With the rapid expansion of IoT and digital transformation, data centers demand ever-higher storage densities. Heat-assisted magnetic recording (HAMR) employs L1₀-FePt, whose magnetic anisotropy is an order of magnitude larger than that of conventional CoCrPt, enabling ultrahigh-density recording. Achieving areal densities beyond 4 Tbit/in² requires granular films with grain sizes ~4.3 nm and narrow inter-grain pitch with ~1 nm [1]. However, conventional segregant systems such as FePt-C and FePt-BN have not yet simultaneously satisfied the required microstructural and magnetic criteria.
​​ To overcome this limitation, we introduced a data- driven materials design framework using the NIMS Research Data Express (RDE) platform. By collecting experimental datasets and applying machine learning to FePt-C and FePt-BN systems, we predicted sputtering conditions that led to FePt-BN-C granular films with sub-6 nm grain sizes and coercivities up to 3.7 T. Although iterative prediction cycles improved the microstructure to 4.9 nm grains, the results also clarified the intrinsic difficulty of meeting all 4 Tbit/in² requirements within this materials system alone.
​​ Beyond materials optimization, three-dimensional magnetic recording offers an additional pathway toward higher areal density. As a proof of concept, FePt-C/Ru-C/FePt-C trilayers were fabricated, demonstrating epitaxial stacking and distinct magnetic switching behaviors arising from different ordering states in the upper and lower FePt layers [2]. Strategies to improve the structural and magnetic quality of the upper layer will be discussed.
​​
​​ [1] D. Weller et al., IEEE Trans. Magn. 50, 3100108 (2014).
​​ [2] P. Tozman et al.,Acta Mater. 271, 119869 (2024).
​​
​​
​​
More information :https://www.spintec.fr/seminar-material-development-for-hamr-and-its-prospects/​​
​
Visioconference : https://univ-grenoble-alpes-fr.zoom.us/j/98769867024?pwd=dXNnT3RMeThjYStybGVQSUN0TVdJdz09​​
​​