Magnetic skyrmions are “topologically protected” solitons, that have recently received increased attention due to their potential application as information carriers and for unconventional computing 1. While the current-driven dynamics has been explored deeply, the theory of temperature gradient-induced dynamics - Skyrmion-Caloritronics 2 - is still at its early stages of development but it is particularly promising due to its low energy consumption. In this work, we study the effects of thermal gradients on skyrmion motion in different systems (single-layer FM with interfacial Dzyaloshinskii-Moriya interaction (IDMI), multilayer, and synthetic antiferromagnets (SAFs)) inspired by the experimental results of Ref. 2 and from a fundamental point of view. We observe that, when driven by the entropic torque, skyrmions in single layers move from the cold to the hot region with a finite skyrmion Hall angle (Fig. 1), while in multilayers they move in the opposite direction (from hot to cold region) (Fig. 2). The latter is in qualitative agreement with the experimental observations (Fig. 3a in Ref. 2). We demonstrate that this difference is due to the distinct scaling relations characterizing the two systems, and to the existence of a magnetostatic field gradient linked to the variation of saturation magnetization (MS) which cannot be neglected in magnetic multilayers. The numerical results are corroborated by a generalized Thiele’s equation developed for this scenario. Moreover, we show that the skyrmion Hall angle is completely suppressed in SAFs, similarly to the current-driven skyrmions 3,4. Our results have fundamental implications in the future development of skyrmionic devices combining thermal gradients and SOTs where the proper temperature dependence of the parameters should be taken into account. See Ref. 5 for founding statement.

References 1 J. Zázvorka, et al., Nat. Nanotechnol. 14, 658 (2019).

2 Z. Wang, et al., Nat. Electron. 3, 672 (2020).

3 X. Zhang, et al., Nat. Commun. 7, 10293 (2016).

4 R. Tomasello, et al., J. Phys. D. Appl. Phys. 50, 325302 (2017).

5 This work was supported by the Project No. PRIN 2020LWPKH7 funded by the Italian Ministry of University and Research, and by the PETASPIN association (www.petaspin.com). JB acknowledges support from the Royal Society through a University Research Fellowship. E.R. acknowledges the economical support of the COST Action CA 17123 (MAGNETOFON) within the STSM program. We would like to acknowledge networking support from COST Action No. CA17123 “Ultrafast opto magneto electronics for nondissipative information technology.”

6 W. Li, et al., Adv. Mater. 31, 1807683 (2019).