratio to verify MC at RT.20-30 nm-thick Mn4-xCoxN films with x = 0-1.3 were epitaxially grown on SrTiO3(001) substrates by molecular beam epitaxy. SiO2 or Ta capping layers were sputtered in-situ on the surface to prevent oxidation. X-ray absorption spectroscopy (XAS) and XMCD measurements were performed at BL-16A of KEK-PF for x = 0.2 and 1.3. In these measurements, we applied an external magnetic field of 5 T perpendicular to sample surfaces. The incidence angle of the circularly polarized x-ray was 54.7°(magic angle) to the plane in order to simplify the sum rule calculation. For x = 0.8, the XMCD measurements were carried out at BL23SU of SPring-8[4].Figures 1(a)-1(c) show the XAS and XMCD spectra of Mn in Mn4-xCoxN at x = 0.2, 0.8[4], and 1.3, respectively. In these figures, the sharp peak near 640 eV comes from Mn atoms at corner sites (I sites), and the broad peak near 643 eV originates from those at face-centered sites (II sites)[4]. We observed the sign reversals of XMCD signals between x = 0.2 and 0.8, and also between x = 0.8 and 1.3. Similar sign reversals were also observed in the XMCD signals of Co atoms. These results indicate that MC occurs twice in the range of x = 0-1.3 in Mn4-xCoxN. Besides, the XAS spectra of Co indicate that Co preferentially occupied the corner sites. We also calculated the mean magnetic moment of Co by using the sum rule analysis. Figure 2 shows the expected magnetic structures derived from the XMCD measurements. Around x = 0.2, Co atoms preferentially replaced Mn(I) and thus the total magnetic moment of I site and that of face-centered site (II site) became closer. With further increasing x, the total magnetic moment became zero. In this manner, the first MC occurred between x = 0.2 and 0.8. After the first MC, the total magnetic moment of II sites became larger and this led to the reversal of magnetic moments of all atoms in order to minimize the Zeeman energy. Above x = 0.8, Co atoms gradually got to occupy II sites and thus the second MC occurs. As far as we investigated, Mn4-xCoxN is the only material which has two MC points at RT. This property can be useful for spintronic devices, for example, the use of compensation in a wide range of composition.The XMCD experiment for x = 0.2 and 1.3 was performed with the approval of the Photon Factory Program Advisory Committee (Proposal No. 2019G574). **

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call