Abstract
Exchange bias (EB) is a fundamental phenomenon in widespread information technologies. However, a comprehensive understanding of its microscopic origin remains a great challenge. One key issue in the debate is the role of frustration and disorder in the EB mechanism, which motivates the exploration of the EB effect in spin glass (SG) systems. Here, in the SG state of Cr-doped Hund’s metal $CsFe_2As_2$, we discover a giant EB effect with a maximum bias field of ~ 2 Tesla, which is almost two orders of magnitude larger than that of traditional alloy SGs. Our results indicate that the giant EB effect should originate from the exchange interactions at the natural boundaries between the tunable ferromagnetic-like (FM) regions around the Cr dopants and the SG matrix, via which the FM spins are strongly pinned by the frozen spins in the SG matrix. In addition, the temperature-dependent and cooling-field-dependent EB behaviors could be interpreted well by the SG model with frustrated FM/SG boundaries, which provides an intuitive and explicit understanding of the impact of glassy parameters on the EB effect. All these results suggest that Hund’s metal or other correlated metals are promising directions for exploring the EB effect in the SG state.
Introduction
Exchange bias (EB) is a cornerstone of modern magnetic technology, essential for the functionality of read heads in hard drives and magnetic random-access memory (MRAM). This phenomenon occurs when the magnetic orientation of a "soft" ferromagnetic layer is pinned by an adjacent "hard" magnetic layer, resulting in a shift of the material's magnetic hysteresis loop. Despite its widespread application, the microscopic origins of EB—specifically the roles of frustration and disorder—remain poorly understood.
Spin glasses (SGs), characterized by magnetic moments "frozen" in random orientations due to competing interactions, serve as ideal systems for studying these effects. However, EB fields in pure spin-glass systems are typically negligible (around 0.01 T). This research investigates whether Hund’s metals—materials with strong electron correlations—can host a more robust, "giant" exchange bias.
Methods
The researchers synthesized high-quality single crystals of Cr-doped and Co-doped $CsFe_2As_2$ using a self-flux method, focusing on doping levels of $x = 0.06$ and $x = 0.10$. Magnetic characterization was performed using SQUID-VSM and PPMS-VSM in fields up to 14 Tesla. The team employed Field-Cooling (FC) and Zero-Field Cooling (ZFC) protocols to determine the spin-freezing temperature ($T_f$). Additionally, thermoremanent magnetization (TRM) measurements were used to observe the slow relaxation characteristic of spin glasses, and training effect cycles were analyzed to confirm the intrinsic nature of the observed exchange bias.
Results and Discussion
The study revealed a giant exchange bias effect in the $x = 0.06$ Cr-doped sample. At 2 K, the material exhibited a maximum exchange bias field ($H_{EB}$) of approximately 2 Tesla, nearly two orders of magnitude larger than traditional spin-glass alloys. This puts the system among the highest-performing inhomogeneous EB materials known.
The origin of this giant EB is attributed to "natural" boundaries between ferromagnetic-like (FM) clusters formed around Cr dopants and the surrounding spin-glass matrix. Unlike manufactured bilayers, these internal interfaces provide strong pinning. The researchers utilized a model distinguishing between "frozen" spins (providing the bias) and "rotatable" spins (contributing to the coercive field). They found that frozen spins relax significantly slower than rotatable ones, ensuring a stable bias.
The EB field is highly tunable via the cooling field ($H_{FC}$). However, excessive cooling fields can cause FM regions to coalesce, destroying the SG matrix and reducing the bias. Furthermore, increasing Cr doping to $x = 0.10$ decreased the EB field, as the expanded FM regions began to overwhelm the pinning matrix.
Conclusion
The discovery of a 2 Tesla exchange bias in a doped Hund’s metal provides a new pathway for developing spintronic materials. While the effect is currently restricted to low temperatures (below the 14 K freezing temperature), it demonstrates the potential of correlated metals. Future research will aim to increase the blocking temperature, perhaps by integrating these materials into artificial heterostructures with high magnetic anisotropy.
Summary updated