Research profile: Mark Johnson

Spintronics is a multidisciplinary field that includes materials physics, condensed matter physics (magnetism) and device physics. It is a dynamic field where the results of basic research can have implications for production lines on a time as small as two or three years. In broad terms, Spintronics uses conduction electron spin as a state variable in digital electronic devices. The dominant commercial application is Hard Disk Drive (HDD) information storage. An important emerging technology is integrated magnetic random access memory (MRAM). Many new applications also are the subject of research.

 

Applied research is very active and is dominated by industrial labs including Western Digital (WD), Hitachi Global (now owned by WD), IBM and Seagate. A relatively large portion of basic research is driven by curiosity only, but an equal portion of basic research is very closely related to applications. To a large extent, then, both basic and applied research are closely related to applications. This has been the foundation for the success of Spintronics, and this will continue for the next 5 years or more.

 

Applied research is driven by the competitive need to make small devices (nanometer size scale) that are capable of switching their magnetization states at fast speeds. It follows that the topics of greatest interest in basic research involve (i) the fabrication and characterization of nanostructures, (ii) studies of high-speed magnetization dynamics and (iii) thermal effects related to magnetization dynamics.

 

In the HDD research area, `perpendicular media’ (magnetic media with magnetization perpendicular to the plane of the media) have been used in production for several years. The focus for coming years is to create media with stronger magnetic anisotropies and smaller domain sizes, perhaps using media that are lithographically patterned and have a very small feature size. A closely related goal is the development of read/write heads with smaller dimensions, small enough to write to, and read from, the media with nanometer sized encoded bits.

 

Memory is an important emerging technology, and MRAM has found commercial success in several niches that require nonvolatility, high performance and high endurance. Industry’s search for the ideal memory has been active for the last decade. The competition to create a universal memory will continue for more than 5 years. The success of MRAM will require development of a low power write process, and there is tremendous activity involving the topic of the spin transfer torque (STT) mechanism.

 

The most promising MRAM approaches involve magnetic tunnels junctions (MTJs) with MgO barriers and ferromagnetic electrodes using ferromagnetic metals and alloys having magnetization perpendicular to the film plane. Prototype perpendicular MTJ memory cell arrays have been fabricated and successfully operated with a feature size as small as 7.5 nm and a cell size of 10 f2 or less (where f is the lithographic feature size). Because MRAM cells have nearly infinite endurance and read and write times of approximately 1 nsec, these prototype MRAM arrays are highly promising candidates for a universal memory.

 

A closely related approach is thermally assisted MRAM. Prototype devices have been fabricated with feature sizes of 10 nm or less and successfully operated. Evidence has been shown that these structures are scalable to f = 10 nm. The `racetrack memory’ is a completely different approach to integrated MRAM, and has been promoted by IBM for the last ten years. In the present state of research, racetrack memory prototypes are fabricated on top of the wafer plane, have relatively large feature sizes (about 60 nm) and have read/write speeds that are not yet competitive. However, interest in the racetrack memory has motivated basic research on topics of domain wall nucleation, motion and pinning.

 

Mark Johnson