以氧化物作為穿隧阻絕層的磁性穿隧接面被廣泛的應用於磁電阻式隨機存取記憶體中。因為磁性穿隧接面透過不同的磁矩組態產生極大的穿隧磁阻值，形成磁性記憶體的最小單元，而磁矩組態可以透過外加磁場或利用自旋極化電流控制。自旋極化電流會在磁性穿隧接面中產生自旋轉移力矩，進而影響磁性材料中的磁矩方向來達成控制磁矩組態的目的。 隨著長晶技術的成長，因為氧化鎂晶體自身獨特的電子結構特性，氧化鎂為穿隧阻絕層的磁性穿隧接面成為相當經典的研究題材，無論在實驗與理論中都發現了極大的穿隧磁阻值。而在這項研究中，我們利用第一原理計算來分析鐵/氧化鎂/鐵磁性穿隧接面的電子結構特性以及電子傳輸特性。氧化鎂因為軌域對稱性所形成的Δ_1能階，擁有一個獨特的傳輸通道，所以在這個系統中我們透過傳輸機率可以估計出大約3000%的穿隧磁阻值，並且我們也透過我們所新開發的JunPy來分析系統中的自旋力矩，了解其中控制磁矩方向之運作特性，以及與其他實驗與理論的研究結果相互比較，驗證我們所開發的軟體之可信度，為未來研究其他不同材料的基石。另一方面,我們也嘗試選擇氧化銪為穿隧阻絕層的磁性穿隧接面,其與氧化鎂有相同的Δ_1能階特性同時本身具有自旋極化的特性，我們透過第一原理分析其電子結構特性，未來希望能透過JunPy程序進一步分析其自旋傳輸特性。 ;Magnetic tunnel junctions (MTJs) have aroused intensive studies for applications in non-volatile magnetic random access memories (MRAMs), which have been widely used, and the most important property of MTJs is the tunneling magnetoresistance (TMR) effect. The giant magnetoresistance (MR) ratio originates from the tunneling current depending on the relative orientation of two FM electrodes that can be controlled by a spin-polarized current via the so-called spin torque effect. The spin-transfer torque (STT) and the field-like spin torque (FLST) are two components of the spin torque effect. We employ the first-principles calculation with the non-equilibrium Green’s function (NEGF) method to analyze the electronic and transport properties. For Fe/MgO/Fe MTJs, the orbital symmetry of ∆_1 state in MgO-barrier is responsible for giant TMR effect that is estimated over 3000% at zero bias. Moreover, our newly developed “JunPy” package is employed to investigate both STT and FLST in non-collinear magnetic configurations, which comparable with previous experimental measurements and theoretical calculations and demonstrate the validity of “JunPy” package. These two components, STT and FLST, are responsible for the magnetization switching, which can control the magnetization configurations in MTJs. As a future work, we propose the EuO-based MTJs and expect the ∆_1 state and spin-polarized EuO-barrier may play a significant role on the enhancement for both TMR and FLST, which may create the next-generation FLST-MRAM.
