Vol.2 No.3 2009

Research paper : Creating non-volatile electronics with spintronics technology (S. Yuasa et al.)−201−Synthesiology - English edition Vol.2 No.3 (2009) first-generation TMR head, and when combined with the latest perpendicular magnetic recording medium, realized an ultra-high-density HDD with a recording density above 250 Gbit/inch2 (twice the previous value)[15]. The future potential for development extends up to the 1 Tbit/inch2 next-generation HDD. By achieving such high HDD recording density, the 3.5 inch drives that are the main devices in use can be replaced with 2.5 inch drives that have adequately high capacity. As a result, 2.5 inch drives are expected to come into mainstream use, even in the market for large-capacity HDD. As described in section 1.1, 2.5 inch drives consume only one-fifth power of the 3.5 inch drive, so replacing 3.5 inch drives with 2.5 inch drives will greatly reduce overall HDD power consumption. The HDD industry has a huge market that compares in scale with DRAM and CPUs (about 3 trillion yen per year). The magnetic head is the most expensive component in a disk drive, and the market value for the magnetic head alone is a huge six hundred billion yen per year. The fact that nearly all of the drives currently being produced are equipped with MgO-TMR heads is the best indication of the social impact of these research results.3.2 Ultimate non-volatile memory “spin-RAM” R & DWith the objective of implementing the “spin RAM” ultimate non-volatile memory that provides large capacity, high speed and high reliability and will serve as the core technology of non-volatile electronics, AIST is working together with Toshiba Corporation and other companies on the NEDO Spintronics Non-volatile Devices Project. Spin-RAM is an MRAM that uses a new physical phenomenon called “spin torque switching” in the data write technique, achieving a higher capacity than the conventional magnetic field write MRAM. To realize a gigabit-class large capacity spin-RAM requires both increased read out by the giant TMR effect of the MgO-MTJ device and a low-power write technique that uses spin torque switching. Spin torque is the torque generated by the transfer of spin angular momentum from the conduction electrons to the local magnetic moment of the magnetic layer when a current flows in the MTJ device. That spin torque can be used to reverse the direction of spin of the ferro-magnetic electrode (i.e., write). Spin torque switching has been implemented before using GMR devices and Al-O barrier MTJ devices, but the current density needed for switching is very high, so achieving a practical effect was considered difficult. AIST implemented spin torque switching in a MgO-MTJ device for the first time in the world in 2005[16][17]. Furthermore, AIST and Osaka University are collaborating to develop a method for quantitative estimation of spin torque[18][19] and have succeeded in verifying a high output microwave oscillator that uses spin torque[20]. Also, Sony, Tohoku University and many others have been doing vigorous research and development on attaining spin torque switching at lower current.Currently, AIST is moving forward with development of the ultimate spin RAM that uses a perpendicularly magnetized MgO-MTJ device that combines newly developed perpendicularly magnetized electrodes and a crystalline MgO tunnel barrier in collaboration with Toshiba Corporation and with the support of the NEDO project. Because this project involves industry, the government and academia and is currently in progress, there are many confidential aspects, so we must omit the details of research and development here, but we have already verified low current and high speed spin torque write operations and excellent data retention characteristics. We are continuing this R & D with the near-term objective of an ultimate non-volatile memory that uses this perpendicularly magnetized MgO-MTJ device and with the long-term objective of implementing ultimate green IT devices through normally-off computer.4 ConclusionWe have described here our impressions based on our own experience of Full Research achievements. The Full Research scenario put forth by AIST is ultimately a conceptualism, and the actual specific methods for executing it must rely on a groping search by the individual persons engaged in R & D in the laboratories. Although conceptualisms are also important, they are not immediately useful in the R & D labs where difficult and pressing problems are being dealt with. Actually linking the results of basic research to commercialization, involves the problem of matching research with social needs and problems of dealing with complex elements such as the difference in interests between organizations and interpersonal relations in addition to the technological problems. Particularly the difference of interests between organizations may create a deadlock, even when the matter is left to the upper levels of the organizations, and there is no progress. In the end, the people in the labs must work within organizations to solve problems, so a relationship of trust among the people doing the R & D in the laboratory is important. That is to say, it is ultimately an interpersonal problem.Speaking of the technical problems, the “valley of death” that lies between basic research and product development is wider and deeper than is imagined, and it is probably impossible for AIST to cross it alone. Particularly in mature industries Fig.11 Structure of the CoFeB/MgO/CoFeB MTJ device. (a)Annealing at 250 ℃ or above initiates crystallizationOriented polycrystallineMgO (001)Amorphous CoFeBOriented polycrystallinebcc CoFeB (001)Oriented polycrystallineMgO (001)Oriented polycrystalline bcc CoFeB (001)(b)Amorphous CoFeB


page 28