Illustration: Courtesy of R. McMichael and M. Stiles

Figure 1: Micromagnetic model of a vortex-type domain wall being driven down a magnetic ($Ni80Fe20$) wire by an applied magnetic field. The wire is $20nm$ wide and $20nm$ thick. (a) The color indicates the in-plane angle of the magnetization, and the arrows indicate the approximate magnetization direction. (b) The motion of the domain wall and the reversal of the core polarity at the wire edge in a constant applied field of $μ0Happl=3mT$. Colors indicate the in-plane magnetization angle, and the altitude corresponds to the out-of-plane component of the magnetization. From left to right, the vortex is near the center of the stripe moving downward [$0.0ns$, same as (a)], approaches the bottom edge $(2.5ns)$, annihilates $(3.5ns)$ and reforms with negative polarity while moving away from the edge $(5.0ns)$. (c) Shows the trajectory of the vortex core as it traces out a zigzag path down the wire, changing polarity and direction at the edge. The red (blue) parts of the trajectory indicate that the core is pointing into (out of) the plane. The arrows indicate the direction of motion with time. The heavy black dots indicate the positions of the core in the images in (b). Both the vortex core polarity and the transverse vortex velocity, $vy$, change sign at the wire edge. Therefore, the predicted domain-wall motion voltage does not change sign because it is proportional to the product of the two.