Unveiling the role of Co-O-Mg bond in magnetic anisotropy of Pt/Co/MgO using atomically controlled deposition and in situ electrical measurement

Abstract

Despite the crucial role of interfacial perpendicular magnetic anisotropy in Co(Fe)/MgO based magnetic tunnel junction, the underlying mechanism is still being debated. Here, we report an anatomical study of oxygen and Mg effect on Pt/Co bilayers through repeated in-situ anomalous Hall effect measurements, controlled oxygen exposure and Mg deposition in an ultrahigh vacuum system. We found that chemisorbed oxygen not only quenches the effective magnetic moment of the Co surface layer, but also softens its magnetic anisotropy. However, a subsequent Mg dusting on the oxygen pre-exposed Pt/Co surface can recover the magnetic anisotropy. The ab initio calculations on the exchange splitting and

Electric-field control of ferromagnetism through oxygen ion gating

“…11 ), and the surface anisotropy energy density can be changed by about 7% while switching between the HRS and the LRS (Supplementary Note 5 ). We note that previous calculations show that the switch of Co-3 d and O-2 p hybridization can lead up to 0.35 erg/cm 2 variation for the surface anisotropy 31 , which would naturally explain the clear magnetic modulation through the electric field controlled interface oxygen ion modulation. Furthermore, we notice that a similar electric field-induced magnetic anisotropy change has been achieved experimentally in the MgO/Fe system 32 , which has been attributed to the electric field controlled orbital hybridization between the Fe 3 d and O 2 p orbitals 32 , 33 .…”

In-situ study of oxygen exposure effect on spin-orbit torque in Pt/Co bilayers in ultrahigh vacuum

“…The layer sequence of the sample is as follows: Si substrate/SiO 2 /Ta (3)/Pt (2)/Co ( t ), where the number inside the parentheses indicates the layer thickness in nm (from now onwards, the Ta seed layer is omitted for simplicity). As we reported previously 30 , Pt (2)/Co ( t ) exhibits in-plane magnetic anisotropy (IMP) when t = 0.2–0.5 nm and clear perpendicular magnetic anisotropy (PMA) when t is in the range of 0.6–1.2 nm beyond which it goes back to IMP. Since this work is concerned about Co films with PMA, we set t at 0.8 nm.…”

“…Subsequently, we exposed the sample to oxygen using the in-situ oxygen source by gradually increasing the dose from 5.5 L to 300 L (here, L is the Langmuir unit with 1 L corresponding an exposure of 1.33 × 10 −6 mbar for 1 s). If we assume a unity sticking coefficient, a full coverage of the energy favourable hcp hollow sites on Co surface requires an oxygen exposure dose of about 5.2 L, and therefore 5.5 L is presumably sufficient to cover the entire Co surface 30 . During the process, the oxygen exposure is paused at selected doses to allow electrical measurements.…”

“…After the oxygen exposure, the Co surface is covered by adsorbed oxygen atoms, though the exact coverage is unknown. Due to the large electronegativity of oxygen, some of the electrons will be transferred from Co to O 30,49 , which results in a larger local asymmetry of charge distribution and thus increases the electric field near the Co surface. The increase of electric field at Co surface naturally leads to a larger Rashba splitting and an enhanced REE, which is directly reflected in the increased FL and DL SOT.…”

“…In all the experiments reported so far, oxygen is incorporated in the form of oxide, and therefore, it is difficult to ascertain if the effects observed are due to the oxide layer in direct contact with the ferromagnet or atomic/molecular oxygen. This is important because in our previous studies we have unveiled that the effects of oxygen and oxide are different when it comes to the influence on perpendicular magnetic anisotropy (PMA) 30 ; oxide capping layer ( e.g ., MgO) enhances PMA of Co whereas oxygen adsorbed on the surface of Co brings about the different effect, i.e ., decreasing the magnetic moment. Therefore, it would be of interest to know if similar effects are present for SOT when the sample is either exposed to oxygen or capped by an oxide.…”