Vol.2 No.3 2009

Research paper : Creating non-volatile electronics with spintronics technology (S. Yuasa et al.)−199−Synthesiology - English edition Vol.2 No.3 (2009) a preferential orientation of (001) crystal plane formed by sputtering. That device attained a room temperature MR ratio of 220 % (Fig. 9, 3))[9]. Viewed microscopically, the textured MTJ device has basically the same structure as the single-crystal MTJ device, so it can be considered to manifest the giant TMR effect by the same mechanism. To distinguish it from the conventional TMR effect, we call this kind of very large TMR effect exhibited by the crystalline MgO tunnel barrier the giant TMR effect (nomenclature of the authors). The results shown in Fig. 9 1), 2), and 3) (references[7]-[9]) are also presented in the 2007 Nobel Prize in Physics paper[1] and recognized worldwide as historical papers.AIST is conducting various types of Type 1 Basic Research using a high-quality single-crystal MgO tunnel barrier and has succeeded in observing new phenomena not seen with an amorphous Al-O tunnel barrier, such as the oscillation phenomenon of the TMR effect relative to the thickness of the MgO barrier[8][10], interlayer exchange coupling mediated by the tunnel electron[11], and a complex spin-dependent tunneling spectra[12] in addition to the giant TMR effect. Progress in understanding the physical mechanism of these phenomena should lead to further development of the physics of the tunneling effect.2.3 Mass production technology for crystalline MgO-MTJ devicesAs described above, AIST achieved the landmark Type 1 Basic Research result of the giant TMR effect in 2004, but at that time the skeptical outlook on industrial applications for the crystalline MgO-MTJ device was still dominant. The main reason is that neither the single-crystal MgO-MTJ device developed by AIST nor the textured MgO-MTJ device developed by IBM had a device structure suitable for application in practical devices. For application to HDD magnetic heads or MRAM, the lower structure, the “SyF type pin layer” (Fig. 4(c)), is necessary (details are omitted here). However, the basic structure of the SyF-type pin layer is (111) oriented face-centered cubic (fcc) crystal, which has three-fold rotational in-plane symmetry, and it is not possible to grow a MgO (001) layer of different crystal symmetry (an in-plane four-fold rotational symmetry structure) over it. That fundamental problem in crystal growth was a serious problem for fabrication of a MgO-MTJ device over this “practical lower structure”.At first, AIST took the viewpoint of developing a new lower structure that had in-plane four-fold rotational symmetry and tried to sell device manufacturers the idea of MgO-MTJ device technology. However, the reaction from device manufacturers was, “The reliability of the lower structure is directly related to product (HDD and MRAM) reliability. The highly reliable SyF-type pin layer lower structure is the result of ten years or so of research and development, so the development of a new lower structure now is not possible (there is no margin for it).” To be sure, commercialization depends on satisfying many requirements. For example, even for new technology that has outstanding capabilities, just one fatal flaw would prevent commercialization and bring death to the technology, as the word in “valley of death” implies. While we know that to be the greatest difficulty to overcome in Full Research in our minds, the reality of the matter comes when we actually experience it. At the time, there were two straightforward solutions: (i) brute-force development of a new lower structure that has in-plane four-fold rotational symmetry structure and (ii) development of a new tunnel barrier that can be formed above a SyF-type pin layer that has in-plane three-fold rotational symmetry. Either of those solutions, however, would require at least five to ten years of development. Furthermore, the requirement of device manufacturers for “mature technology at a level that can be immediately developed and introduced to the production line” meant that it was practically impossible for AIST to independently develop such a solution. Faced with that situation, AIST began joint development with the production system manufacturer Canon ANELVA Corporation and achieved the “landmark solution” described below.Although we might think of production system manufacturers Fig. 8 Single-crystal Fe/MgO/Fe-MTJ device cross-sectional transmission electron microscope (TEM) image [8].Fig.9 History of room temperature MR ratio improvement.2 nm Fe (001)single-crystalelectrodeMgO (001)single-crystaltunnel barrierFe (001)single-crystalelectrode : : : CoFeB / MgO / CoFeB-MTJTextured MgO - MTJSingle-crystal MgO-MTJCrystalline MgO (001) barrierAmorphous Al-O barrier1) Reference [7]Reference [2]4) Reference [13]3) Reference [9]2) Reference [8]5006004003002001000YearMR ratio at room temperature (%)200520001995Reference [3]


page 26