Paper detail

HAMR Thermal Reliability via Inverse Electromagnetic Design

Heat-Assisted Magnetic Recording (HAMR) has promise to allow for data writing in hard disks of beyond 1 Tb/in2 areal density, by temporarily heating the area of a single datum to its Curie temperature while simultaneously applying a magnetic field from a conventional electromagnet. However, the metallic optical antenna or near-field transducer (NFT) used to apply the nano-scale heating to the media may self-heat by several hundreds of degrees. With the NFT reaching such extreme temperatures, demonstrations of HAMR technology observe write-head lifetimes that are orders of magnitude less than that required for commercial product. Hence, thermal reliability of the NFT is of upmost importance. In this paper, we first derive fundamental limits on the self-heating of the NFT to drive design choices for low temperature operation. Next, we employ Inverse Electromagnetic Design software, which provides deterministic gradient-based optimization of electromagnetic structures with thousands of degrees of freedom using the adjoint method. The Inverse Design software solves for unintuitive solutions to Maxwells equations, and we present computationally generated structures for HAMR write-heads that offer a 40 percent or 170 C reduction in NFT self-heating compared to typical industry designs.