Determining how far a spin current propagates in a given material is a key fundamental endeavor in spintronics. In a typical ferromagnetic metal, a coherent spin current polarized transverse to the magnetization decays within just ~1 nm due to rapid spin dephasing. By contrast, theory predicts much longer transverse spin coherence lengths in in metals with alternating magnetic moments (e.g., antiferromagnets and compensated ferrimagnets), which effectively suppress spin dephasing [1]. This mechanism appears to be supported qualitatively by a recent experiment reporting a spin coherence length of >10 nm in ferrimagnetic CoTb [2]. Yet, it remains an open question whether or how a spin current can stay coherent over such a long length scale, particularly considering the well-known strong spin-orbit coupling in CoTb.

In this talk, I will present a quantitative test for the proposed extended spin coherence length (more specifically, dephasing length) in ferrimagnetic alloys. Our test consists of spin pumping measurements on a series of ferrimagnetic CoGd spin sinks [3], which exhibit weaker spin-orbit coupling than CoTb. Our experimental results, corroborated by a modified drift-diffusion model [4], reveal that spin dephasing is indeed suppressed in nearly compensated CoGd – with the dephasing length extended to ~5 nm. Thus, although the magnitude of the apparent spin coherence length evidently depends on experimental details, our findings validate the enhancement of transverse spin coherence enabled by antiferromagnetic order.  

[1] A. S. Núñez et al., Phys. Rev. B 73, 214426 (2006); V. Baltz et al., Rev. Mod. Phys. 90, 015005 (2018).
[2] J. Yu et al., Nat. Mater. 18, 29 (2019).
[3] Y. Lim et al., arXiv:2001.06918.
[4] T. Taniguchi et al., IEEE Trans. Magn. 44, 2636 (2008); K.-W. Kim. Phys. Rev. B 99, 224415 (2019).