An Experimental Investigation on Curie Temperature Variations and Related Switching Field Distributions in Heat-Assisted Magnetic Recording Media

Abstract

Heat assisted magnetic recording (HAMR) is considered to be the next major hard disk drive (HDD) technology for areal densities beyond ∼1.5 Tb/in2 [1]. In HAMR, each high anisotropy magnetic grain is heated close to, or above, its Curie temperature (Tc), then freezed in a new magnetization direction with an effective magnetic field as the grain cools down. As recent theoretical recording simulations suggested, grain-to-grain variability in T C will contribute more to switching field distribution (SFD) than grain-to-grain variability in magnetic anisotropy (H K ) [2]. There were a few methods to measure Tc distribution of HAMR media as previously reported in reference [3] and [4]. In this paper, we report a different measurement method and the study of Tc variations and the related switching field distributions in L1 0 ordered FePt based granular HAMR media. As shown in Fig. 1 (a), a small AC or sine-wave perpendicular magnetic field (or switching field Hsw) is applied to HAMR media that is heated to a range of temperatures by a pump laser beam. The MOKE signal change with increasing pump laser powers Pw (or temperatures T) is measured by a probe laser beam, and finally the peak-to-peak MOKE signal amplitude (which is the total signal from each grain in the entire measured area where each grain has its own coercivity Hc and Tc) is plotted as a function of Pw or T. As schematically shown in Fig. 1 (b), the measured MOKE signal (magnetization change (ΔΜ)) remains small for low temperature and then starts to increase at the 1st temperature (T1) of minimum Tc among the total number of measured grains and finally to decrease to negligence around 2nd temperature (T2) of maximum Tc among the total number of measured grains. There is a maximum MOKE signal amplitude change at the 3r temperature (T3) (between T1 and T2) which is close to the average Tc of the total number of measured grains. This is proposed to be due to 2 competing factors: 1). Media coercivity (Hc) decreases with temperature (so that Hc of some grains is smaller than applied Hsw), resulting in more and more grains just starting to flip (till all grains are switched when T ≥ T3), and thus to make ΔΜ (and measured MOKE signal) to increase; 2). Saturation magnetization (Ms) of flipping grains is decreasing with temperature, and is dropping significantly when approaching Tc, leading to decreased ΔΜ (and measured MOKE signal). The advantages of the proposed method include higher measurement speed without sweeping of large magnetic field and long time heating, and is also a non-destructive measurement approach without troublesome sample preparation.