Home >  About ISSP >  Publications > Activity Report 2018 > Xxxx Xxxx

Visualization of Electrically Controllable Magnetic Domains in a Quasi-One-Dimensional Quantum Antiferromagnet BaCu2Si2O7

PI of Joint-use project: Kenta Kimura
Host lab: Masuda Group

Quasi-one-dimensional quantum antiferromagnets (q1D-QAFMs) consist of chains of magnetic ions with a small spin quantum number (S = 1/2 and 1), where the antiferromagnetic interactions within the chains are much stronger than the interchain interactions. q1D-QAFMs have attracted attention for their potential to exhibit exotic phenomena (e.g., spin liquid behavior and topological spin excitations) and their possible application in quantum spintronic technology. Like other types of antiferromagnets, q1D-QAFMs will form a pair of domain states when undergoing an antiferromagnetic transition due to interchain interactions, typically represented by an “up-down-up-down” state and a “down-up-down-up” state. These domains are randomly distributed in a single crystal sample, and their observation and manipulation are critical for device applications. However, observing domain patterns in q1D-QAFMs appears to be challenging, not only because of the absence of net magnetization, but also because of strong quantum fluctuations that significantly reduce the ordered components of individual spins. Indeed, there has been no experimental observation of the domain pattern in q1D-QAFMs.

In the present study [1], we successfully visualize antiferromagnetic domains in one of the most representative spin-1/2 q1D-QAFMs, BaCu2Si2O7, which exhibits an antiferromagnetic order below TN = 9.2 K with reduced ordered moments of only 0.1 μB per Cu2+ ion [2]. We use a recently developed optical imaging technique based on nonreciprocal directional dichroism (NDD) [3]. The NDD is a change in optical absorption by reversing the direction of light propagation or the sign of magnetic order parameters. As a result, antiferromagnetic domain patterns can be visualized as a difference in the intensity of transmitted light, as illustrated in Fig. 1(a). For materials to exhibit the NDD, both space-inversion and time-reversal symmetries must be broken. For BaCu2Si2O7, we expect such a symmetry requirement to be fulfilled by a combination of the antiferromagnetic order and a zigzag chain arrangement of the magnetic Cu2+ ions. In the following experiments, we used single crystals of BaCu2Si2O7 grown by the Masuda group using a floating zone method.

masuda-a2-fig1.png
Fig. 1. (a) Schematic illustration of the visualization of the antiferromagnetic domains in a quasi-one-dimensional quantum antiferromagnet via the nonreciprocal directional dichroism. Yellow transparent cylinders denote light beams propagating along the direction of the thick black arrow. The diameter of the cylinders represents the intensity of the light beams. In the antiferromagnetic phase, the spins or magnetic moments (green arrows) are in principle antiparallel. However, the parallel alignment of the magnetic moments can appear as a defect called a domain wall (red line). (b) Optical microscopy image of a sample at 5 K. (c) Visualization of electric-field-driven displacement of antiferromagnetic domain walls. (d) Schematics of a domain wall parallel (DW||) and perpendicular (DW⊥) to the spin chains.

Figure 1(b) shows the optical microscopy image of a thin plate sample taken at 5 K below TN. A clear two-level contrast, which was absent in the paramagnetic phase, appears in the antiferromagnetic phase, indicating the coexistence of the opposite antiferromagnetic domains in the sample. The image reveals that the domain walls (DWs) separating opposite antiferromagnetic domains run predominantly along the direction of the spin chains. When the sample was heated above TN and then cooled back, the domain pattern was changed but the direction of the domain walls was maintained. This indicates that the anisotropy of the DWs is robust for BaCu2Si2O7. Furthermore, we have shown that the antiferromagnetic domains can be controlled by an applied electric field in combination with a small bias magnetic field. As highlighted by the dotted lines in Fig. 1(c), the application of positive and negative electric fields shifts the DWs in opposite directions. Such electric-field induced displacement of the magnetic domains can be explained by a magnetoelectric coupling allowed by broken space-inversion and time-reversal symmetries. Importantly, the direction of the DWs is maintained during the displacement. This indicates that the DW anisotropy also governs the electric-field driven DW motion in BaCu2Si2O7.

On a qualitative level, the observed DW anisotropy can be explained in terms of the anisotropy of the magnetic interactions. Figure 1(d) shows schematic spin arrangements near the DW parallel and perpendicular to the spin chains, denoted by DW|| and DW⊥, respectively. The formation of DW⊥ costs the energy governed by the strong intrachain interaction, which is 2 orders of magnitude larger than the interchain interactions governing DW||. This suggests that DW|| is preferable, in agreement with the experimental observation. However, we find that the anisotropy of the magnetic interactions is much larger than the anisotropy of the DWs, which is estimated as the ratio of the lengths of DW|| (LDW||) and DW⊥ (LDW⊥): LDW||/ LDW⊥ ~ 10. This implies that another factor should also be considered to understand the DW formation. To our knowledge, there is no relevant theory available. Future work is needed to understand the microscopic origin of the DW anisotropy in the q1D-QAFMs.

In conclusion, the present study will contribute to the understanding of the domain physics of q1D-QAFMs. It also raises an interesting question as to whether the domain pattern observed in BaCu2Si2O7 is a material specific property or intrinsic to q1D-QAFMs.


References
  • [1] M. Moromizato T. Miyake, T. Masuda, T. Kimura, and K. Kimura, Phys. Rev. Lett. 133, 086701 (2024).
  • [2] M. Kenzelmann, A. Zheludev, S. Raymond, E. Ressouche, T. Masuda, P. Böni, K. Kakurai, I. Tsukada, K. Uchinokura, and R. Coldea, Phys. Rev. B 64, 054422 (2001).
  • [3] K. Kimura and T. Kimura, APL Mater. 11, 100902 (2023).
Authors
  • K. Kimuraa
  • aOsaka Metropolitan University