UC Berkeley

The hard disk drive (HDD) is still the dominant technology in digital data storage due to its cost efficiency and long term reliability compared with other forms of data storage devices. The HDDs are widely used in personal computing, gaming devices, cloud services, data centers, surveillance, etc. Because the superparamagnetic limit of perpendicular magnetic recording (PMR) has been reached at the data density of about 1 Tb/in^2 , heat assisted magnetic recording (HAMR) is being pursued and is expected to help increase the areal density to over 10 Tb/in^2 in HDDs in order to fulfill the future worldwide data storage demands.

In HAMR, the magnetic media is heated locally (~50nm x 50nm) and momentarily (~10ns) to its Curie temperature (~750K) by a laser beam. The laser beam is generated by a laser diode (LD) and focused by a near field transducer (NFT). But the energy and mass transfer at high temperature from the laser heating can cause potential reliability issues. The design temperature of the NFT is much lower than the media’s Curie temperature. However, the distance between the NFT and the media is less than 10nm. As a result, the heat can flow back from the media to the NFT, which is called the back-heating effect. This can cause undesired additional temperature increase on the NFT, shortening its lifetime. Additionally, depletion, evaporation and degradation can happen on the lubricant and the carbon overcoat (COC) layer of the media. The material can transfer from the media to the head at high temperature and cause solid contamination on the head, adversely affecting its reliability.

Since the laser heating in HAMR happens at nanoscale spatially and temporally, it is difficult to measure experimentally. In this dissertation, a comprehensive experimental stage, called the Computer Mechanics Laboratory (CML)-HAMR stage, was built to study different aspects of HAMR systems, including the heat and mass transfer in the head-disk interface during laser heating. The CML-HAMR stage includes an optical module, a spinstand module and a signal generation/acquisition module. And it can emulate the HAMR scenario.

The head’s temperature was measured during the laser heating using the stage and heads with an embedded contact sensor (ECS). It was estimated, based on a linear extrapolation, that the ECS temperature rise is 139K, 132K, 127K and 122K when the disk is heated to the Curie temperature (~750K) and the head-disk clearance is 0nm, 1nm, 2nm and 3nm, respectively. The heating effect of the ECS was also studied and a related heat transfer experiment was performed. The normalized ECS self heating temperature rise, an indicator of the heat transfer in the head-disk interface (HDI), was measured. It was concluded that the heat transfer coefficient across the HDI strongly depends on the width of the gap size, especially when the gap size is smaller than 1nm.

The head disk interaction during the laser heating was studied using a waveguide head, i.e., a HAMR head without the NFT. It showed that the laser heating can cause head surface protrusion. This lowers the fly-height (FH) and results in early touchdown (TD). It was shown that the ratio of touchdown power (TDP) change to the laser current is 0.3mW/mA. The dynamics of the head also changes during the laser heating. It was found that the magnitude of the 1st-pitch-mode vibration on the head increases over time both in short term and long term. The accumulation of material transferred to the head was also investigated. It was found that the solid contamination caused by the laser heating forms in the center of the waveguide. The round-shaped contamination formed on the head surface after laser heating.

Finally the disk lubricant reflow after laser heating was studied. In the experiment, a beam of free space laser shines on the rotating disk at different laser powers, disk rotating speeds and repetitions. Then the disk was examined by an optical surface analyzer (OSA). It was found that 80% of the displaced lubricant recovers within 20 minutes. A simulation was also performed. The experiments and the simulation are in good agreement.