µMAG at MMM'98, Miami

A special evening session "Standard Problems in Micromagnetics" was held at the 43rd Annual Conference on Magnetism and Magnetic Materials in Miami, FL on November 11, 1998. Several previous µMAG workshops have been held in conjunction with major conferences, but this is the first time the workshop has been included as a regular session with papers reviewed and published in the conference publication. Roughly 100 people attended the session despite the lure of a tropical Miami evening.

Pre-session -- OOMMF public code

The workshop opened with a pre-session presentation by Mike Donahue and Don Porter describing and demonstrating the capabilities of the OOMMF public code. This package has been ported to a large number of operating systems and while it has been designed to allow programmers to write and substitute modules, programming skills are not required to use the package with its large number of options. The OOMMF distribution is currently limited to 2-D arrays of spins, but 3-D arrays are planned for future releases.

MMM'98, Session FZ

Standard Problem #1

The first two talks of the formal session related to solutions of the µMAG Standard Problem #1. Tony Arrott gave a very energetic presentation outlining a number of pitfalls that can occur in micromagnetic computations, including the artificial formation of vortices on Neel walls when the grid is too coarse. Tony also underscored the importance of the propagation of solitons along the particle edges during magnetization reversal.

Standard Problem #2

The last three presentations of the session were on solutions to µMAG Standard Problem #2. Luis Lopez Diaz presented solutions calculated on a 2D square grid using different solvers, and found essentially identical results, but varying performance depending on the state of the magnetization and the particle size. Luis performed a fairly detailed calculation of the remanent state in the range 0.1 < d/l_ex < 30. In particular, his data show a transition in the remanent state from a symmetric end structure to an asymmetric end domain near d/l_ex = 10.

Bob McMichael presented solutions using a 2D square grid using different magnetostatic field algorithms. Constant magnetization methods and constant charge methods give very similar results when extrapolated to zero cell size, but without edge corrections, the constant charge method gives results that depend more strongly on cell size. Results taken over the range 3 < d/l_ex < 75 show different reversal mechanisms for d/l_ex < 30, including reversal initiated by switching of the end domains and vortex formation in the end domains.

Thomas Schrefl presented Problem #2 solutions calculated using a finite element method, comparing results obtained with the field directed along the [111] direction (as in the problem specification) with results obtained with the field along the [100] direction. Vortex formation is suppressed for fields along [111], giving a weak dependence of H_c on the particle size up to d/l_ex = 30. In contrast, for fields along [100], vortices form, and H_c drops for particle sizes greater than d/l_ex = 20.

The session concluded with two transparencies comparing data from the three Problem #2 solutions. The close agreement between these solutions received an enthusiastic round of applause.

µMAG organization / NIST CTCMS / rmcmichael@nist.gov
03-DEC-1998