At the recent LithoVision technical symposium Aaron Bowser, Senior Lithography Manager at Seagate Technology, shared valuable insights into the unique challenges and innovations in lithography for leading-edge hard disk drives (HDD). He also identified a number of commonalities with IC manufacturing.
Bowser highlighted that hard disk drives are enabling the expanding global datasphere, with growth projected to skyrocket from 16ZB in 2016 to 163ZB in 2025 (Reference: IDC Data Age 2025 study, sponsored by Seagate www.DataAge2025.com). To support this growth trajectory, areal density must continue to increase as it has in the past. However, the growth rate for traditional perpendicular magnetic recording (PMR) is actually slowing as it reaches the thermal stability limits of smaller bit sizes using current recording media. To overcome this challenge, Seagate’s recording technology is transitioning to heat assisted magnetic recording (HAMR), which involves temporarily heating the disk material during the write process. HAMR technology has the capability to extend the areal density to 5TB/in2 by 2025 (Figure 1A).
During the HAMR process, a near field transducer (NFT) is used to temporarily heat the media, decrease its coercivity, and enable creation of strong magnetic fields in smaller regions than hitherto possible. This enables continued bit size scaling for HDD manufacturing (Figure 1B). Key litho requirements from this are CD resolution and edge placement; these are being driven to levels that are similar to those in logic and memory manufacturing.
For CD resolution, Bowser explained that within the thin film head (TFH) design, the reader device remains the most aggressive CD requirement, as its width (Wrdr) is only 50% of the media bit track width (Figure 2A). Reader width patterning will need to transition from current widths of around 25 nm to 15 nm and smaller in the future to support the necessary areal density growth. Similarly, the critical lithography edge placement requirement is the positioning between the NFT peg and the reader (Figure 2B). HAMR’s stringent placement requirements will drive on product overlay mean + 3sigma well below 10 nm in years to come.
Beyond the shrinking resolution and overlay requirements that are familiar to IC makers, Bowser highlighted a number of specialized lithography challenges associated with their processes (Figure 3). The most noteworthy of those is the use of AlTiC substrates. These 200 mm diameter substrates are 1.2 mm thick and drive additional substrate conductivity/thermal-related overlay issues. The unique lapping process employed in HDD TFH manufacturing places further edge placement error (EPE) restrictions as well. HDD manufacturing also uses very small die sizes with a large number of dies per wafer, in addition to small lot sizes and low wafer starts. These factors enable more relaxed requirements for defectivity and productivity compared to IC manufacturing.
A number of techniques are used to achieve the necessary isolated line resolution scaling, including employing chrome-less phase edge lithography (CPEL) masks and post-litho etch trimming. These require low line width roughness (LWR) resist materials. eBeam direct write and nanoimprint lithography methods are considered to be encouraging solutions as well. As with semiconductor manufacturing, satisfying the increasingly aggressive HDD edge placement and on-product overlay (OPO) requirements relies on a complete fab ecosystem as well (Figure 4A). This will require integration of metrology tooling, correction for asymmetric marks, advanced process control (APC) for overlay as well as field stitching across the bar, higher order process corrections, and thermal management for the AlTiC wafers. In addition, by transitioning semi-critical lithography layers to tools such as the cost-effective Nikon NSR-SF155 scanner field stepper that provides excellent overlay, Seagate is able to reduce cost of ownership (CoO).
Bowser explained that in the future, tightening overlay requirements will necessitate new lithography tooling for hard disk manufacturing (Figure 4B). He prioritized the HDD litho market’s needs as affordable CoO, support for 200 mm AlTiC substrates, and single machine overlay of less than 3 nm. Defectivity and productivity are not high priorities due to the specialized HDD manufacturing processes.
Seagate HDDs are transitioning to HAMR technology to meet the exploding worldwide demand for data storage. Although this is a new process, Bowser reiterated that the fundamental challenges for thin film head manufacturing will continue to be CD scaling of isolated lines, as well as aggressive layer-to-layer and within layer edge placement error specs. In order to achieve Seagate Technology’s future objectives, they will require further overlay improvements on cost-effective lithography tool platforms that are compatible with 8” AlTiC substrates, in addition to even lower LWR resists, and advanced trimming processes.