Magnetic states of Mn and Co atoms atCo2MnGe/MgOinterfaces seen via soft x-ray magnetic circular dichroism

Abstract

The magnetic states of Mn and Co atoms in Co-rich Co 2 MnGe Heusler alloy thin films facing an MgO barrier were studied by means of soft x-ray magnetic circular dichroism ͑XMCD͒. In particular, the Co 2 MnGe film-thickness dependence of the Mn and Co magnetic moments was investigated. With a decrease in the Co 2 MnGe film thickness to 1-2 monolayers ͑MLs͒, the spin magnetic moment of Mn decreased and the Mn L 2,3 -edge x-ray absorption spectra ͑XAS͒ showed a Mn 2+ -like multiplet structure in MnO, in contrast to samples thicker than 4 ML, indicating that the Mn atoms of the 1 and 2 ML samples were oxidized. The Co spin magnetic moment increased slightly with decreasing thickness. A Co 2+ -like multiplet structure in CoO was not observed in all the Co L 2,3 -edge XAS and XMCD, indicating that, even in the ultrathin samples, the Co atoms were not oxidized, and were more strongly spin polarized than those in the thicker samples. Co spin magnetic moments of 1.40-1.77 B larger than the theoretical value for ideal stoichiometric Co 2 MnGe ͑ϳ1 B ͒ and the Co-rich film composition imply the presence of Co antisites that would lower the spin polarization.

Effects of off-stoichiometry on the spin polarization at the Co2MnβGe0.38/MgO interfa

“…According to the site occupation model employed here (Table I) show a shoulder-like structure observed on the higher energy side of the Co L 3 -edge XAS, common to thin film bulk samples. 27,33 Co-oxide-like multiplet structures 20 were not observed for any samples, meaning that Co atoms were not oxidized even in the interfacial region, consistent with the previous report. 27 To determine the Co magnetic moment, we again used the sum rules 36,37 assuming the 3d hole number (n h ) to be 2.2.…”

“…27,33 Co-oxide-like multiplet structures 20 were not observed for any samples, meaning that Co atoms were not oxidized even in the interfacial region, consistent with the previous report. 27 To determine the Co magnetic moment, we again used the sum rules 36,37 assuming the 3d hole number (n h ) to be 2.2. 20,24,27 As shown in Fig 4(c), we experimentally determined the Co spin magnetic moment m spin (Co) to be in the range of 1.2 μ B -1.4 μ B , which is obviously larger than the theoretically predicted bulk value of 1.06 μ B .…”

“…We have assumed that the contribution of the dipole operator term T z in the spin sum rule is negligibly small because the Mn site in bulk CMG has T d symmetry. 27 Since the L 3 and L 2 regions in Mn L 2,3 XAS and XMCD overlap due to the strong exchange interaction between the Mn 2p core hole and the Mn 3d electrons, we have divided the obtained spin magnetic moment by the correction factor 0.68 given by Teramura et al 38 We have assumed a 3d hole number (n h ) of 4.5 for Mn 20,24,27 on the basis of the band structure calculations. The spin magnetic moment m spin (Mn) thus deduced is plotted as a function of β in Fig.…”

“…33 In the XAS spectra for β = 1.40, 1.60, and 1.80, multiple fine structures were observed, implying that some of the interfacial Mn atoms are oxidized to Mn 2+ as reported in the previous work. 27 Therefore, we have subtracted the oxide component Mn 2+ -like multiplet structures such as MnO from the raw data for the β = 1.40, 1.60, and 1.80 films, as described in detail in the Appendix. The bulk properties of these films could be different from those of the interfacial region.…”

“…27 To determine the Co magnetic moment, we again used the sum rules 36,37 assuming the 3d hole number (n h ) to be 2.2. 20,24,27 As shown in Fig 4(c), we experimentally determined the Co spin magnetic moment m spin (Co) to be in the range of 1.2 μ B -1.4 μ B , which is obviously larger than the theoretically predicted bulk value of 1.06 μ B . 7,39 This can be explained by the presence of Co Mn antisites, which are predicted to have larger spin magnetic moments than Co atoms at the regular sites.…”