figshare
Browse
Liang_cmu_0041E_10275.pdf (7.47 MB)

Integrated HAMR Light Delivery System via High Q Resonators

Download (7.47 MB)
thesis
posted on 2018-07-24, 00:00 authored by Hongliang LiangHongliang Liang
In the age of Big Data, the demand for higher capacity hard disk drives (HDDs) is ever increasing. To enable higher recording densities, heat assisted magnetic recording (HAMR) is being investigated as a promising approach. This technology uses a high coercivity material for the recording medium to overcome the superparamagnetic limit and a laser in the recording head to temporarily heat the medium above the Curie temperature at locations where data needs to be written. Because the generated heat spot must be in the tens of nanometers, the laser must be coupled to a near field transducer (NFT) to reduce its spot size. While numerous laser-to-NFT coupling schemes have been proposed, they involve little monolithic integration and have low coupling efficiency. We desire to improve the laser-to-NFT coupling efficiency such that optical power output requirement decrease by an order of magnitude to as low as 1 mW. We present a HAMR light delivery scheme in which we first optimize our laser through mode stabilization then couple it to the NFT by means of a waveguide and a high Q dielectric resonator, with all components placed onto the same chip. This level of integration may lead to lower manufacturing costs and a reduction in system complexity.
We begin our study by fabricating a Fabry-Perot GaAs/AlGaAs quantum dot bar laser. To optimize our laser, we first address the issue of power fluctuation as it directly affects the quality of the data written in HAMR. We minimize power fluctuation by enforcing single mode operation on the laser, achieved through a novel method of coupling the laser to a periodic array of gold NFT-like metal nanostructures. We found that at coupling distance of 50 nm the array was able to completely suppress any appearance of secondary modes as observed through the laser emission spectrum. We continue our optimization by examining what is needed to allow the laser to operate in continuous-wave (CW) rather than pulsed to enable continuous data writing. Based on the pump duty cycle and thermal resistance of our structure, we examine different heat sinking topologies and estimate the amount of heat sinking necessary to make possible CW operation. We determine CW operation can be achieved by either using a copper heat sink or reducing the laser ridge width down to 1 μm.
The final study examines the coupling efficiency of the waveguide-resonator-NFT system using Si3N4 core/SiO2 cladding. We achieve critical coupling between a rectangular waveguide and a disk resonator using a coupling distance of 100 nm, disk radius of 11 μm, and waveguide width of 1 μm. By examining the effect of the NFT on the quality factor of the resonator, we estimate a coupling efficiency from laser to NFT of 50%. Thermal characterization through resistance measurements is performed to verify launch of plasmonic modes on the NFT. The measurement results combined with simulations in COMSOL suggest a propagation length of 6 to 8 μm, which is consistent with that of a mode being launched.
In summary, this work presents an analysis of a complete HAMR light delivery system design and supports key aspects of that design with experimental confirmation of design assumptions. Specifically, the design includes a mode stabilized laser operating in CW and coupled to a NFT through a waveguide-resonator system. The coupling efficiency is improved such that the power requirements of the source laser decreases by more than an order of magnitude compared to those of current coupling schemes. The level of integration of the system can lead to lower costs and complexity for the HAMR recording head.

History

Date

2018-07-24

Degree Type

  • Dissertation

Department

  • Electrical and Computer Engineering

Degree Name

  • Doctor of Philosophy (PhD)

Advisor(s)

James Bain

Usage metrics

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC