Development of perpendicular magnetic anisotropy of Co-based alloy for magnetic tunnel junctions

Abstract

The magnetic thin film applied to the free layer, which is a component in constructing the MTJ applied to the storage layer of the memory data in the STT-RAM (Spin Torque Transfer-Random Access Memory), is one of the most important part to achieve high integration. The magnetic layer of the conventional MTJ is a material and a structure having in-plane magnetic anisotropy. In order to achieve high integration to replace the memories like DRAM, and NAND-Flash, currently occupying the market, it is necessary to have a high thermal stability and an aspect ratio of 1. However, when the aspect ratio is set to 1 in the in-plane MTJ, multi-domain could be formed by a strong anti-magnetic field, which causes a serious problem in realizing a uniform switching operation. Therefore, in order to secure stable switching characteristics, the aspect ratio must be made large, which is a disadvantage for integration. In order to solve this problem, introduction of material having perpendicular magnetic anisotropy into the reference layer and the free layer has been proposed. In the case of perpendicular MTJ, magnetization twist phenomena such as multi-domain formation and vortex are not observed even if the aspect ratio is set to 1. In principle, it is also advantageous in lowering the write current and strong hybridization between the insulating layer and ferromagnet such as Co, and Fe. Due to the strong interfacial energy induced by the hybridization, high thermal stability can be achieved by having large anisotropic energy. This is expected to satisfy the scaling requirement of MTJ for high integration, and therefore, the aim of this research is to develop magnetic thin film materials and structures with superior perpendicular magnetic properties. Although there are many materials exhibiting perpendicular magnetic anisotropy, few materials meet the requirements for application to MRAM devices. Rare earth-transition metal (RE-TM) alloys have a large damping constant and low spin polarization efficiency. In addition, magnetic alloys containing L10 alloy and noble metal have high saturation magnetization values and large attenuation constants, respectively. All of these characteristics require a large write current density value when inputting data. Therefore, when using the materials as a magnetic layer, especially in the case of a free layer of MTJ, there is a problem of power consumption due to high write current, breakdown problems, difficulties in compatibility with CMOS with low write current density. It is known that the perpendicular magnetic anisotropy is induced in CoFeB in the junction structure of CoFeB, which is an amorphous alloy added with boron, and MgO, which is a tunnel barrier, and provides a technique to solve the problems. As a result, CoFeB is used as a general amorphous magnetic material as a magnetic thin film of MTJ. However, the addition of B, a non-magnetic element, leads to a loss of magnetization efficiency of CoFe, leading to a decrease in the polarization ratio which determines the TMR of MTJ. That is, there is a need to develop a new material having more excellent properties to solve such a problem. In this study, we use the Zirconium (Zr), expected to have a high polarization for applying as a soft magnetic material because its d-orbit is partially empty. We have developed amorphous magnetic materials (CoZr, CoFeBZr) with perpendicular magnetic properties by introducing transition metal of Zr having excellent soft magnetic properties into Co, Fe or CoFe based magnetic materials. When the size of the device is scaled to a pattern of several tens of nanometers, the problem of the thermal stability being rapidly deteriorated is emerging. In order to ensure the reliability of the device for 10 years, it is necessary to obtain a thermal stability constant of 40 or more in the nanopattern size. However, in the present system, it is difficult to realize a thermal stability constant of 40 or more in a pattern of several tens of nanometers. Therefore, it is necessary to find a way to improve it. The first way is to increase the interfacial magnetic anisotropy energy induced by the hybrid coupling of the ferromagnet (FM)-3d orbitals with the O-2p orbitals, and the second is to increase the effective thickness of the magnetic layer. In this study, we use two methods to improve the thermal stability. As a first study, development of perpendicular magnetic anisotropy using Fe/MgO structure was performed. In this study, PMA characteristics were obtained by using the interfacial anisotropy energy induced by the hybrid coupling at the interface between MgO and magnetic layer Fe, which is used as a tunnel barrier for separating free layer and reference layer. It is known that the interfacial anisotropy energy induced between Fe/MgO is theoretically about 1.6 times larger than that induced by Co/MgO. Therefore, When this structure is applied to the MTJ, higher PMA formation can be expected compared with an alloy such as CoFe (B) in which Co is mixed. Thus, Co/Fe dual magnetic thin film structure is used in this study. Co and Fe layers, which are not the conventional CoFe (B) alloy structure, are deposited, separately. Since Co has a lower saturation magnetization than Fe, introduction of a Co layer could reduce the magnetic moment per unit volume and at the same time enables the development of a PMA with better characteristics than the existing one using the large interfacial anisotropy induced at the Fe/MgO interface. In addition, it is expected to have a high polarization due to the absence of nonmagnetic material such as B, while minimizing the magnetic moment increase due to the formation of the CoFe alloy. This, in turn, maintains compatibility with existing MRAM processes and implements PMA magnetic films that meet the requirements for MTJ integration. As a second study, development of perpendicular magnetic anisotropy using a double-MgO structure was conducted. When this structure is applied as a free layer of MTJ, there are three advantages in comparison with the single-MgO structure. First, there is an effect of increasing spin efficiency. When a free layer is deposited between MgO layers and two reference layers having a different direction are stacked on the MgO surface opposite to the free layer, the amount of accumulated spin charge is increased, resulting in an increase in spin efficiency .Second, because there is an insulating layer on both sides of the free layer, there is no seed layer, and the effect of spin pumping, which causes an increase in damping constant, can be eliminated. And this structure can also enhance thermal stability. Since this structure forms the interface between the free layer and MgO on both sides, it is possible to increase the interfacial energy induced at the interface. By using a thin film spacer, the free layer can be made thicker and the volume of the magnetic layer can be increased. Increasing the interfacial energy and increasing the volume are the most important factors for improving the thermal stability. Therefore, when used, the thermal stability can be improved by about 90% when compared with the single-MgO structure. In this study, it is expected to develop the structural core technology of MTJ for commercialization, aiming at lowering the write current for data input by using double-MgO by improving the thermal stability by securing strong perpendicular magnetic property. Therefore, this thesis has proved the possibility of replacing existing technologies and applying them to high-efficiency and high-integration devices by developing perpendicular magnetization characteristics through research of magnetic materials and structures. These findings are expected to add to the competitiveness of STT-RAMs as new memories in core memory industry that are far from being integrated.