Visually informed guesses of Chain-Of-States methods to find saddle points of the energy landscape in chiral magnetic nanodisks.

Magnetic systems with an interfacially-induced Dzyaloshinskii-Moriya Interaction (DMI) are currently the focus of active research as candidates for devices in information technologies [1-6]. A crucial figure of merit is the height of the minimum energy barrier that separates adjacent minima, since it determines stability against thermal fluctuations. It is difficult to determine analytically the transition states for thermal activated switching. This difficulty is exacerbated in the presence of DMI which roughens the energy landscape and introduces many minima. Numerically, a family of numerical algorithms relies on the evolution of chains-of-states to identify transition states [7-9]. These methods use initial guesses for transition paths that evolve downwards in energy until convergence is achieved. We highlight the importance of considering visual features of energy minima in the initial guesses for chain-of-states. Two advantages of our improved guesses are a dramatic reduction of relaxation steps, and avoidance of topological defects that correspond to singularities of the energy landscape.